right hemisphere stroke case study

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

The PMC website is updating on October 15, 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Taylor & Francis Open Select

What are the functional outcomes of right hemisphere stroke patients with or without hemi-inattention complications? A critical narrative review and suggestions for further research

Maria stella stein.

a Department of Clinical Sciences , Brunel University , London, UK

Cherry Kilbride

Frances ann reynolds.

Purpose : There is widespread acceptance that patients demonstrating neglect/hemi-inattention (HI) following right hemisphere stroke (RHS) underachieve functionally compared to their counterparts without neglect. However, empirical evidence for this view needs examination. The purpose of this review is to critically appraise relevant studies that compared outcomes from RHS patients with/without hemi-attention and suggest more robust follow-up research. Method : Twelve studies published in 1995–2013 were critically reviewed. Two independent reviewers appraised design features including sample representation, assessment and data analysis methods. Strengths and limitations were highlighted. Results : Results were largely inconsistent. Considerable heterogeneity within patient groups and across studies complicated interpretation. Evidence suggested average group disparity in scores between patients with and without HI at discharge but the cause of functional disparity could not be attributed specifically to HI from the data and modelling results available. Conclusion : The relationship between HI status and functional recovery warrants further investigation in studies with stronger methodology to ensure rigour and robustness in the results. Pending further research, HI status should not be regarded as a key predictor of functional recovery or rehabilitation potential in patients with RHSs. This group should continue to receive appropriate therapeutic intervention aimed at maximising their functional recovery post-stroke.

Implications for Rehabilitation

• Findings from this review demonstrate a paucity of evidence to support the presence of hemi-inattention as a key predictor of functional recovery in patients with right hemisphere stroke; as such, practitioners should take this into consideration when planning rehabilitation programmes of their patients.
• In the initial months following right hemisphere stroke, there are wide-ranging differences in the rate and amount of functional recovery in patients, with and without hemi-inattention. Practitioners should not limit the aspirations of their patients based on the presence or absence of hemi-inattention.
• This review has identified a number of measurement limitations in commonly employed assessment tools for hemi-inattention and overall functional recovery. As such, practitioners should take the limitations of specific measures into account when interpreting the results contextually and with respect to their patients’ situation.

Introduction

Hemi-inattention (HI), commonly referred to as “neglect”, is a complex, heterogeneous and disabling condition which acutely affects up to 80% of patients with right hemisphere stroke (RHS) dysfunction [ 1 , 2 ]. Despite considerable research and advances in the field, HI remains poorly defined as a condition per se . This is supported by the use of multiple descriptors (e.g. unilateral neglect, unilateral inattention) and taxonomies in the literature [ 3–5 ].

Clinically, HI is characterised by reduced attention and/or spatial awareness to details in the environment (commonly towards the left side of the body). HI can affect one or more functional domains (e.g. sensory-motor, visual-spatial) [ 6 , 7 ] and often co-exists with anosognosia and depression [ 8 , 9 ]. HI has been regarded as responsible for delayed and challenging rehabilitation, reduced safety awareness, poor functional outcomes, increase in dependency levels and risk of institution care [ 10–12 ].

Historically, findings from published studies have reported disparity in functional ability scores; with patient groups affected by HI (HI+) underachieving compared to those without (HI−) [ 11 , 13 , 14 ]. Traditionally the cause of this disparity has been largely attributed to the presence of HI, although findings from predictive models have been conflicting and inconclusive [ 10 , 13 , 15–17 ]. This has led to considerable confusion and uncertainty about the clinical importance and significance of differences thought to be associated with HI [ 9 , 18–20 ]. The paucity of relevant evidenced-based reviews has not helped to clarify the predictive role of HI or to promote good rehabilitation practice.

The last systematic review was undertaken by Jehkonen et al. [ 21 ]. The authors focused on the methodological quality of 26 studies published in 1996–2005, which evaluated the impact of Neglect on functional ability in predominantly generic stroke patient samples with mixed lesion sites. Jehkonen et al. [ 21 ] highlighted as an issue considerable differences in patient samples and inconsistencies in results but nonetheless concluded that HI had a significant negative impact on functional outcome, either as an independent predictive factor or in the presence of other variables. Their findings corroborated those of earlier reviews [ 2 , 22 ] which were not specifically focused on the relationship between HI and functional ability. Jehkonen et al. [ 21 ] recommended further research on homogeneous patient groups with respect to right/left hemispheric lesions to improve consistency in the results.

Given the paucity of research in this area, an in-depth evidence-based critical review of relevant studies is offered here with a different approach to that taken by Jehkonen et al. [ 21 ]. However, considering the extent of methodological differences between studies [ 21 ], a narrative review was appropriate. This enabled the inclusion of sufficient, relevant contemporary studies which would have otherwise been excluded by the more stringent selection criteria of a pure systematic review. Narrative reviews “lay out the most recent and best knowledge of various aspects of a problem” [ 23 , p. 427], and are considered appropriate when a diversity of research methods are used in the studies considered as relevant (rather than focusing only on randomised controlled trials), where studies have used different outcome measures and/or non-equivalent samples [ 24 ] and when studies are of relatively poor methodological quality [ 25 ].

The current review examined traditional claims made by previous studies and reviews [ 21 , 22 ] about the negative impact of HI on function; more specifically the strength of the relationship between HI status and functional recovery following RHS. Another aim was to estimate the magnitude of functional differences between HI± patient groups. The current review extended the work carried out by Jehkonen et al. and used a more rigorous and systematic approach to the selection of studies and the review process. Consequently it included fewer ( n  = 12) but more homogenous studies with RHS patient samples. Theoretically similarly designed studies tend to be more comparable than heterogeneous stroke studies.

Both the discrepancy in HI± patient scores and the relationship between HI and functional recovery are of interest to rehabilitation professionals. Together with other indicators (e.g. stroke severity) they may be used to predict likely change in function with time since stroke. This knowledge can guide rehabilitation decisions, e.g. as to which patients are suitable for early supported home discharge schemes. Currently there is an urgent need for reliable predictors and indicators to support the transfer of in-patient rehabilitation services to appropriate stroke survivors in the community. The final aim was to formulate more robust research strategies based on the limitations of studies to date.

A literature search was conducted from 1995 to February 2015 of the databases MEDLINE, AMED, CINAHL, PsycINFO and COCHRANE using several descriptors of neglect subtypes in the literature including HI , spatial, visual, unilateral, personal, extra-personal, motor, sensory, hemi and representational. The words; stroke, CVA, functional* and activities of daily living (ADL) were added to the final search so that studies focused on specific functional activities were included. Children or young adults (≤18 years) and non-human samples were excluded.

The search yielded three Cochrane reviews and 195 publications; AMED (70), CINAHL (86), MEDLINE (102) and PsycINFO (57). In line with the aims of the review, and supported by recommendations from Jehkonen et al. [ 21 ], only studies that compared the homogeneous patient groups with respect to hemispheric lesion site (RHS) and presenting comparisons of patients with or without HI were selected (including intervention designs); all other studies with the heterogeneous patient samples and no HI group comparison were excluded. In addition, functional ability had to be quantifiably measured so that the HI± group differences in scores could be calculated. Two reviewers read the abstracts and, when in doubt, the publication to determine relevance. This process led to the selection of 12 international studies.

The following information (source, aims, design type, demographic data, assessment tools, data analysis method, results and findings) were extracted from each study and are summarised in Table 1 . The critical evaluation process was guided by a checklist described in Appendix 1. Importantly, it focused on the extent of representation of the RHS patient sample with respect to stroke and HI severity levels, time to baseline assessment and follow-up observations, the type of data collected and appropriateness of assessment tools, extent of statistical data analysis undertaken including modelling specifications and, where appropriate, the extent of adjustment undertaken for established confounding factors (e.g. age, stroke severity, time since stroke) and handling of missing data. Each study’s strengths and limitations were identified as part of the review process, these are summarised in Table 1 ; also included is the authors’ assessment of the methodological quality of each study. This was graded (A–D) according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) [ 26 ].

On the GRADE scale; A is high and assigned to well-performed Randomised controlled trials (RCTs) and observational studies with consistent results and/or strong effects. B is moderate – serious flaws in the design in which the estimated effect is likely to be considerably different than the true effect. C is low – studies with serious limitations in which the true effect is likely to be very different than the estimated, e.g. failure to include relevant confounding factors. D is very low – as in C but any estimated effect is very uncertain and highly unlikely to reflect the true effect.

Population studied and demographics

Geographically, studies were undertaken in Canada (1), Italy (3), Israel (2), UK (2), USA (3), USA & Italy (1) with local RHS populations. Two of the studies probably used the same population [ 16 , 27 ].

Age and gender were described consistently; study [ 11 ] made reference to educational background and family burden. The age range varied from 57 (SD 10) [ 10 ] to 60–69 [ 11 , 14 , 16 , 27 , 28 ] and 70–76 years [ 12 , 15 , 29–32 ]. In addition, there was considerable variation in age within specific studies, e.g. 33–88 years [ 11 ] and 40–99 years [ 29 ]. Gender tended to be equally distributed. Morbidity was documented in studies [ 16 , 29 ]; stroke was associated with hypertension, diabetes and heart disease. Stroke severity was not always made clear. It was reported as moderate in two studies [ 15 , 16 ], whilst study [ 30 ] indicated that patients with severe stroke were excluded. Two studies [ 11 , 14 ] recruited only patients with (perceived) good rehabilitation potential. Stroke severity was unreported in seven studies [ 10 , 12 , 27 , 28 , 29 , 31 , 32 ].

Definition of function and “Neglect/HI” syndrome

Conceptually, functional ability/outcome was rarely defined but inferred from ADL measurement scales, mainly the Barthel Index (BI) [ 33 ] and Functional Instrumental Measure (FIM) [ 34 ]. “Neglect/HI” tended to be traditionally defined as a failure to orient, report or respond to stimuli located on the opposite side to the site of the brain lesion which cannot be explained by either primary sensory or motor deficits [ 35 ]. Different studies referred to HI sub-types interchangeably, e.g. visual neglect [ 15 ], unilateral spatial neglect [ 12 ] but effectively measured the same condition because the measurements used cannot differentiate between sub-types of neglect (e.g. visual, spatial and unilateral) [ 3 , 4 , 36 ] but provide an overall measure of the degree or profundity of the condition.

Research settings

Research settings were insufficiently described to allow clear comparison between countries, e.g. termed as a rehabilitation facility or hospital in Israel and a stroke unit in England. They tended to be either acute in-patient hospital and/or community rehabilitation facilities which would suggest research on samples assessed at varying intervals after stroke onset.

Nine studies [ 10 , 14–16 , 27 , 28 , 30–32 ] employed a prospective design, two studies [ 12 , 29 ] employed a retrospective design and one study [ 11 ] employed both. Study [ 11 ] employed a cross-sectional design; most studies [ 5 , 12 , 16 , 28–30 , 32 ] employed a serial design characterised by variable baseline (T0) measures and one follow-up at discharge. Four studies [ 10 , 14 , 27 , 31 ] included up to three follow-up observations. The longest follow-up period was one year since stroke [ 10 ]. All other follow-ups were not fixed in time but varied relative to the discharge point.

Selection criteria

Inclusion criteria tended to be vague; 10/12 studies [ 10–12 , 15 , 16 , 27–29 , 31 , 32 ] included only patients with “good rehabilitation potential” which was not clearly defined. However, by inference it would appear that severely cognitively impaired patients and those with common (age-related) morbidities (e.g. cardio-pulmonary) were automatically excluded early on from most of the studies.

Confirmation of stroke

Stroke was reportedly confirmed by a neurologist in all the studies and by radiological means in 42% [ 10 , 14 , 16 , 28 , 31 ]. Stroke severity was measured in three studies [ 15 , 16 , 27 ] but the scale score was only reported once in [ 16 ] (using the Canadian Neurological Scale – CNS = 7). Aetiologically, infarct was predominant but the majority of studies also included haemorrhage.

Time to first (1st) observation

Baseline measures were taken at variable (non-comparable) times across the studies, making direct comparison difficult. Time to 1st observation was not reported in study [ 29 ] and unclear in [ 32 ]. In hospital settings, initial measurement varied from 7 to 15 days since stroke in studies [ 12 , 15 ], up to 30 days [ 10 , 28 , 30 , 31 ], up to 40 days [ 14 , 16 ] but occurred after 2–6 months in community rehabilitation facilities [ 11 , 27 ].

Sample size

Sample size and composition varied considerably; from 16 participants [ 32 ] to 178 [ 16 ]. Six studies [ 11 , 12 , 15 , 16 , 29 , 30 ] had more than 100 participants; three studies [ 14 , 16 , 28 ] reported between 50 and 100; three studies [ 10 , 31 , 32 ] had less than 50 participants. The proportion of HI+ to HI− patients also varied; 7/12 studies had less than 50% HI+ in the sample, the smallest being 19% [ 10 ] and largest 60% [ 29 ] (the latter being therefore more adequately powered to statistically detect differences between the HI± groups).

Attrition rates

Attrition rates of 1%, 16%, 7.8% and 11% were reported [ 33 , 14 , 16 , 31 ], respectively. Reasons for attrition were due to a combination of factors (incomplete documentation at discharge, loss to community follow-up and mortality).

Assessment of Neglect/HI

In regard to HI, both diagnostic tools and frequency of assessment varied; most studies [ 11 , 12 , 14–16 , 27–30 ] assessed only at baseline (which differed in time across studies), whilst three studies [ 10 , 31 , 32 ] assessed patients at admission and discharge (which also differed in time since stroke). In half of the studies [ 10 , 14 , 28 , 30–32 ], HI was identified and assessed by a validated test battery – the Behaviour Inattention Test (BIT) [ 37 ]. Single letter cancellation and line-bisection (pen and paper tasks) were used in two studies [ 12 , 15 ], respectively, whilst three used various standardised Neglect-specific tests [ 11 , 16 , 27 ]. Study [ 29 ] relied on mention of Neglect in the medical documents.

Other assessments

Functional ability was assessed by the FIM in all but four studies [ 15 , 16 , 27 , 31 ] which used different versions of the BI. Most studies measured or recorded additional factors which ranged from length of stay (LOS), discharge destination outcome, continence status, aspects of cognitive-motor function including perception, muscle strength, balance and tactile sensation. Validated measurement scales were generally used for these purposes.

Statistical data analysis

Data tended to be summarised by group (HI±) scores. Median or mean statistics were frequently reported with standard deviation (SD), and inter-quartiles to a lesser extent. Data distribution was rarely described but inferred from the summary statistic used. Rasch data transformation was undertaken by two studies [ 29 , 30 ] in an attempt to “normalise” a skewed data distribution. Estimation and management of missing data were not specifically reported, with one exception [ 30 ].

Type of data analysis

The type and extent of data analysis varied substantially. For clarity, only general tendencies are described in this section; for specific details refer to Table 1 .

The majority of studies carried out preliminary tests for uni/bivariate associations (e.g. neglect × functional ability) and/or group (HI±) score comparison, e.g. [ 11 , 12 , 30 ]. The correlation coefficients used were Spearman’s rho or Pearson’s r , whereas t -test, Mann–Whitney U and Chi square test were frequently used for group comparisons. In order to minimise type I error (i.e. a false statistically significant result), one study [ 31 ] adjusted for multiple testing of the same participants over time by means of Bonferroni correction. Adjustment for small sample size was reported in two studies [ 15 , 32 ] but the adjustment method (Pillai’s trace) was only described once in study [ 32 ].

Eight studies [ 10–12 , 15 , 16 , 28–30 ] used regression methods to evaluate various relationships between predictor or explanatory variables (IV’s) with one or more dependent variables (DVs), including functional ability. However, the type of model used was not clearly identified (predictive versus associative model). Therefore, it was difficult to assess the suitability of the models employed for the purpose of answering the question posed. For example, Paolucci et al’s study [ 16 ] modelled the impact on later function of a large number and combination of IV’s (admission stroke severity score, gender, type of lesion, hypertension, diabetes, heart disease, unilateral spatial neglect, depression, epileptic seizures post-stroke, family support, education level, discharge destination in various combinations) but the extent of adjustment for confounding factors was variable – stroke severity was inconsistently adjusted for and no adjustment for differences in age was undertaken. Therefore, it was difficult to tease out specific relationships and infer cause from complex regression models. Differences in age, gender and duration of in-patient stay were adjusted in some models evaluated by studies [ 11 , 16 , 30 ]. Furthermore, the rationale behind the choice, order of entry and measurement level (continuous/categorical) of IV’s was rarely stated (e.g. study [ 10 ] used stepwise methods, whereas study [ 16 ] used forward stepwise), which complicated understanding and interpretation of the results.

Three out of four studies with more than one follow-up point evaluated change in functional ability over time by means of ANOVA’s and/or multiple regression analysis [ 10 , 14 , 27 ]; study [ 10 ] specified repeated measures ANOVA. However, both methods have considerable limitations which potentially impact on study power and accuracy of results, especially in serial models with more than one follow-up and several repeated measures on the same individual. To this end, ANOVA requires complete data sets, which is problematic in stroke research due to the likelihood of missing data in the long-term. Ordinary single or multivariate regression analysis does not take into account correlation generated by multiple responses from the same individual on the same assessment measure/s (this violates the statistical assumption of independent observations in regression analysis [ 38 , 39 ]). Consequently, both the validity and accuracy of ordinary regression results are threatened including any inferences based on the results.

Main results and findings

A substantial number of findings were reported across studies, only those pertaining to functional outcome and HI are summarised in this section (refer to Table 1 for details by study).

Disparity between the HI± group scores

All the studies found statistically significant disparities in average HI± group scores wherein the HI+ patients scored less that the HI− in overall functional ability and sensory-motor components on the BI, FIM and RMI (Rivermead Mobility Index) [ 40 ], at discharge and up to one year post-stroke onset. Relatively less (statistically non-significant) disparity was found on cognitive FIM sub-scale scores in three studies [ 14 , 29 , 30 ]. Nevertheless, the magnitude of differences reported across studies was considerably variable even when the same measurement scales were used at discharge. Study [ 15 ] reported a difference between the HI± groups of 2 BI units (10%); whilst other studies reported differences of 7 BI units (35%) [ 31 ], 10 FIM units (8%) [ 14 ] and >30 FIM units (24%) [ 30 ]. However, it must be pointed out that time to 1st observation and discharge point were also considerably variable across the studies (see later section), therefore it is difficult to extrapolate further from the findings to isolate specific influences of HI.

Progress rates

In general, similar rates of progress between the HI± groups were found prior to discharge but again these rates varied across studies even though the samples were homogenous with respect to lesion side (RHS). Four studies [ 10 , 14 , 27 , 31 ] followed up patients beyond discharge; study [ 27 ] found that specific HI training improved functional ability of the HI+ group but gains were not maintained by the end of the study (estimated from highly variable published data; recorded about 6–10 months after stroke).

Length of in-patient stay and discharge destination outcome

The HI+ patient group tended to have longer LOS/days but this varied considerably across studies, e.g. HI+ 64, HI− 36 days [ 15 ], HI+ 31, HI− 25 days [ 12 ] and HI+ 79, HI− 52 days [ 31 ]. On average, levels of community support and rates of institutional care were higher in the HI+ patients. However, entry of both the groups to institutional care was variable; ranging from 1/40 (2.5%) [ 10 ], 32/178 (18%) [ 16 ] and 6/28 (22%) [ 31 ].

Modelling results

In regard to multiple-regression modelling results, six out of eight studies reported that HI significantly and negatively predicted outcome in a variety of models of differing complexity and specification [ 10 , 16 ]. However, without a clear approach to modelling, such as identifying the strategy (e.g. step-up/down) used to build the models, the technique used to implement that strategy (forward/backward step-wise) and the decision criteria (e.g. theory driven) used within the technique, it was difficult to follow the modelling process and know the true effect of individual variables (inclusive of HI) as evident from the following examples.

Whilst four studies [ 10–12 , 16 ] reported independent prediction of HI for the measured functional outcomes, they did not clarify what “independent” means (e.g. explains more than 10% of the DV). In contrast, two studies [ 15 , 29 ] did not find a significant relationship between HI status and functional ability in any of the models evaluated – using either FIM or BI (0–20 scale) scores as the DV and HI as a categorical or continuous predictor variable, respectively. Notably the measurement of HI was itself a variable, since different studies used different measurement levels and scales.

In regard to covariates, pre-stroke function was unrelated to functional ability at admission (T0) according to three studies [ 15 , 16 , 30 ]. Baseline function was significantly, positively related to functional outcome at discharge according to three studies [ 12 , 28 , 29 ]. Age was significantly, negatively related to change in functional ability according to studies [ 28 , 29 ] but unrelated to final functional ability in models evaluated by studies [ 15 , 30 ]. Cognitive ability was unrelated to ADL as measured by BI and FIM in studies [ 15 , 30 ], respectively, but positively related in study [ 10 ].

Overall model fit

HI generally explained little of the final variance in functional ability (DV). Study [ 30 ] reported the largest amount of variance explained in the DV (discharge FIM scores) as 49%; 44% of which were explained by admission FIM scores and a further 5% by BIT scores (HI levels). A key limitation of these analyses was that none of the studies undertook basic sensitivity analysis, e.g. how well the models met regression assumptions, such as normal distribution of residuals. In addition, confidence intervals (CIs) around regression coefficients or standard error (SE) sizes were rarely reported which complicated the interpretation of regression coefficients.

Overall quality

Taking everything into account, individual studies were graded as C/D on the GRADE scale. This reflected (but was not limited to) serious limitations in the design and data analysis methods described in previous sections. In particular, the exclusion of RHS patients with higher levels of stroke severity and HI, effectively limited representativeness of the samples and generalisation to the wider clinical population. There was a lack of adjustment of potential confounding factors (such as stroke severity, age and time since stroke) in statistical models estimating specific effects of HI on outcomes; and insufficient data/details to enable the reader to make informed decisions (e.g. omission of confidence intervals around regression coefficient estimates which enable accurate interpretation of the result and lack of sensitivity analysis to support validity and conclusions made from the findings).

The purpose of the review was threefold – to test traditional claims that neglect/HI has a deleterious impact on functional ability after stroke, to assess the extent of differences between RHS patient groups (with and without HI) and clarify the relationship between HI and functional recovery. In addition, suggestions for more rigorous research were formulated based on a critique of these studies.

Based on findings from the 12 reviewed studies, it is apparent that the presence of HI+ is linked to poor functional outcomes in RHS patients when compared to their counterparts without HI impairments (HI−). However, it was not possible to assess the extent of differences between the HI± groups at specific points in time (elapsed after stroke) because of considerable variation and inconsistency in design within and across studies – not just in assessment measures but also in sample mix (age, stroke severity); time to 1st observation as well as follow-up assessments; and methods of data analysis. Neither was it possible to draw firm conclusions on the relationship between HI status and change in basic functional ability over time. In other words, the specific contribution of HI to the functional ability and outcome of RHS patients cannot be known with certainty from the results of the reviewed studies. It is possible that other influential factors, such as stroke severity, age and time since stroke substantially, explain the discrepancy in the HI± patient group scores reported in the literature. To this end, the implications of poor methodology in the field of stroke and HI were also highlighted in previous reports [ 2 , 21 , 22 ]. There is also recent discussion [ 38 ] of the importance of optimising data analysis methods in the field of neglect/HI to enhance the accuracy of predictive models.

Aside from clear methodological limitations, such as exclusion of patients with severe stroke and marked HI+ at baseline, it is likely that unmeasured patient factors also influenced the results and are responsible for a proportion of unexplained variance in functional ability post-stroke. These include pre-stroke educational levels and personality characteristics which would be expected to vary across patients. Furthermore, seven studies [ 12 , 15 , 16 , 28–30 , 32 ] used discharge as the only follow-up point. Although discharge is a key point of change in the stroke recovery pathway, it is subject to natural variation in stroke recovery rate and initial severity. Furthermore, the optimal discharge point may be influenced by differing cultural expectations as to when the recovering patient is ready to be discharged to community care and where this will be undertaken, i.e. home or institution. From the description available in past reviewed studies, it would appear that contextual features, such as stroke unit versus acute general rehabilitation hospital, were not sufficiently comparable in care and rehabilitation provision across different countries especially at the time that they were undertaken [ 10 , 12 , 14–16 ]. This lack of comparability in context of care is in itself a source of important variation when evaluating functional outcomes from different countries. In relation to the issues associated with discharge as an informative follow-up point, these can be minimised by including at least one standardised fixed follow-up point for all the patients, e.g. baseline – discharge – follow-up 6 months after stroke.

Sample representation and size

Although all patients in the studies selected had RHS, there were considerable sample differences in age and recruitment settings which were not directly comparable. Furthermore, time to 1st observation and patient selection criteria were substantially different; and patients with severe cognitive impairment, and probably severe neglect/HI+ levels, tended to be automatically excluded at the recruitment phase. This strongly suggests that the samples were not sufficiently representative of the RHS population which limits the generalisation of any findings. It is imperative that future studies are conducted with representative (severity-inclusive) samples especially in predictive/explanatory relationship studies. This can be achieved by augmenting the design to suit specific requirements of severely cognitively impaired patients and then applying advanced methods of data analysis (e.g. multi-level modelling – MLM). Specific advantages of MLM over traditional methods of regression analysis and ANOVA are highlighted later on.

The importance of a large enough sample size in relation to study power and reducing type I error cannot be overemphasised. For instance, two studies [ 10 , 27 ] followed up patients for longer periods (up to one year after stroke) but with relatively small starting samples (40 and 59, respectively). Although attrition rate was not reported, this can be substantial at one year post-stroke which further reduces study power to detect differences between the HI± groups. Therefore, future studies should ensure sufficient initial sample size, particularly in longer term follow-up designs (lasting more than 3 months after stroke) to account for possible attrition; an allowance of 15–20% is cited in the literature [ 41 , 42 ].

Measurement

The accuracy of results depended heavily on the precision of instruments and consistency in measurement. Only 50% of studies followed evidence-based recommendations in favour of diagnosing neglect/HI with validated test batteries rather than single cancellation task/s; the rationale [ 35 , 43 , 44 ] being that several tests are more likely to detect neglect than a single test. To this end, the proportion of the HI+ patients detected in studies that assessed with the BIT is considerably higher than in those that used single tests. Like other test batteries, the BIT has measurement limitations in that it is restricted to assessment of HI in activities conducted within close proximity (arm’s length) of surrounding body space and possibly mental representation of objects. In addition, similar to other test batteries (e.g. Chaterine Bergago scale [ 45 ]), the BIT cannot distinguish between different sub-types of the disorder [ 3 , 36 ] because all the assessed tasks inherently draw on visual, sensory, motor, spatial, perceptual and cognitive function to varying degrees. These components are intricately inter-twined and dependent on one another which make it virtually impossible to determine the relative contribution of individual components to the results or behaviour observed. Consequently (and of relevance to clinicians and researchers), it is not possible to accurately explain test results in terms of specific sub-types of HI in routine rehabilitation settings. That being said, assessment batteries provide an overall score of HI severity and are indicative of the type of perceptual, cognitive and/or motor impairments present.

Pending the development of more comprehensive and practical assessment tools, future researchers and clinicians should follow evidence-based assessment guidelines, i.e. use of validated assessment batteries for HI. Interpretation of results will also be enhanced by reporting the proportion of patients with mild compared to severe neglect/HI in the sample by categorising the HI+ patients at baseline. This is possible with the BIT [ 36 , 46 ].

Functional ability in the reviewed studies was assessed with at least two versions of the BI (using scales of 0–20 and 0–100) with differing sensitivity and precision [ 47 , 48 ], making comparison difficult. Furthermore, neither version assesses social/communicative components. They are therefore possibly more limited than the FIM in this respect [ 46 ]. In comparison, the Extended Barthel Index version has been used in several stroke studies [ 49 , 50 ]. It is a validated measure and overcomes some of the limitations of other BI versions in that it includes assessment of social and communicative components and accounts for time taken to complete ADL tasks.

Ideally future studies should measure functional ability with comprehensive evidence-based measures; however it must be acknowledged that current choices are limited not only by content in terms of validity but also reliability over time and applicability across acute and community settings.

The position, interval and frequency of observations

In relation to design, the position, interval and frequency of observations is of critical importance in recording important changes in ability at the points when they are likely to occur. Furthermore, the amount and quality of information collected partly determines the type and extent of data analysis that is possible in order to answer specific research questions [ 39 , 51 , 52 ]. This ability was greatly compromised within and across the reviewed studies by the variability in time around the initial measurement point and follow-up, the number and frequency of observations made over time, and the duration of individual studies.

In some studies, the extent of variation around the mean LOS blurred the beginning and end of the study (e.g. study [ 16 ] reported mean LOS (HI+) 117 ± 61 days versus (HI−) 81 ± 38, and study [ 29 ] reported (HI+) 29 versus (HI−) 22, range 3–75 days). This would suggest considerable differences in LOS and patient exposure to care within and across studies which have implications for assessing recovery.

At baseline, the total variation in initial measurement points across studies was 7 days to 6 months post-stroke in studies [ 27 , 32 ], respectively. During this period, both spontaneous and rehabilitation-driven recovery processes are known to actively contribute to outcome [ 53–55 ]. Based on known recovery patterns, the impact of spontaneous recovery processes is expected to confound rehabilitative processes in at least 11/12 studies. Future studies should aim to collect initial data as early as is pragmatically possible, ideally within the 1st week of stroke. This would potentially streamline data collection procedures, enhance comparison of results across studies and minimise the effect of spontaneous recovery processes on outcome so that cause (e.g. influence of HI) can be more confidently inferred from regression results.

Future studies could also consider fixing the last follow-up point whilst still recording HI and functional change at discharge. Better still would be to statistically adjust for the confounding effect of time elapsed post-stroke. An optional growth curve modelling (MLM) approach could be used to model change over time in time-variant factors – this approach has been applied in a predictive stroke patient study [ 56 ]. The advantages afforded by MLM in relation to analysis of stroke data are next summarised in this review.

Multi-level modelling approaches to data analysis

The advantages of multi-level modelling over traditional regression methods of analysis should not be overlooked when modelling dependency in the data due to multiple measurements from the same patient over time and interdependency between potential explanatory factors associated with functional ability (e.g. cognitive-motor processes). MLM can also handle data missing at random and unequal interval observations – all of which were potentially problematic features of the reviewed studies. As a result, MLM regression estimates are highly accurate and relatively stable compared to estimates obtained by ordinary multivariate regression analysis [ 51 , 52 ]. MLM has other useful features, such as estimation of unexplained variance within and across patients over time (in serial studies). It is imperative that future studies consider the advantages afforded by novel advanced statistical techniques in serial designs, which provide rich information but need to be appropriately statistically analysed for optimal results [ 38 , 57 , 58 ].

From the discussion so far it can be deduced that regression estimates in 8/12 studies which modelled outcome are likely to be overinflated with under-estimated standard errors. In turn, this increases the risk of type 1 error which has adverse implications for interpretation of results, not least the predictive importance of HI on functional ability.

Confounding factors in stroke functional outcome studies

Another issue complicating the inference of causality from regression results is the lack of statistical adjustment for the established confounding effect of stroke severity and to a lesser extent chronological age. Besides the natural variation of both factors in different samples, stroke severity and age are both associated with HI [ 21 , 59 , 60 ]. Therefore, it is important that these particular variables are modelled in studies exploring multiple predictor variables and functional outcome [ 39 , 61 ]. As already stated, the use of MLM can greatly help with teasing out complicated relationship dynamics. By the same token, the need for an adequately powered study is emphasised. Considering that only six studies started with approximately 100 patients and some cases [ 10 , 27 ] also analysed predictive multivariate models, it is likely that respective studies were underpowered to detect true effects. This increases the uncertainty of the modelled results and detracts from the validity of their findings.

Strengths and limitations of the review

The current review extended the work carried out by Jehkonen et al. by reviewing a homogeneous (versus mixed) patient sample with respect to hemispheric lesion (RHS). Further, a relatively more rigorous and systematic approach to study selection and the overall review process was undertaken.

The current review focused on group comparisons of functional outcomes of RHS patients with and without HI and the relationship of HI status with functional change with time since stroke, whereas Jehkonen et al. undertook a generalised review of the methodological issues from a wider range of studies which did not always include patient comparisons in the design (HI±). Group comparisons allow for the calculation of mean differences between patient groups and estimation of the modelled relationship between HI (group) status and functional outcome. This resulted in new insights into the data, such as the lack of adjustment for established confounding factors in past studies (e.g. stroke severity, time since stroke and age), leading to a different conclusion, i.e. no inferences could be made from the data available on either the relationship between neglect and functional outcome or the magnitude of difference in measureable scores between the patients with and without HI due to considerable heterogeneity in design and methods used to statistically analyse the data. Currently there is an urgent need for valid predictors and indicators to support the transfer of in-patient stroke rehabilitation services to appropriate stroke survivors living in the community.

Furthermore, due to the structure and layout of the review by Jehkonen et al., it is not possible to tell which studies found what and when in terms of results and time since stroke. The current review is more detailed in this respect. It is also more systematic and transparent both in the selection of studies and their evaluation (a checklist was used to ensure parity and consistency during the review process). The methodological quality of each study was separately graded. All these factors contribute to the rigour of the review and the validity of the findings.

The main limitation of the review is the possibility that relevant studies may have been missed due to indexing and multiple terms used to describe the “neglect” syndrome although a thorough search of the literature has been conducted several times. In addition, only English language publications or translations were considered.

The review does not address HI conditions arising from specific cortical or sub-cortical structures or anterior/posterior circulation division or left hemisphere lesions. Although HI is associated with different brain areas, HI emanating from the right hemisphere is commonly encountered in rehabilitation settings and challenging to treat [ 4–7 , 9 ]. In addition, the interpretation of results from assessments of HI in the left hemisphere lesions is confounded by speech and language impairments which frequently accompany LHS conditions. Non-language based clinical assessment for HI is urgently needed for this purpose. There is a lack of comparative group studies specifically on HI/neglect emanating from sub-cortical structures [ 63 , 64 ] probably because this is harder to identify in routine clinical settings and possibly less commonly encountered.

Conclusions

To conclude, based on findings from this review, the evidence for an important relationship between HI status and functional outcome in RHS patients (with or without consideration of other influential factors) cannot be substantiated from the data available. This is due to significant methodological limitations of published studies highlighted in this review and the relatively few studies published in this field in the last decade [ 29–32 ]. The evidence that differences exist between the HI± patient groups on discharge and in the early months that follow tends to be better substantiated and is of interest to rehabilitation practitioners. However, on its own, this knowledge does not advance practice in the field.

In light of findings from this review, it is recommended that rehabilitation practitioners not only take into account the severity of HI in decision making but also recognise that evidence is weak for it being an independent marker of future functional recovery. Given the extent of differential progress possible in patients with RHS with and without HI, it is recommended that rehabilitation practitioners consider both the rate and amount of functional recovery over time. This is likely to be more informative overall than isolated measurements of HI soon after stroke. In relation to measurement, it is recommended that rehabilitation professionals sufficiently consider the psychometric limitations of commonly employed assessments of HI and overall functional ability (e.g. the BIT, BI and FIM) when interpreting the results.

Further research is warranted to identify the magnitude of the differences within and between the patient HI± groups and importantly the precise relationship between HI status and functional recovery with time since stroke. For this reason, well-designed, longitudinal, serial studies on large RHS patient samples are required which can withstand the scrutiny of future reviews on the subject.

It is recommended that future rehabilitation research in the field takes into account important methodological features to improve the quality and robustness of their findings. These include but are not limited to adequate sample size to detect differences in groups of patients with and without HI, with a full spectrum of stroke severity and HI, sufficient amounts of data by means of standardised methods repeated over time rather than isolated data points, and with sufficient follow-up to allow functional recovery to occur over time since stroke.

Declaration of interest

This work received no external funding. The authors report no conflicts of interest.

Appendix 1.

Abbreviations – CT = computer tomography, MRI = magnetic resonance imaging.

Content was adapted from the Critical Appraisal Skills Programme [ 62 ].

Appendix 2.

• Stone SP, Patel P, Greenwood RJ. The incidence of neglect phenomena and related disorders in patients with an acute right or left hemisphere stroke. Age Ageing. 1993; 22 :46–52. [ PubMed ] [ Google Scholar ]
• Bowen A, McKenna K, Tallis CT. Reasons for variability in the reported rate of occurrence of unilateral spatial neglect after stroke. Stroke. 1999; 30 :1196–202. [ PubMed ] [ Google Scholar ]
• Plummer P, Morris ME, Dunai J. Assessment of unilateral neglect. Phys Ther. 2003; 83 :732–40. [ PubMed ] [ Google Scholar ]
• Singh-Curry V, Husain M. Hemispatial neglect: approaches to rehabilitation. Clin Rehabil. 2010; 24 :675–84. [ PubMed ] [ Google Scholar ]
• Karnath H, Rorden C. The anatomy of spatial neglect. Neuropsychologia. 2012; 50 :1010–17. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Mark VW. Acute versus chronic functional aspects of unilateral spatial neglect. Front Biosci. 2003; 8 :172–89. [ PubMed ] [ Google Scholar ]
• Parton A, Malhotra P, Husain M. Hemispatial neglect. J Neurol Neurosur Ps. 2004; 75 :13–21. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Appelros P, Karlssonb GM, Hennerdal S. Anosognosia versus unilateral neglect. Coexistence and their relations to age, stroke severity, lesion site and cognition. Eur J Neurol. 2007; 14 :54–9. [ PubMed ] [ Google Scholar ]
• Kortte K, Hillis AE. Recent advances in the understanding of neglect and anosognosia following right hemisphere stroke. Curr Neurol Neurosci Rep. 2009; 9 :459–65. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Katz N, Hartman-Maeir A, Ring H, et al. Functional disability and rehabilitation outcome in right hemisphere damaged patients with and without unilateral spatial neglect. Arch Phys Med Rehabil. 1999; 80 :379–84. [ PubMed ] [ Google Scholar ]
• Buxbaum LJ, Ferraro MK, Veramonti T, et al. Hemispatial neglect: subtypes, neuroanatomy, and disability. Neurology. 2004; 62 :749–56. [ PubMed ] [ Google Scholar ]
• Gillen R, Howard Tennen H, Tara McKee T. Unilateral spatial neglect: relation to rehabilitation outcomes in patients with right hemisphere stroke. Arch Phys Med Rehabil. 2005; 86 :763–7. [ PubMed ] [ Google Scholar ]
• Pedersen PM, Jørgensen HS, Hirofumi N, et al. Hemineglect in acute stroke – incidence and prognostic implications: the Copenhagen Stroke Study1. Am J Phy Med Rehab. 1997; 76 :122–7. [ PubMed ] [ Google Scholar ]
• Cherney L, Halper AS, Kwasnica CM, et al. Recovery of functional status after right hemisphere stroke: relationship with unilateral neglect. Arch Phys Med Rehabil. 2001; 82 :322–8. [ PubMed ] [ Google Scholar ]
• Kalra L, Perez I, Gupta S, et al. The Influence of visual neglect on stroke rehabilitation. Stroke. 1997; 28 :1386–91. [ PubMed ] [ Google Scholar ]
• Paolucci S, Antonucci G, Grasso MG, et al. The role of unilateral spatial neglect in rehabilitation of right brain–damaged ischemic stroke patients: a matched comparison. Arch Phys Med Rehabil. 2001; 82 :743–9. [ PubMed ] [ Google Scholar ]
• Nijboer T, Van de Port I, Schepers V, et al. Predicting functional outcome after stroke: the influence of neglect on basic activities in daily living. Front Hum Neurosci. 2013; 7 :1–6. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Plummer P, Morris ME, Hurworth RE, et al. Characterisation of unilateral neglect by physiotherapists. Disabil Rehabil. 2006; 28 :571–7. [ PubMed ] [ Google Scholar ]
• Barrett AM, Buxbaum LJ, Coslett HB, et al. Cognitive rehabilitation interventions for neglect and related disorders: moving from bench to bedside in stroke patients. J Cogn Neurosci. 2006; 18 :1223–36. [ PubMed ] [ Google Scholar ]
• Menon-Nair A, Korner-Bitensky N, Wood-Dauphinee S. Assessment of unilateral spatial neglect post stroke in Canadian acute care hospitals: are we neglecting neglect? Clin Rehabil. 2006; 20 :623–34. [ PubMed ] [ Google Scholar ]
• Jehkonen M, Laihosalo M, Kettunen JE. Impact of neglect on functional outcome after stroke – a review of methodological issues and recent research findings. Restor Neurol Neurosci. 2006; 24 :209–15. [ PubMed ] [ Google Scholar ]
• Lincoln N, Bowen A. The need for randomised treatment studies in neglect research. Restor Neurol Neuros. 2006; 24 :401–8. [ PubMed ] [ Google Scholar ]
• Dijkers M. Task force on systematic reviews and guidelines the value of traditional reviews in the era of systematic reviewing. Am J Phys Med Rehabil. 2009; 88 :423–30. [ PubMed ] [ Google Scholar ]
• Pai M, McCulloch M, Gorman J, et al. Systematic reviews and meta-analyses: an illustrated, step-by-step guide. Natl Med J India. 2004; 17 :86–95. [ PubMed ] [ Google Scholar ]
• Nicolson A, Moir L, Millsteed J. Impact of assistive technology on family caregivers of children with physical disabilities: a systematic review. Disabil Rehabil Assist Technol. 2012; 7 :345–9. [ PubMed ] [ Google Scholar ]
• Guyatt G, Oxman AD, Vist GE. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008; 336 :924–6. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Paolucci S, Antonucci G, Guariglia C, et al. Facilitatory effect of neglect rehabilitation on the recovery of left hemiplegic stroke patients: a cross-over study. J Neurol. 1996; 243 :308–14. [ PubMed ] [ Google Scholar ]
• Ring H, Feder M, Schwartz J, et al. Functional measures of first-stroke rehabilitation inpatients: usefulness of the Functional Independence Measure total score with a clinical rationale. Arch Phys Med Rehabil. 1997; 78 :630–5. [ PubMed ] [ Google Scholar ]
• Odell K, Wollack J, Flynn M. Functional outcomes in patients with right hemisphere brain damage. Aphasiology. 2005; 19 :807–30. [ Google Scholar ]
• Di Monaco M, Schintu S, Dotta M, et al. Severity of unilateral spatial neglect Is an independent predictor of functional outcome after acute inpatient rehabilitation in individuals with right hemispheric stroke. Arch Phys Med Rehabil. 2011; 92 :1250–6. [ PubMed ] [ Google Scholar ]
• Stein MS, Maskill D, Marston L. Impact of visual-spatial neglect of stroke functional outcomes, discharge destination and maintenance of improvement post-discharge. BJOT. 2009; 72 :219–25. [ Google Scholar ]
• Timbeck R, Spaulding SJ, Klinger L, et al. The effect of visuospatial neglect on functional outcome and discharge destination: an exploratory study. Phys Occup Ther Geriatr. 2013; 31 :37–46. [ Google Scholar ]
• Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Maryland State Med J. 1965; 14 :61–5. [ PubMed ] [ Google Scholar ]
• Granger CV, Cotter AC, Hamilton BB, et al. Functional Assessment Scales: a study of persons after stroke. Arch Phys Med Rehabil. 1993; 74 :133–8. [ PubMed ] [ Google Scholar ]
• Heilman KM, Valenstein E, Watson RT. The neglect syndrome. In: Vinken PJ, Bruyn GW, Klawans HL, editors. Handbook of clinical neurology 45. Amsterdam: Elsevier Science Publishers; 1985. pp. 152–83. [ Google Scholar ]
• Menon A, Korner-Bitensky N. Evaluating unilateral spatial neglect post stroke: working your way through the Maze of Assessment Choices. Top Stroke Rehabil. 2004; 11 :41–66. [ PubMed ] [ Google Scholar ]
• Wilson B, Cockburn J, Halligan PW. Behavioural inattention test, manual. Titchfield, Hampshire: Thames Valley Test Company; 1987. [ Google Scholar ]
• Goedert K, Boston R, Barrett A. Advancing the science of spatial neglect rehabilitation: an improved statistical approach with mixed linear modelling. Front Hum Neurosci. 2013; 7 :1–15. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Shmueli G. To explain or to predict? Stat Sci. 2010; 25 :289–310. [ Google Scholar ]
• Collen C, Wade DT, Robb DT, et al. The Rivermead Mobility Index: a further development of the Rivermead Mobility Assessment. Int Disabil Stud. 1991; 13 :50–4. [ PubMed ] [ Google Scholar ]
• Snijders AB, Bosker RJ. Standard errors and sample sizes for two-level research. J Educ Stat. 1993; 18 :237–59. [ Google Scholar ]
• Walker J, Almond P. Interpreting statistical findings: a guide for health professionals and students. Berkshire: Open University Press; 2010. [ Google Scholar ]
• Halligan P, Marshall JC, Wade DT. Visuospatial neglect: underlying factors and test sensitivity. Lancet. 1989; ii :908–11. [ PubMed ] [ Google Scholar ]
• Lopes M, Ferreira HP, Carvalho JC, et al. Screening tests are not enough to detect hemineglect. Arq Neuropsiquiatr. 2007; 65 :1192–5. [ PubMed ] [ Google Scholar ]
• Azouvi P, Samuel C, Louis-Dreyfus A, et al. Sensitivity of clinical and behavioural tests of spatial neglect after right hemisphere stroke. J Neurol Neurosurg Psy. 2002; 73 :160–6. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF Activity. Disabil Rehabil. 2005; 27 :315–40. [ PubMed ] [ Google Scholar ]
• McDowell I. Measuring health: a guide to rating scales and questionnaires. 3rd. New York: Oxford University Press; 2006. [ Google Scholar ]
• Quinn TJ, Langhorne P, Stott DJ. Barthel index for stroke trials: development, properties, and application. Stroke. 2011; 42 :1146–51. [ PubMed ] [ Google Scholar ]
• Geshwinder H, Rettkeb H, Van den Heuvelc W, et al. Rehabilitation in acute stroke patients in German-speaking Switzerland. Swiss Med Wkly. 2007; 137 :205–11. [ PubMed ] [ Google Scholar ]
• Schuster C, Butler J, Andrews D, et al. Comparison of embedded and added motor imagery training in patients after stroke: results of a randomised controlled pilot trial. Trials. 2012; 13 :11 (1–19). [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Cheng J, Edwards LJ, Maldonado-Molina MM, et al. Real longitudinal data analysis for real people: building a good enough mixed model. Statist Med. 2010; 29 :504–20. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Snijders AB, Bosker RJ. Multilevel analysis: an introduction to basic and advanced multilevel modeling. 2nd. London: Sage Publishers; 2012. [ Google Scholar ]
• Grefkes C, Fink GR. Reorganization of cerebral networks after stroke: new insights from neuroimaging with connectivity approaches. Brain. 2011; 134 :1264–76. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kwakkel G, Kollen BJ. Predicting activities after stroke: what is clinically relevant? Int J Stroke. 2013; 8 :25–32. [ PubMed ] [ Google Scholar ]
• Rehme AK, Grefkes C. Cerebral network disorders after stroke: evidence from imaging-based connectivity analyses of active and resting brain states in humans. J Physiol. 2013; 591 :17–31. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Tilling K, Sterne J, Wolfe C. Multilevel growth curve models with covariate effects: application to recovery after stroke. Statist Med. 2001; 20 :685–704. [ PubMed ] [ Google Scholar ]
• Royston P, Moons K, Altman DG, et al. Prognosis and prognostic research: developing a prognostic model. BMJ. 2009; 338 :1373–7. [ PubMed ] [ Google Scholar ]
• Kwakkel G, Kollen B, Twisk J. Impact of time on improvement of outcome after stroke. Stroke. 2006; 37 :2348–53. [ PubMed ] [ Google Scholar ]
• Appelros P, Gunnel M, Karlsson A, et al. Prognosis for patients with neglect and anosognosia special reference to cognitive impairment. J Rehabil Med. 2003; 35 :254–8. [ PubMed ] [ Google Scholar ]
• Appelros P, Karlsson GM, Hennerdal S. Anosognosia versus unilateral neglect. Coexistence and their relations to age, stroke severity, lesion site and cognition. Eur J Neurol. 2007; 14 :54–9. [ PubMed ] [ Google Scholar ]
• Twisk JW. Applied multilevel analysis: a practical guide. New York: Cambridge University Press; 2006. [ Google Scholar ]
• Critical Appraisal Skills Programme. Available from: http://www.casp-uk.net/# !casp-tools-checklists/c18f8
• Punt TD, Riddoch MJ. Motor neglect: implications for movement and rehabilitation following stroke. Disabil Rehabil. 2006; 28 :857–64. [ PubMed ] [ Google Scholar ]
• Sampanis DS, Riddoch J. Motor neglect and future directions for research. Front Hum Neurosci. 2013; 7 :1–2. [ PMC free article ] [ PubMed ] [ Google Scholar ]

• Become a member

Unilateral Spatial Neglect Recovery Poststroke

Information & authors, metrics & citations, view options, graphical abstract.

Time Course

Mechanisms of recovery, acute recovery, subacute and chronic recovery, predictors of recovery, augmenting recovery: state of the evidence on usn intervention, nonpharmacological interventions, behavioral cueing, prism adaptation, noninvasive brain stimulation, virtual reality, robot-assisted intervention, pharmacological interventions, article information, supplemental material, eletters (0).

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.

Information

Published in.

Permissions

• hemispatial neglect
• randomized controlled trial
• recovery of function
• rehabilitation
• Cognitive Impairment

Affiliations

Disclosures, sources of funding.

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.

• Alix Gouret,
• Solène Le Bars,
• Thibault Porssut,
• Florian Waszak,
• Sylvie Chokron,
• Masaki Kondo,
• Kamal Narayan Arya,
• Shanta Pandian,
• Divya Pandey,
• G. G. Agarwal,
• Neera Chaudhary,
• Jean-Baptiste Brunet de Courssou,
• Christian Denier,
• Margot Juliëtte Overman,
• Elena Binns,
• Elise T. Milosevich,
• Nele Demeyere,
• David R Painter,
• Michael F Norwood,
• Chelsea H Marsh,
• Trevor Hine,
• Christie Woodman,
• Marilia Libera,
• Daniel Harvie,
• Kelly Dungey,
• Julie Bernhardt,
• Leslie Gan,
• Susan Jones,
• Heidi Zeeman,
• Abhishek Jaywant,
• Alexandra Keenan,
• Grigorios Nasios,
• Lambros Messinis,
• Efthimios Dardiotis,
• Markos Sgantzos,
• Alex R. Carter,
• A.M. Barrett,
• Bara’ah Bsharat,
• Ahmad Al-Duhoun,
• Parisa Ghanouni,

View options

Pdf and all supplements, login options.

Check if you have access through your login credentials or your institution to get full access on this article.

Purchase Options

Purchase this article to access the full text.

Purchase access to this article for 24 hours

Purchase access to this journal for 24 hours

Restore your content access

Enter your email address to restore your content access:

Note: This functionality works only for purchases done as a guest. If you already have an account, log in to access the content to which you are entitled.

Share article link

Copying failed.

Submit a Response to This Article

Compose eletter, contributors, statement of competing interests, previous article, next article, comment response.

• Type 2 Diabetes
• Heart Disease
• Digestive Health
• Multiple Sclerosis
• Diet & Nutrition
• Health Insurance
• Public Health
• Patient Rights
• Caregivers & Loved Ones
• End of Life Concerns
• Health News
• Thyroid Test Analyzer
• Doctor Discussion Guides
• Hemoglobin A1c Test Analyzer
• Lipid Test Analyzer
• Complete Blood Count (CBC) Analyzer
• What to Buy
• Editorial Process
• Meet Our Medical Expert Board

Right-Sided Strokes: What to Expect

They can cause specific problems that you should know about

Frequently Asked Questions

Strokes are defined as right-sided or left-sided based on which hemisphere (side) of the brain is affected. Since different regions of the brain control specific functions, the effects of a stroke correlate to the damaged area of the brain.

A right-sided stroke can cause many symptoms . The most noticeable are those that affect the left side of the body, which is controlled by the right side of the brain.

Verywell / Ellen Lindner

Types of Right Hemisphere Strokes 

Any stroke, including a right-sided stroke, can occur due to either a blood clot, bleeding, or both. 

An ischemic stroke is caused by a decrease in blood flow to an area of the brain. Even a few minutes of inadequate blood flow can cause damage to the brain tissue. 

An ischemic stroke can be caused by a blood clot in a large blood vessel or a small blood vessel. Usually, blockage of blood flow in a small vessel causes less damage than blockage of blood flow in a larger vessel. 

Often, small vessel strokes occur due to atherosclerosis of an artery in the brain. Atherosclerosis is a combination of blood vessel damage and a buildup of material that can eventually lead to complete blockage of blood flow.

Sometimes strokes are caused by a blood clot that traveled from the heart or the carotid artery to the brain . This is more common with large vessel strokes . 

Hemorrhagic 

A hemorrhagic stroke occurs when a blood vessel leaks into the brain. Blood causes harmful irritation to the brain tissue, and the bleeding also deprives the nearby area of the brain of adequate blood supply. 

A right-sided stroke can occur suddenly, and it can cause: 

• Sudden weakness of the face, arm, or leg
• Severe dizziness, balance problems, and difficulty walking 
• Head pain, especially from a hemorrhagic stroke 

Sometimes a stroke can evolve rapidly, and the symptoms can seem confusing and overwhelming. If you or someone else experiences any of these signs, get medical attention immediately. 

The specific effects that you experience from a right-sided stroke can become more obvious to you as you become more medically stable in the days after the initial event. Effects can persist for years, and sometimes the effects can improve over time.

A right-sided stroke causes immediate and lasting effects that differ from those of a left-sided stroke. 

Hemiplegia on the Left Side

Hemiplegia is paralysis (complete loss of movement) on one side of the body. A right-sided stroke can cause hemiplegia of the whole left side of the body.

More commonly, this type of stroke causes left-side hemiparesis , which is diminished strength, without total paralysis. It usually affects only the face, arm, or leg—not necessarily the whole left side.

Sometimes, months or years after the stroke, spasticity (muscle stiffness or rigidity) can develop in the weak muscles. This occurs when a stroke affects the right motor strip of the cerebral cortex (which helps control movement) or the right internal capsule (nerve fibers from the motor strip run through this area).

Diminished Sensation on the Left Side 

After a right-sided stroke, it is possible to have diminished sensation or loss of sensation on the left side of the body. Sometimes paresthesias (numbness, tingling, or other unusual sensations) or pain can develop in the areas of the body that have diminished sensation. This usually begins after weeks, months, or longer. 

Sensory disturbances on the left side of the body can occur due to a stroke in the right sensory strip of the cerebral cortex or the right thalamus.  

Prosopagnosia

One of the rare effects of a right-sided stroke is prosopagnosia , which is an inability to recognize faces. This can occur due to a stroke affecting the right fusiform gyrus, an area near the back of the brain that works to help identify faces.

Left Neglect

One of the distressing characteristics of a right-sided stroke is deceased attention to the left side of the body or an inability to recognize the area of the body impacted by the stroke. As with other effects of a right-sided stroke, the severity of this problem can range from mild to severe.  

Neglect can occur when a stroke affects the right parietal lobe (a back part of the brain). 

Challenges of Neglect

Neglect after a right-sided stroke can make it especially difficult to participate in physical therapy and other aspects of rehabilitation.

Homonymous Hemianopia

A right-sided stroke can cause loss of vision on the left side from both eyes. This can affect the whole left side, or only the upper or lower part of vision on the left side. When it affects the whole left side it is called left homonymous hemianopia . If it involves just the upper or lower part of vision it is a quadrantinopia.

A stroke affecting the right occipital lobe, which is the farthest back region of the brain, can cause left homonymous hemianopia. 

Anosognosia

This complex effect is the inability of a person to recognize that they have a disability from a stroke. It is similar to neglect, but there are some subtle distinctions because a person who is experiencing anosognosia may recognize the impaired area of the body, but can be unable to recognize the impairment. 

Anosognosia can occur due to damage in the right parietal, temporal, or frontal lobe of the brain. 

Pseudobulbar Affect 

This condition can occur due to a number of different neurological conditions, including a right-sided stroke. The symptoms of pseudobulbar affect include episodes of uncontrollable emotional outbursts, such as laughing or crying. They may be inappropriate, as the emotions come out at random times and don't always make sense.

It can be embarrassing for some people who may be distressed by their own lack of emotional control. People who have had a very large stroke might not notice the effects or might not be distressed about it. 

There are several treatments for a stroke . When the symptoms first begin, treatment can include blood pressure control, fluid management, and sometimes blood thinners. These interventions can reduce the damage of a stroke and improve survival. 

After the acute stage of a right-sided stroke, treatment involves rehabilitation. This can include physical therapy, speech and swallow therapy, cognitive therapy, and occupational therapy to help maximize movement and self-care. 

Prevention 

After a stroke, prevention of further strokes is important. Diagnostic testing involves tests that assess stroke risk factors . Prevention is focused on managing risk factors to reduce the chances of another stroke. 

Prevention includes:

• Maintaining optimal blood pressure 
• Diet modification and medical treatment to achieve healthy cholesterol and triglyceride levels 
• Control of diabetes 
• Blood thinners if there is a high risk of blood clots
• Treatment of heart problems, such as valve disease, coronary artery disease, and irregular heart rhythms 
• Smoking cessation

Prevention involves consistent surveillance of risk factors and assessment of risk factor control. 

A stroke can have many different effects, depending on which side of the brain is affected. A right-sided stroke can cause left-sided weakness, left-sided sensory loss, loss of vision from the left side of both eyes, personality changes, neglect of the left side of the body, and lack of recognition of the stroke.

The risk of having a stroke can be reduced if risk factors are identified and managed. Often, a stroke can be treated, but there can be residual effects. The larger a stroke, the more substantial the effects. Stroke rehabilitation is an important part of recovery.

The difference is that a right-sided stroke affects the right side of the brain, while a left-sided stroke affects the left side of the brain. They each can cause weakness and diminished sensation on the opposite side of the body. A right-sided stroke also can cause a lack of awareness of the weak side of the body, and this can make rehabilitation more difficult. 

It depends on many factors. It can take longer to recover from a large stroke, especially if you have had other strokes before or if you have health problems, such as severe heart or lung disease. 

This type of stroke can be caused by blockage of blood flow or from a bleeding blood vessel. Risk factors include high blood pressure, heart disease, smoking, uncontrolled diabetes, and high cholesterol. 

Ishii D, Ishibashi K, Yuine H, Takeda K, Yamamoto S, Kaku Y, Yozu A, Kohno Y. Contralateral and ipsilateral interactions in the somatosensory pathway in healthy humans . Front Syst Neurosci . 2021;15:698758. doi:10.3389/fnsys.2021.698758

National Institute of Health. Prosopagnosia .

Gillespie DC, Cadden AP, Lees R, West RM, Broomfield NM. Prevalence of pseudobulbar affect following stroke: A systematic review and meta-analysis. J Stroke Cerebrovasc Dis . 2016;25(3):688-94. doi:10.1016/j.jstrokecerebrovasdis.2015.11.038

By Heidi Moawad, MD Dr. Moawad is a neurologist and expert in brain health. She regularly writes and edits health content for medical books and publications.

LOGIN TO YOUR ACCOUNT

• Email* Password* Forgot password? Reset it here Keep me logged in Fields with * are mandatory Don't have an account? Create one here

Change Password

Your password must have 8 characters or more and contain 3 of the following:.

• a lower case character, 
• an upper case character, 
• a special character 

Password Changed Successfully

Your password has been changed

Create a new account

Can't sign in? Forgot your password?

Enter your email address below and we will send you the reset instructions

If the address matches an existing account you will receive an email with instructions to reset your password.

Request Username

Can't sign in? Forgot your username?

Enter your email address below and we will send you your username

If the address matches an existing account you will receive an email with instructions to retrieve your username

• NEWLY PUBLISHED
• RECOMMEND TO A LIBRARIAN

Speech-Language Pathology Practices for Adults With Right Hemisphere Stroke: What Are We Missing?

• Ashley Ramsey , and
• Margaret Lehman Blake

Department of Communication Sciences & Disorders, University of Houston, TX

Google Scholar

More articles by this author

• Add to favorites
• Download Citation
• Track Citations

Limited evidence exists to guide the assessment and treatment of cognitive-communication disorders associated with right hemisphere stroke. The purpose of this study was to obtain information about speech-language pathologists' (SLPs') clinical practices and decision making for this population to understand what practices are being used and identify gaps in clinical practice.

A survey was distributed via online ASHA Communities for the Special Interest Groups and other social media platforms. Respondents included 143 SLPs from across the United States representing 3–50 years of experience and a wide range of practice settings. Survey questions probed assessment practices including how tests are selected, what tests are used to diagnose specific deficits, and how confident SLPs were in their diagnoses. Treatment decisions were queried for a small set of disorders.

SLPs routinely assess cognitive disorders using standardized tests. Communication disorders are less likely to be formally assessed. Three core right cerebral hemisphere deficits—anosognosia, aprosodia, and pragmatic deficits—are either not assessed or assessed only through observation by 80% of SLPs. Evidence-based treatments are commonly used for disorders of attention, awareness, and aprosodia.

Communication disorders are less likely to be formally assessed than cognitive disorders, creating a critical gap in care that cannot be filled by other allied health professionals. Suggestions for free or low-cost resources for evaluating pragmatics, prosody, and awareness are provided to aid SLPs in filling this gap.

Supplemental Material

https://doi.org/10.23641/asha.12159597

• American Speech-Language-Hearing Association . (n.d.). ASHA Community [Homepage] . https://community.asha.org/ Google Scholar

Additional Resources

• Cited by Frontiers in Neurology 15 13 Aug 2024 Stroke rehabilitation: from diagnosis to therapy Xiaohong Li , Yanjin He , Dawu Wang and Mohammad J. Rezaei Glossa :140 (25-51) 2 Aug 2024 Le test de compréhension de l’IRonie et des Requêtes Indirectes – version courte (IRRI-C) : développement, validité de contenu et données normatives préliminaires. The IRony and Indirect Request comprehension test - short version (IRRI-C): development, content validity and preliminary normative data. Natacha Cordonier , Maud Champagne-Lavau and Marion Fossard Journal of Speech, Language, and Hearing Research 67:2 (511-523) 12 Feb 2024 Identifying Spatial Neglect in Chronic Right Hemisphere Stroke Survivors Using the RHDBank Outcomes Jamila Minga , Timothy Rich , Olga Boukrina , Peii Chen and Kimberly Hreha American Journal of Speech-Language Pathology 32:5 (2351-2373) 11 Sep 2023 Perspectives of Speech-Language Pathologists and Students on Providing Care to People Living With Dementia: A Scoping Review Jamie F. Mayer , Makenna R. Green , Laura W. White and Trey Lemley Applied Neuropsychology: Adult (1-9) 30 Mar 2023 Undetected language deficits in left or right hemisphere post-stroke patients Maria Martzoukou , Anastasia Nousia and Grigorios Nasios American Journal of Speech-Language Pathology 32:1 (96-106) 11 Jan 2023 The Functional External Memory Aid Tool Version 2.0: A How-To Clinical Guide Alyssa M. Lanzi , Anna K. Saylor , Robert F. Dedrick , Michelle S. Bourgeois and Matthew L. Cohen Melissa Johnson and Jessie Preston (2023) Assessing Discourse Ability in Adults with Right Hemisphere Damage Spoken Discourse Impairments in the Neurogenic Populations 10.1007/978-3-031-45190-4_15 Disability and Rehabilitation 44:26 (8524-8538) 18 Dec 2022 Management of communication disability in the first 90 days after stroke: a scoping review Caroline Baker , Abby M. Foster , Sarah D’Souza , Erin Godecke , Ciara Shiggins , Edwina Lamborn , Lucette Lanyon , Ian Kneebone and Miranda L. Rose International Journal of Language & Communication Disorders 14 Dec 2022 Optimizing our evidence map for cognitive–communication interventions: How it can guide us to better outcomes for adults living with acquired brain injury Sheila MacDonald and Elyse Shumway Communication Sciences & Disorders 27:3 (606-616) 30 Sep 2022 The Korean Version of the Right Hemisphere Language Battery: Adaptation and Reliability Jihyeon Yun , Jee-Hye Chung , Yeong-Wook Kim and Il-Young Jung American Journal of Speech-Language Pathology 31:5 (1949-1962) 7 Sep 2022 Clinical Guidelines for Eliciting Discourse Using the RHDBank Protocol Jamila Minga , Melissa D. Stockbridge , Alexandra Durfee and Melissa Johnson PLOS ONE 17:8 (e0271727) 11 Aug 2022 Effect of right hemispheric damage on structured spoken conversation Yeo Jin Kim , Hye Yeong Jeong , Hui-Chul Choi , Jong-Hee Sohn , Chulho Kim , Sang-Hwa Lee , Joon Soo Shin , So Ra Chin , Yoon Kyoung Lee , So Jung Oh , Ji Hye Yoon and Daniel Mirman American Journal of Speech-Language Pathology 31:4 (1653-1671) 12 Jul 2022 Survey Results of Speech-Language Pathologists Working With Cognitive-Communication Disorders: Improving Practices for Mild Cognitive Impairment and Early-Stage Dementia From Alzheimer's Disease Alyssa M. Lanzi , Anna K. Saylor and Matthew L. Cohen Aphasiology 36:6 (669-686) 3 Jun 2022 Test item priorities for a screening tool to identify cognitive-communication disorder after right hemisphere stroke Amanda Love , Petrea Cornwell , Ronelle Hewetson and David Shum American Journal of Speech-Language Pathology 31:2 (689-704) 10 Mar 2022 Conceptual Framework Behind the Development of a Level of Confidence Tool: The Pediatric Videofluoroscopic Swallow Study Value Scale Leann Schow Smith and Julie M. Barkmeier-Kraemer Brain Sciences 11:5 (667) 20 May 2021 Explicit Training to Improve Affective Prosody Recognition in Adults with Acute Right Hemisphere Stroke Alexandra Zezinka Durfee , Shannon M. Sheppard , Erin L. Meier , Lisa Bunker , Erjia Cui , Ciprian Crainiceanu and Argye E. Hillis American Journal of Speech-Language Pathology 30:2S (986-992) 16 Apr 2021 Confidence and Training of Speech-Language Pathologists in Cognitive-Communication Disorders: Time to Rethink Graduate Education Models? Emily L. Morrow , Lyn S. Turkstra and Melissa C. Duff Topics in Language Disorders 41:1 (99-122) 1 Jan 2021 Making Sense of Right Hemisphere Discourse Using RHDBank Jamila Minga , Melissa Johnson , Margaret Lehman Blake , Davida Fromm and Brian MacWhinney American Journal of Speech-Language Pathology 29:4 (1821-1832) 12 Nov 2020 Potential for Cognitive Communication Impairment in COVID-19 Survivors: A Call to Action for Speech-Language Pathologists Amy E. Ramage American Journal of Speech-Language Pathology (1-8) The Right ICD Code, Right Now: A Call to Action for Pragmatic Language Disorders After Right Hemisphere Stroke Jamila Minga , Shanika Phillips Fullwood , Deborah Rose and Danai Kasambira Fannin
• Get Permissions

• Received: Sep 6, 2019
• Revised: Dec 19, 2019
• Accepted: Jan 2, 2020
• Published online: Apr 24, 2020
• Published in issue: May 8, 2020
• asha-topics
• asha-article-types

Copyright © 2020 American Speech-Language-Hearing Association

Loading ...

Painting after Right-Hemisphere Stroke – Case Studies of Professional Artists

• February 2007
• Frontiers of Neurology and Neuroscience 22:1-13
• This person is not on ResearchGate, or hasn't claimed this research yet.

• Universität Heidelberg

Discover the world's research

• 25+ million members
• 160+ million publication pages
• 2.3+ billion citations

No full-text available

To read the full-text of this research, you can request a copy directly from the authors.

• PHYS LIFE REV

• Jozsef Arato

• Shole Vatanparasti
• Anoshirvan Kazemnejad

• Eugen Bogdan Petcu
• Katherine Sherwood

• Rodica Ileana Miroiu

• Julien Bogousslavsky
• Dahlia W. Zaidel
• Eunjoo Rhee
• Connie Hong

• Byung-Chul Lee
• NEUROPSYCHOLOGIA

• Britt Anderson

• PSYCHIAT DANUB

• Marina Roje Bedeković

• Maria Bazan
• Philos Ethics Humanit Med

• OCCUP MED-OXFORD
• Mike McKiernan

• Luis Fornazzari

• Isik Bayraktar

• Thomas Meissner
• Janet W H Sit
• Amy W. H. Chan

• ART PSYCHOTHER

• Elena Sinforiani

• Geoffrey D Schott

• REV CLIN ESP
• Recruit researchers
• Join for free
• Login Email Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google Welcome back! Please log in. Email · Hint Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google No account? Sign up

American Speech-Language-Hearing Association

• Certification
• Publications
• Continuing Education
• Practice Management
• Audiologists
• Speech-Language Pathologists
• Academic & Faculty
• Audiology & SLP Assistants

Right Hemisphere Disorder

View All Portal Topics

The scope of this Practice Portal page is right hemisphere disorder (RHD) —a unique constellation of deficits associated with acquired right-side brain injury in adults.

See ASHA’s Right Hemisphere Disorder Evidence Map for summaries of the available research on this topic. See also ASHA’s Evidence Maps on Stroke and Traumatic Brain Injury for research related to right hemisphere damage in these populations.

RHD is most commonly caused by a stroke or other acquired brain injury (e.g., stroke, tumor) that impacts the right hemisphere of the brain. RHD is a constellation of changes in

• pragmatics —the ability to convey or comprehend meaning or intent of a message;
• discourse —the ability to understand or produce verbal and written language in units longer than single sentences; and
• cognitive-communication skills —the cognitive skills that are needed for effective, clear communication, including attention, memory, executive function, visual-perceptual skills, and/or awareness of deficits.

Communication deficits caused by brain injury often co-occur with other cognitive deficits. These include the following:

• Anosognosia —reduced awareness of neurological deficits and other changes following brain injury.
• Egocentric unilateral spatial neglect (i.e., reduced awareness of visual stimuli to one side of the individual’s midline) is the most common (Kleinman et al., 2007).
• Neglect may involve visual, auditory, somatosensory, or kinetic modalities.
• This may co-occur with neglect dyslexia —misreading or not detecting text on the left side of the page or on the left side of words (Siéroff, 2017).

Word retrieval, syntax, morphology, and phonological processing are not typically affected by injury to the right hemisphere. However, these deficits occur with right hemisphere stroke in a small percentage of patients. This phenomenon is called crossed aphasia . This condition may occur in people with language dominance in the right hemisphere at baseline. Most people are left hemisphere dominant for language, so crossed aphasia is rare.

Although the deficits associated with RHD may be subtle in highly structured contexts, they are often more apparent during dynamic and/or complex tasks such as conversation (Ferré et al., 2011). These deficits can significantly impact functional performance in social and vocational settings (Blake, 2006; Lehman & Tompkins, 2000).

Realizing the potential impact of RHD on daily functioning is particularly important, as the deficits experienced by people with acute RHD often go unrecognized and undiagnosed (Edwards et al., 2006). This may lead to reduced referrals for speech-language pathology or other rehabilitation services and prolonged, negative impacts for people with RHD. In addition, RHD can lead to disrupted social relationships (Hewetson et al., 2021) and difficulty maintaining jobs and other social activities (Tompkins, 2012).

Incidence and Prevalence

Incidence is the number of new cases of a disorder or condition identified in a specific time period.

Prevalence is the number of individuals who are living with a disorder or condition in a given time period.

There are no population-level statistics on the incidence and prevalence of RHD. Therefore, the following estimates should be interpreted with caution. Statistics are most often reported within stroke populations because of the localized nature of RHD.

Hospital-based studies have reported that 42.3%–47.2% of people presenting with stroke have right hemisphere strokes (Deb-Chatterji et al., 2022; Hedna et al., 2013; Portegies et al., 2015). However, these statistics may be an underestimate. Patients may be less likely to seek medical attention with right hemisphere stroke due to diminished insight into their new deficits and the reduced recognizability of RHD symptoms (e.g., visual or cognitive-communication deficits) when compared to those of the left hemisphere (e.g., aphasia; Deb-Chatterji et al., 2022; Foerch et al., 2005).

The reported rates of disorders of the right hemisphere are as follows:

• Anosognosia is frequently present in individuals with traumatic brain injury; however, prevalence rates are not available for individuals with right hemisphere damage in isolation. Across multiple studies, 76.9%–97.1% of individuals with traumatic brain injury had limited awareness of deficits at discharge (Steward & Kretzmer, 2022). Persistent deficits were noted in 66.2%–87.5% of these individuals in the subacute stage, and 35%–60% of individuals continued to have reduced insight at 6–10 months post-injury (Steward & Kretzmer, 2022).
• Aprosodia is estimated to be present in 50%–70% of individuals with right hemisphere brain damage (Sheppard et al., 2020; Ukaegbe et al., 2022).
• Cognitive-communication deficits are estimated to occur in 50%–90% of all individuals with right hemisphere brain damage (Ferré & Joanette, 2016; Hewetson et al., 2017).
• Crossed aphasia (i.e., aphasia that results from right hemisphere brain damage) is rare. Historical reports have estimated that crossed aphasia occurs in under 3% of individuals with vascular etiologies (Lahiri et al., 2019). However, one recent study reported crossed aphasia in 6.73% of individuals with first-time strokes (Lahiri et al., 2019).
• Pragmatic deficits were reported in 16.3%–29.6% individuals with right hemisphere brain damage in two retrospective analyses of an inpatient rehabilitation unit (Blake et al., 2002, 2003). However, these numbers likely underestimate the prevalence of pragmatic deficits in this population due to reduced SLP referral and a general lack of comprehensive evaluation of pragmatic skills (Blake et al., 2002).
• Spatial neglect is estimated to occur in 33%–82% of individuals with right hemisphere stroke (Barrett, 2021). The overall prevalence of neglect dyslexia in individuals with RHD is not known. However, one study of 138 individuals with right hemisphere stroke found that 22.5% exhibited neglect dyslexia during the acute phase of their recovery (Lee et al., 2009).

Signs and Symptoms

RHD results in a collection of symptoms that vary in severity and in domains affected depending on the site and extent of injury. For a detailed discussion of signs and symptoms associated with RHD, see Blake (2018).

Below are examples of symptoms grouped by domain. Individuals may not present with all symptoms.

Apragmatism

Apragmatism is when a person has difficulty conveying or comprehending the meaning or intent of a message within a specific context. Contexts can include the conversational partner(s), environment, culture, or goals of the interaction. Apragmatism is a primary communication impairment in RHD (Minga et al., 2023).

Apragmatism can be divided into three areas: Linguistic, paralinguistic, and extralinguistic.

• have trouble using or recognizing sarcasm, jokes, figurative language, or information that can be interpreted in multiple ways (Lundren & Brownell, 2016);
• have difficulty making inferences or understanding global meanings of discourse—such as the implied main idea or the overall gist of the story or discussion (Tompkins et al., 2004, 2008); and/or
• be tangential or verbose and may interrupt—or may have reduced verbal output (Blake, 2006).
• have aprosodia —or the inability to understand and express meaning and emotion through the use of variations in pitch, loudness, intonation, and rhythm (Stockbridge et al., 2022).
• variations in facial expressions,
• body language, and
• the use of gestures or eye contact.

Cognitive Communication

RHD affects aspects of cognitive communication that impact how the person interacts with others and with their environment. Common areas of impairment include

• awareness of deficits;
• attention (all forms, including unilateral neglect);
• executive functioning (e.g., working memory, inhibitory control, cognitive flexibility);
• problem solving;
• reasoning and judgment; and
• sequencing.

For more information about the executive functioning deficits that occur across brain injuries of varying etiologies, please see ASHA’s Practice Portal page on Executive Function Deficits .

RHD also affects discourse —language units larger than a sentence that have a specific purpose and meaning together.

Discourse-level communication involves cognition and language —both of which are commonly impaired in RHD. As such, subtle deficits in cognition and/or language may be more apparent in discourse-level tasks than in discrete tasks.

Cognitive deficits in RHD may make efficient and appropriate discourse management challenging. However, healthy aging can also contribute to changes in discourse over time. Clinicians can determine if there are any baseline or age-related factors that may influence a patient’s discourse-level communication to understand the true impact of RHD.

Discourse-level changes may impact the following:

• May include too little relevant information about stories and procedures.
• May disproportionately exclude inferred content (vs. explicit content).
• May have difficulty suppressing information about themselves or about a topic for which they feel strongly.
• May struggle to present information sequentially.
• track references from sentence to sentence,
• plan and adapt to listener knowledge, and
• monitor listener understanding.
• May have difficulty with ambiguity, such as when to use “a” versus “the” to convey introduced content (Barker et al., 2017). They may vary the label when referring to the same subject, which can lead to confusion for their conversational partner (Stockbridge et al., 2022).
• May have difficulty identifying instances of communication breakdown and misunderstanding or may have difficulty achieving effective conversational repair when they do identify a breakdown.

Other Deficits

Other deficits that may be associated with RHD include

• pseudobulbar affect, which can cause lability (e.g., crying or inappropriate laughing); difficulty interpreting and conveying emotions;
• reduced empathy;
• egocentrism, or the use of language that is excessively self-focused and preoccupied with the person’s own thoughts, feelings, and needs;
• dysarthria; and
• hemiparesis/hemiplegia.

See ASHA’s Practice Portal pages on Adult Dysphagia and Dysarthria in Adults .

RHD may result from a variety of changes in the structure or function of the right hemisphere of the brain. These can range in severity and may result in chronic or acute deficits. Changes in the brain include tumors, surgery, infection, stroke, seizure, neurodegenerative conditions, and traumatic brain injury.

Please see the following Practice Portal pages for further information: Head and Neck Cancer and Traumatic Brain Injury in Adults .

Roles and Responsibilities

SLPs play a central role in the screening, assessment, diagnosis, and treatment of persons with RHD. The professional roles and activities in speech-language pathology include

• clinical/educational services (diagnosis, assessment, planning, and treatment);
• prevention, counseling, and advocacy; and
• education, administration, and research.

See ASHA’s Scope of Practice in Speech-Language Pathology (ASHA, 2016).

Appropriate roles for SLPs include, but are not limited to, the following:

• Screening individuals with a history of right hemisphere brain injury, and determining the need for further assessment and/or referral for other services.
• Conducting a culturally and linguistically relevant, comprehensive assessment of language, communication, and cognition.
• Diagnosing cognitive and communication disorders, the characteristics of these disorders, and their functional impact.

Counseling and Education

• Educating and counseling people with RHD and their care partners on cognitive and communication-related issues, and facilitating participation across contexts (i.e., family, vocational, and community).
• Providing prevention information to individuals and groups known to be at risk for conditions associated with RHD (e.g., stroke and traumatic brain injury).
• Educating other professionals on the needs of persons with RHD and the role of SLPs in diagnosing and managing deficits associated with this disorder.
• Making decisions about the management of disorders related to RHD in collaboration with the patient, family, and interprofessional treatment team. See ASHA’s resources on interprofessional education/interprofessional practice (IPE/IPP) and person- and family-centered care .
• Developing person-centered treatment plans, providing intervention and support services, documenting progress, and determining appropriate dismissal criteria.
• Consulting and collaborating with other professionals to facilitate program development and to provide supervision, evaluation, and/or expert testimony, as appropriate.
• Advocating for individuals with RHD and their families at the local, state, and national levels.
• Remaining informed of research in the area of RHD and helping advance the knowledge base related to the nature and treatment of RHD.

As indicated in the ASHA Code of Ethics (ASHA, 2023), individuals who hold the Certificate of Clinical Competence shall engage in only those aspects of the professions that are within the scope of their professional practice and competence, considering their certification status, education, training, and experience.

See ASHA’s Right Hemisphere Disorder Evidence Map for pertinent scientific evidence, expert opinion, and client/caregiver perspective. See also ASHA’s Evidence Maps on Stroke and Traumatic Brain Injury for information related to right hemisphere brain damage in these populations.

The clinician considers the following factors that may have an impact on screening and comprehensive assessment:

• language(s) used
• concurrent motor speech impairment (e.g., dysarthria)
• hearing loss and auditory agnosia —the inability to recognize or differentiate between sounds or the brain’s neurological inability to process sound meaning
• visual acuity deficits, visual field cuts, and visual agnosia —the inability to recognize or interpret visual stimuli (e.g., objects, faces)
• upper extremity hemiparesis (may affect the ability to write, sign, or gesture or to access augmentative and alternative communication devices)
• presence of chronic pain (from either preexisting or new conditions) and/or acute pain
• endurance and fatigue (testing may need to be broken into shorter sessions)
• potential impact of prescription drugs on the individual’s presentation and test performance (e.g., excessive drowsiness, exacerbation of cognitive problems secondary to polypharmacy)
• emotional and psychological status (e.g., poststroke depression)
• premorbid functional status and social determinants of health
• anticipated/preferred discharge setting (may indicate the level of supervision and independence that would be required to succeed)

If the individual with RHD wears prescription glasses and/or hearing aids, then they should wear these items during assessment.

If additional hearing and/or visual deficits resulted from the neurological event—and physical or environmental modifications (e.g., large-print material, attention to placement of test stimuli, modified lighting, amplification devices) are not sufficient to compensate for these changes—then the individual should be referred for complete audiologic and/or vision assessments prior to testing. The individual should be referred to psychology, psychiatry, and/or neuropsychiatry if there are signs or reports of depression, emotional lability, or other psychological issues.

Screening is a procedure for identifying the need for further assessment and does not provide a detailed description of the severity and nature of the deficits associated with RHD. Screening is conducted in the language(s) used by the person, with sensitivity to cultural and linguistic diversity. Screening may be completed by the SLP or another appropriately trained professional. Standardized and nonstandardized methods are used to screen oral motor functions, speech production skills, comprehension and production of spoken and written language, pragmatic language skills, and other cognitive skills (attention, memory, and executive function) as they relate to communication, swallowing, unilateral neglect, and hearing. Deficits in these skills may be related to RHD.

Screening often incorporates the use of targeted questionnaires with the individual and family members. Keep in mind that changes after RHD are not always recognized by the individual or family members.

Screening may result in

• a recommendation for comprehensive speech, language, swallowing, and cognitive-communication assessments and/or
• a referral for other examinations or services (e.g., complete audiologic assessment and/or vision testing as well as assessment by a psychiatrist or neuropsychologist).

Comprehensive Assessment

Effective RHD assessment relies on patient and care partner interviews to establish baseline communication function and to highlight behavioral changes. Assessment also considers normal age-related changes in cognitive-communication skills when gauging baseline function.

There are few standardized assessments for use with patients with RHD (see, e.g., Joanette et al., 2015). Functional assessment may more accurately predict performance on activities of daily living than standardized assessment and is particularly valuable for individuals with RHD. Assessment considers the impact of RHD on a patient’s quality of life. Appropriate treatment goals target the patient’s cognitive-communication deficits in an effort to restore that person to their maximum functional capacity and quality of life.

Typical components of a comprehensive assessment of deficits associated with RHD include the following:

Case History

• Relevant medical history (history of previous strokes or other neurological disorders).
• Patient interview (educational, social, and occupational history).
• Input from care partners or others close to the patient to identify changes from baseline.
• Impact of deficits on activities of daily living and overall daily functioning.
• Input from other medical professionals (e.g., physical and occupational therapists, neurologist, neuropsychologist, social worker).
• Cultural and linguistic backgrounds, including languages used across different contexts.
• Social determinants of health .

Individual and Care Partner Report

• Functional communication and cognitive-communication struggles and successes.
• Communication difficulties and impact on individual and their family/care partners.
• Contexts in which the individual has difficulty engaging (e.g., social interactions, work activities).
• Goals and preferences of the individual and their care partners.

Oral Mechanism Evaluation

• Strength, speed, and range of motion of components of the oral motor system.
• Sequential/alternating movement repetitions (i.e., diadochokinetic rates).
• Steadiness, tone, and accuracy of movements for speech and nonspeech tasks.
• Motor speech abilities (see the Assessment section of ASHA’s Practice Portal page on Dysarthria in Adults ).
• Phonation, including pitch and volume. Consider using nonlinguistic tasks to differentially diagnosis motor speech limitations from apragmatism.

Language Assessment, Including Discourse

• Discourse assessment in adults may include picture/picture series description, storytelling, and conversational sampling (Coelho et al., 2022). The SLP can use computerized discourse analysis tools (e.g., Computerized Language Analysis [CLAN]) to compare discourse abilities to the norm.
• Functional language use and quality-of-life self-report measures, including proxy measures, can help capture differences between communication challenges that occur in different environments (e.g., work vs. home).
• Reading decoding and comprehension—specifically at the paragraph level or longer—including looking for evidence of neglect dyslexia.
• Language comprehension and production in a variety of contexts (e.g., conversation, storytelling). See the RHDBank protocol for guidance.
• Discourse analysis to assess the strengths and challenges of various forms of discourse and to help identify treatment targets. This includes global coherence (Leaman & Edmonds, 2021; Wright & Capilouto, 2012), main concept analysis (Dalton & Richardson, 2019), and story grammar (Greenslade et al., 2020).

Pragmatics Assessment

• Linguistic Pragmatics—use of language that is appropriate to the context and is able to clearly communicate intended meanings or interpreting and expressing meaning through the use of word selection and grammar.
• Paralinguistic Pragmatics—use of prosody to understand and express feelings, emotion, tone, and implied information.
• Extralinguistic Pragmatics—the use, recognition, and interpretation of nonverbal cues (e.g., facial expression, eye contact, gesture) for communication.
• social participation restrictions and
• satisfaction with social participation across multiple settings.

See Minga et al. (2023) for further information.

Cognitive-Communication Assessment

Areas that are assessed in terms of cognitive communication include the following:

• awareness of deficits
• metacognition, cognitive flexibility, and theory of mind
• impulsivity
• judgment and safety awareness
• memory (verbal and nonverbal; short-term, episodic, and working)
• problem solving and reasoning
• visuospatial awareness in one’s environment (e.g., navigating, finding items on the person’s left side)

Feeding and Swallowing Assessment

Deficits that frequently occur with RHD impact feeding and swallowing. These deficits include

• visuospatial deficits;
• hemispatial neglect; and
• level of arousal and cognitive status,
• poor initiation (e.g., having food in front of them and not eating),
• impulsivity (e.g., pocketing food, eating quickly and choking),
• impaired awareness of deficits (e.g., not following diet recommendations), and
• reduced attention (e.g., difficulty using swallowing strategies).

See the Assessment section of ASHA’s Practice Portal page on Adult Dysphagia .

Audiologic Assessment

Hearing and vestibular testing may be indicated, depending on the individual’s presenting needs. SLPs make referrals to audiologists as appropriate. For details, see the Assessment sections of ASHA’s Practice Portal pages on Hearing Loss in Adults , Balance System Disorders , and Tinnitus and Hyperacusis .

Assessment Results

Assessment may result in one or more of the following:

• Diagnosis of a cognitive-communication disorder and other deficits associated with RHD.
• Description of the characteristics, severity, and functional impact of the disorder.
• Statement regarding prognosis for improvement.
• Recommendations for intervention, support, and community resources.
• Referral for other assessments or services (e.g., neuropsychologist, physical therapist, occupational therapist, vocational counselor, neuro-ophthalmologist, audiologist).

Cultural and Linguistic Considerations

Pragmatic and social norms (e.g., eye contact, turn-taking, nonverbal cues) vary among different cultures . Cultural differences should not be interpreted as pragmatic deficits. See ASHA’s Practice Portal page on Cultural Responsiveness for more information.

When selecting the language of assessment, it is important to consider the patient’s preference, language(s) spoken, age of acquisition of each language, premorbid use of each language, and language(s) needed for return to daily activities. Clinicians should gather data in all languages used by the client and their care partners to determine the degree of impact.

Prompts and cues used in assessment may not carry the same meaning for individuals from one culture to another. Any accommodations and/or modifications to the testing process to reconcile cultural and linguistic variations should be documented. Scores from standardized tests should be interpreted and reported with caution in these cases. See ASHA’s Practice Portal pages on Multilingual Service Delivery in Audiology and Speech-Language Pathology ; Collaborating With Interpreters, Transliterators, and Translators ; and Cultural Responsiveness .

See ASHA’s Right Hemisphere Disorder Evidence Map for pertinent scientific evidence, expert opinion, and client/caregiver perspective. See also ASHA’s Evidence Maps on Stroke and Traumatic Brain Injury for information related to right hemisphere brain damage in this population.

Treatment for RHD is individualized to address areas of need identified in the assessment, considering the goals identified by the individual and their care partners.

Treatment is provided in the language(s) used by the individual with RHD. Services may be provided either by a multilingual SLP or in collaboration with trained interpreters. See ASHA’s Practice Portal page on Collaborating With Interpreters, Transliterators, and Translators .

For a detailed discussion of RHD treatment, see, for example, Blake (2018) and Myers (1999, 2001).

Treatment approaches can be restorative, compensatory, or a combination of the two.

• Restorative —goal is to improve the underlying impairment.
• Compensatory —goal is to use strategies to increase success despite existing impairments.

People with RHD often have limited insight into their deficits. This can decrease participation in either therapeutic approach and can limit an individual’s ability to generalize compensatory strategies taught in treatment sessions.

Treatment selection depends on the person’s communication and activity participation needs, the preferences of the person and their care partners, and the presence of co-occurring conditions. See ASHA’s Practice Portal page on Cultural Responsiveness for more information.

Below are descriptions of treatment options for addressing RHD, although there are few published treatment approaches available.

Apragmatism and Discourse

Treatment for apragmatism and discourse focus on improving communication skills across a variety of settings. Treatment often involves making implicit communication practices explicit. SLPs typically work to increase the individual’s awareness of their pragmatic deficits when compared to social norms. The clinician asks questions to better understand the person’s norms and premorbid communication behaviors. Techniques used to practice these skills include coaching, one-on-one rehearsal, role play, group practice, visual and verbal feedback, and video modeling. There are four aspects of apragmatism and discourse:

• linguistic apragmatism
• paralinguistic apragmatism
• extralinguistic apragmatism
• cognitive-communication skills

Each of these four aspects is discussed in the subsections below.

Linguistic Apragmatism

Treatment considers the person’s language use including appropriateness to context, if they can clearly communicate, and if they can correctly interpret the meaning of others’ language. This includes (a) using conversational skills, (b) applying inference and using global meanings of discourse, and (c) understanding and using alternate meanings.

Conversational Skills

Treatment considers the person’s premorbid behaviors and cultural norms and includes explicit instruction on how to use and monitor strategies for successful conversation —for example:

• improved turn-taking
• reduced interruptions
• reduced tangential comments
• less abrupt beginnings and endings
• context-appropriate social boundaries
• improved question asking

Conversational skills may also be supported by providing direct instruction on theory of mind —which is the ability to understand the mental states of others and how their mental states may differ from your own. It includes

• considering others’ beliefs, attitudes, emotions, and intentions in social situations, and
• understanding that one’s own beliefs may differ from other people’s beliefs.

Inference and Global Meanings of Discourse (Topic, Gist, Big Picture)

People with RHD may have difficulty understanding the key points or the core message or topic of conversation.

Treatment includes

• labeling items in scenes or stories,
• identifying the relevant or significant items,
• explaining the relationship among the words and concepts, and
• organizing printed sentences into a narrative or placing pictures into a logical sequence.

Understanding and Using Alternate Meanings

People with RHD may have trouble understanding ambiguities and nonliteral language as well as recognizing multiple meanings of words.

• grouping words according to their connotative meaning (i.e., alternate or secondary meaning);
• providing multiple meanings for homographs (e.g., “left” = the direction vs. “left” = the verb meaning “went”) or homophones (e.g., “son” vs. “sun”); and
• understand ambiguities (at the word and sentence levels),
• interpret figurative language such as metaphors and figures of speech (Lundgren et al., 2011), and
• generate alternative meanings to ambiguous sentences.

Paralinguistic Apragmatism (Prosody)

Paralinguistic apragmatism, or prosody, is the set of variations in the suprasegmental aspects of language (e.g., rate, pitch or intonation, and intensity).

Prosody can convey linguistic content such as rising intonation for a yes/no question or part of speech (e.g., “PREsent” vs. “present”). Prosody also can convey emotional or affective information (e.g., anger, happiness). Deficits in prosody (aprosodia) can be expressive and/or receptive.

For examples of prosodic treatments, see Leon et al. (2005), Rosenbek et al. (2004, 2006), and Durfee et al. (2021).

Restorative treatment of prosody may include the following:

• Explicit instruction in the features of prosody and their role in supporting meaning (Durfee et al., 2021).
• Cognitive–linguistic approach—wherein people with RHD explicitly define the intended emotion and its features and then practice producing those features, with gradually fading cues (Rosenbek et al., 2004).
• Imitative approach—wherein people with RHD imitate clinician productions of target prosodic variations, with gradually fading cues.
• manipulating and varying prosodic features during imitation, reading tasks, or conversation to match a target meaning or emotion and
• Client: Tom went to a football game.
• Clinician: Tom went to a basketball game?
• Client: No, Tom went to a football game. 
• Clinician: Marissa went to the football game?
• Client: No, Tom went to the football game.
• identifying the emotion(s) conveyed by speakers within the context of audio or video clips;
• engaging in role-playing exercises;
• participating in natural conversation with the clinician or other communication partners; and
• identifying differences and similarities in the prosodic features between multiple audio, video, or spoken samples (e.g., words, phrases, or sentences).

Treatment can also use compensatory strategies, including

• identifying nonprosodic elements that convey emotion (e.g., word choice, facial expressions, body language);
• asking communication partners to state their emotions at the beginning of a conversation to help avoid miscommunication (e.g., “I’m really upset right now”); and
• encouraging or prompting the person with RHD to verbalize their emotional state or intent at the beginning of the conversation.

Prosodic features—and how people use them to convey meaning—vary across languages, and goals may need to be language specific. See ASHA’s page on Phonemic Inventories and Cultural and Linguistic Information Across Languages for more information.

Extralinguistic Apragmatism

Treatment for extralinguistic apragmatism considers the person’s premorbid behaviors and cultural norms and includes explicit instruction on how to use and monitor the following behaviors:

• changing facial expressions based on the tone of the message
• using specific gestures to emphasize meaning
• nodding to indicate understanding as a listener
• turning toward the person who is talking
• considering the other person’s beliefs, attitudes, emotions, and intentions in social situations and
• understanding that one’s own beliefs may differ from those of others.

Cognitive-Communication Skills

Attention, memory, and executive functions (discussed in the subsections below) are common targets for treatment in people with RHD (see, e.g., Tompkins, 2012). These skills influence the dynamic aspects of communication and functional independence.

Restorative approaches are aimed at improving one or more types of attention (e.g., sustained, selective, alternating, divided). Restorative approaches include tasks that require the person to keep a target action or response in mind in the presence of confounding variables—variables such as competing distractors, dual-task training, and task shifting. Treatment options include

• computerized attention training programs (e.g., monitoring a computer screen for a target that appears in one of four quadrants),
• cancellation tasks, and
• recognizing targets or patterns in auditory or visual information.

Task complexity can be varied by the amount of material presented, the rate of presentation, the number of targets, or the relationship between targets (e.g., “Tell me if the next number in the sequence is higher than the one before it”).

Compensatory and metacognitive approaches help the person attend to a task until they complete it. These approaches include the following:

• listing small steps and the expected timeline for each step
• identifying potential distractions and ways to overcome them
• predicting performance before the activity
• self-monitoring (e.g., identifying episodes of reduced attention and potential causes)
• using strategies (e.g., reading aloud, self-talk, timed breaks) throughout, as needed
• comparing expected performance on an activity with actual performance
• evaluating the value of the strategies used
• generating new strategies
• avoiding or reducing competing auditory or visual stimuli (e.g., turning off the TV during conversation or other tasks, avoiding noisy restaurants);
• identifying the best time of day to complete important tasks (e.g., after a nap, before a physical therapy appointment); and
• organizing the workspace and removing distracting items.

People with RHD may experience three types of attention disorders: unilateral neglect, left-sided neglect, and neglect dyslexia.

Unilateral Neglect

Unilateral neglect is an attention disorder that frequently occurs with anosognosia —reduced insight into one’s own deficits. People with RHD may experience the neglect of visual, auditory, and/or tactile stimuli from one side of their body and/or the environment. This includes proprioceptive feedback —or information sent to the brain that allows an individual to know where their body is in space.

Left-Sided Neglect

Left-sided neglect, is common in people with RHD. This section focuses on one particular subtype—called left visual neglect —particularly as it affects communication and functional independence.

It is important to understand the person’s visual acuity (i.e., how sharp someone's vision is at a distance), and the presence of a visual field cut (i.e., missing a part of the area a person can typically see) before selecting treatment strategies. Clinicians may consult with or refer to neuro-optometry or neuro-ophthalmology as appropriate.

Treatment approaches include the following:

• e.g., highlighting the left edge of a screen/paper or placing high-priority items to the person’s left.
• e.g., presenting the person with written sentences in paragraph form, which requires the person to read text in the entire paragraph—including the neglected space—to understand the sentence.
• e.g., engaging in route-finding tasks, reading aisle signs while grocery shopping, interacting with unfamiliar partners in the community.

Neglect Dyslexia

Neglect dyslexia is a reading impairment that can occur in people with RHD wherein they omit or misread text on the left side of the page or on the left side of individual words. Treatment can include compensatory strategies designed to draw attention to the left side of the text (e.g., a bold, red line down the left side of the page). Other treatment approaches have been trialed, although there is no consensus as to the best method for treating this disorder (see, e.g., Gordon et al., 1985; Reinhart et al., 2011).

Compensatory treatment is often used to address memory deficits in people with RHD. This includes the use of external aids and internal strategies.

External aids include

• alarms, timers, and electronic reminders about upcoming tasks or events (e.g., medication reminders, upcoming appointments);
• calendars to track appointments and important events;
• journals to document details of events or activities;
• labels (e.g., on cabinets and drawers) to indicate content;
• note-taking (e.g., during phone calls or meetings);
• photographs and other visual supports (e.g., representing a sequence of steps in a task); and
• to-do lists.

Internal strategies include

• mnemonics —creating an acronym or a phrase using the first letter of each item in a list;
• internal repetition or
• visualization of a task being performed and completed repeatedly; and
• semantic elaboration —identifying and describing salient features of the information to be remembered, and then linking these features with preexisting knowledge.

Restorative treatment is also used to treat memory deficits. Treatment may include

• recalling information during unrelated tasks,
• computer-assisted training,
• spaced retrieval tasks, and
• working memory tasks.

Please see ASHA’s Practice Portal page on Traumatic Brain Injury for further information.

Executive Functions

Executive functions are also common treatment targets for people with RHD. Please see ASHA’s Practice Portal page on Executive Function Deficits for further information.

Awareness of Deficits

Treatment to increase awareness of deficits and their functional impact include the following:

• discussing and predicting how deficits might affect day-to-day function, and identifying ways to minimize negative consequences
• planning and predicting task performance (e.g., using a graphic organizer or chart), and then comparing predicted and actual performance
• providing feedback either during or after a task, reviewing performance, and generating strategies for improvement
• considering and discussing the core skills needed to complete complex tasks, and identifying the person’s current level of independence for each skill (e.g., to drive independently, people have to be able to see signs and oncoming traffic, remember where they are going, and use dashboard information)
• emphasizing safety and risk avoidance (e.g., being aware of cognitive deficits and their impact on successfully navigating environments and complex tasks)
• incorporating family, friends, and care partners, when appropriate, to provide a personal perspective and insight into changes from trusted sources

The goal of treatments that address anosognosia and that involve increasing one’s awareness is to teach an individual how to use metacognitive skills to reflect on their own performance, challenges, and safety. This can drive the use of compensatory strategies, encourage participation in therapy, and limit risk (e.g., fall risk).

Service Delivery

In addition to determining the optimal treatment approach for individuals with RHD, other factors include the availability of specific types of services in a particular region, insurance coverage, pattern of recovery, potential for returning to school or work, and service delivery options, including the following:

• Format —structure of the treatment session (e.g., group vs. individual). Group treatment for people with RHD provides opportunities to target the dynamic aspects of communication and receive peer feedback.
• Provider —person providing the treatment (e.g., speech-language pathologist, multidisciplinary team, trained volunteer, care partner).
• Dosage —frequency, intensity, and duration of service.
• Timing —timing of intervention relative to the onset of RHD.
• Setting — location of treatment (e.g., hospitals, outpatient facilities, skilled nursing facilities, home, community-based settings). Telepractice may provide an opportunity to introduce multiple settings but may be impacted by the presence and severity of unilateral neglect and other cognitive-communication deficits.

ASHA Resources

• Clinician’s Guide to Cognitive Rehabilitation in Mild Traumatic Brain Injury: Application for Military Service Members and Veterans [PDF]
• Evaluating and Treating Communication and Cognitive Disorders: Approaches to Referral and Collaboration for Speech-Language Pathology and Clinical Neuropsychology
• Interprofessional Education/Interprofessional Practice (IPE/IPP)
• Person- and Family-Centered Care
• Person-Centered Focus on Function: Traumatic Brain Injury [PDF]
• Right Hemisphere Brain Damage (RHD) [Consumer Information]
• Traumatic Brain Injury (TBI) [Consumer Information]
• Traumatic Brain Injury: A Primer for Professionals

Other Resources

This list of resources is not exhaustive, and the inclusion of any specific resource does not imply endorsement from ASHA.

• American Stroke Association
• Brain Injury Association of America (BIAA)
• RHDBank (for student and video links: User Name = student; Password = access)
• Right Hemisphere Brain Damage
• Right Hemisphere Stroke
• Social Communication and TBI: A Guide for Professionals

American Speech-Language-Hearing Association. (2016). Scope of practice in speech-language pathology [Scope of practice]. https://www.asha.org/policy/

American Speech-Language-Hearing Association. (2023). Code of ethics [Ethics]. https://www.asha.org/policy/

Barker, M. S., Young, B., & Robinson, G. A. (2017). Cohesive and coherent connected speech deficits in mild stroke. Brain and language , 168 , 23–36. https://doi.org/10.1016/j.bandl.2017.01.004

Barrett, A. M. (2021). Spatial neglect and anosognosia after right brain stroke. Continuum (Minneapolis, Minn.), 27 (6), 1624–1645. https://doi.org/10.1212/CON.0000000000001076

Blake, M. L. (2006). Clinical relevance of discourse characteristics after right hemisphere brain damage. American Journal of Speech-Language Pathology , 15 (3), 255–267. https://doi.org/10.1044/1058-0360(2006/024)

Blake, M. L. (2018). The right hemisphere and disorders of cognition and communication: Theory and clinical practice. Plural.

Blake, M. L., Duffy, J. R., Myers, P. S., & Tompkins, C. A. (2002). Prevalence and patterns of right hemisphere cognitive/communicative deficits: Retrospective data from an inpatient rehabilitation unit. Aphasiology , 16 (4–6), 537–547. https://doi.org/10.1080/02687030244000194

Blake, M. L., Duffy, J., Tompkins, C., & Myers, P. (2003). Right hemisphere syndrome is in the eye of the beholder. Aphasiology, 17 (5), 423–432. https://doi.org/10.1080/02687030344000120

Coelho, C., Cherney, L. R., & Shadden, B. B. (Eds.). (2022). Discourse analysis in adults with and without communication disorders: A resource for clinicians and researchers. Plural.

Dalton, S. G. H., & Richardson, J. D. (2019). A large-scale comparison of main concept production between persons with aphasia and persons without brain injury. American Journal of Speech-Language Pathology , 28 (1S), 293–320. https://doi.org/10.1044/2018_AJSLP-17-0166

Deb-Chatterji, M., Flottmann, F., Meyer, L., Brekenfeld, C., Fiehler, J., Gerloff, C., Thomalla, G., & GSR-ET-Investigators. (2022). Side matters: Differences in functional outcome and quality of life after thrombectomy in left and right hemispheric stroke. Neurological Research and Practice, 4 (1), Article 58. https://doi.org/10.1186/s42466-022-00223-7

Durfee, A. Z., Sheppard, S. M., Meier, E. L., Bunker, L., Cui, E., Crainiceanu, C., & Hillis, A. E. (2021). Explicit training to improve affective prosody recognition in adults with acute right hemisphere stroke. Brain Sciences , 11 (5), 667. https://doi.org/10.3390/brainsci11050667

Edwards, D. F., Hahn, M. G., Baum, C. M., Perlmutter, M. S., Sheedy, C., & Dromerick, A. W. (2006). Screening patients with stroke for rehabilitation needs: Validation of the post-stroke rehabilitation guidelines. Neurorehabilitation and Neural Repair , 20 (1), 42–48. https://doi.org/10.1177/1545968305283038

Ferré, P., & Joanette, Y. (2016). Communication abilities following right hemisphere damage: Prevalence, evaluation, and profiles. Perspectives of the ASHA Special Interest Groups, 1 (2), 106–115. https://doi.org/10.1044/persp1.SIG2.106

Ferré, P., Ska, B., Lajoie, C., Bleau, A., & Joanette, Y. (2011). Clinical focus on prosodic, discursive and pragmatic treatment for right hemisphere damaged adults: What’s right? Rehabilitation Research and Practice , 2011, 1–10 . https://dx.doi.org/10.1155/2011/131820

Foerch, C., Misselwitz, B., Sitzer, M., Berger, K., Steinmetz, H., & Neumann-Haefelin, T. (2005). Difference in recognition of right and left hemispheric stroke. The Lancet , 366 (9483), 392–393. https://doi.org/10.1016/S0140-6736(05)67024-9

Gordon, W. A., Hibbard, M. R., Egelko, S., Diller, L., Shaver, M. S., Lieberman, A., & Ragnarsson, K. (1985). Perceptual remediation in patients with right brain damage: A comprehensive program. Archives of Physical Medicine and Rehabilitation, 66 (6), 353–359.

Greenslade, K. J., Stuart, J. E. B., Richardson, J. D., Dalton, S. G., & Ramage, A. E. (2020). Macrostructural analyses of Cinderella narratives in a large nonclinical sample. American Journal of Speech-Language Pathology , 29 (4), 1923–1936. https://doi.org/10.1044/2020_AJSLP-19-00151

Hedna, V. S., Bodhit, A. N., Ansari, S., Falchook, A. D., Stead, L., Heilman, K. M., & Waters, M. F. (2013). Hemispheric differences in ischemic stroke: Is left-hemisphere stroke more common? Journal of Clinical Neurology , 9 (2), 97–102. https://doi.org/10.3988/jcn.2013.9.2.97

Hewetson, R., Cornwell, P., & Shum, D. (2017). Cognitive-communication disorder following right hemisphere stroke: Exploring rehabilitation access and outcomes. Topics in Stroke Rehabilitation , 24 (5), 330–336. https://doi.org/10.1080/10749357.2017.1289622

Hewetson, R., Cornwell, P., & Shum, D. H. K. (2021). Relationship and social network change in people with impaired social cognition post right hemisphere stroke. American Journal of Speech-Language Pathology , 30 (2S), 962–973. https://doi.org/10.1044/2020_AJSLP-20-00047

Joanette, Y., Ska, B., Côté, H., Ferré, P., Lapointe, L., Coppens, P., & Small, S. (2015). Montreal Protocol for the Evaluation of Communication (MEC). Australasian Society for the Study of Brain Impairment.

Kleinman, J. T., Newhart, M., Davis, C., Heidler-Gary, J., Gottesman, R. F., & Hillis, A. E. (2007). Right hemispatial neglect: Frequency and characterization following acute left hemisphere stroke. Brain and Cognition , 64 (1), 50–59. https://doi.org/10.1016/j.bandc.2006.10.005

Lahiri, D., Dubey, S., Sawale, V. M., Das, G., Ray, B. K., Chatterjee, S., & Ardila, A. (2019). Incidence and symptomatology of vascular crossed aphasia in Bengali. Cognitive and Behavioral Neurology , 32 (4), 256–267. https://doi.org/10.1097/WNN.0000000000000210

Leaman, M. C., & Edmonds, L. A. (2021). Measuring global coherence in people with aphasia during unstructured conversation. American Journal of Speech-Language Pathology , 30 (1S), 359–375. https://doi.org/10.1044/2020_AJSLP-19-00104

Lee, B. H., Suh, M. K., Kim, E.-J., Seo, S. W., Choi, K. M., Kim, G.-M., Chung, C.-S., Heilman, K. M., & Na, D. L. (2009). Neglect dyslexia: Frequency, association with other hemispatial neglects, and lesion localization. Neuropsychologia , 47 (3), 704–710. https://doi.org/10.1016/j.neuropsychologia.2008.11.027

Lehman, M. T., & Tompkins, C. A. (2000). Inferencing in adults with right hemisphere brain damage: An analysis of conflicting results. Aphasiology , 14 (5–6), 485–499. https://doi.org/10.1080/026870300401261

Leon, S. A., Rosenbek, J. C., Crucian, G. P., Hieber, B., Holiway, B., Rodriguez, A. D., Ketterson, T. U., Ciampitti, M. Z., Freshwater, S., Heilman, K., & Gonzalez-Rothi, L. (2005). Active treatments for aprosodia secondary to right hemisphere stroke. Journal of Rehabilitation Research & Development , 42 (1), 93–101.

Lundgren, K., & Brownell, H. (2016). Figurative language deficits associated with right hemisphere disorder. Perspectives of the ASHA Special Interest Groups , 1 (2) , 66–81.

Lundgren, K., Brownell, H., Cayer-Meade, C., Milione, J., & Kearns, K. (2011). Treating metaphor interpretation deficits subsequent to right hemisphere brain damage: Preliminary results. Aphasiology , 25 (4), 456–474. https://doi.org/10.1080/02687038.2010.500809

Minga, J., Fromm, D., Williams-DeVane, C., & MacWhinney, B. (2020) Question use in adults with right-hemisphere brain damage. Journal of Speech, Language, and Hearing Research , 63 (3),738-748. https://doi.org/10.1044/2019_JSLHR-19-00063

Minga, J., Sheppard, S. M., Johnson, M., Hewetson, R., Cornwell, P., & Blake, M. L. (2023). Apragmatism: The renewal of a label for communication disorders associated with right hemisphere brain damage. International Journal of Language & Communication Disorders, 58 (2), 651–666. https://doi.org/10.1111/1460-6984.12807

Myers, P. S. (1999). Right hemisphere damage: Disorders of communication and cognition . Singular.

Myers, P. S. (2001). Communication disorders associated with right hemisphere damage. In R. Chapey (Ed.), Language intervention strategies in aphasia and related neurogenic communication disorders (pp. 809–828). Lippincott Williams & Wilkins.

Portegies, M. L. P., Selwaness, M., Hofman, A., Koudstaal, P. J., Vernooij, M. W., & Ikram, M. A. (2015). Left-sided strokes are more often recognized than right-sided strokes: The Rotterdam Study. Stroke , 46 (1), 252–254. https://doi.org/10.1161/STROKEAHA.114.007385

Reinhart, S., Schindler, I., & Kerkhoff, G. (2011). Optokinetic stimulation affects word omissions but not stimulus-centered reading errors in paragraph reading in neglect dyslexia. Neuropsychologia, 49 (9), 2728–2735. https://doi.org/10.1016/j.neuropsychologia.2011.05.022

Rosenbek, J. C., Crucian, G. P., Leon, S. A., Hieber, B., Rodriguez, A. D., Holiway, B., Ketterson, T. U., Ciampitti, M., Heilman, K., & Gonzalez-Rothi, L. (2004). Novel treatments for expressive aprosodia: A phase I investigation of cognitive linguistic and imitative interventions. Journal of the International Neuropsychological Society , 10 (5), 786–793. https://doi.org/10.1017/S135561770410502X

Rosenbek, J. C., Rodriguez, A. D., Hieber, B., Leon, S. A., Crucian, G. P., Ketterson, T. U., Ciampitti, M., Singletary, F., Heilman, K. M., & Gonzalez Rothi, L. J. (2006). Effects of two treatments for aprosodia secondary to acquired brain injury. Journal of Rehabilitation Research & Development , 43 (3), 379–390.

Sheppard, S. M., Keator, L. M., Breining, B. L., Wright, A. E., Saxena, S., Tippett, D. C., & Hillis, A. E. (2020). Right hemisphere ventral stream for emotional prosody identification: Evidence from acute stroke. Neurology, 94 (10), e1013–e1020. https://doi.org/10.1212/WNL.0000000000008870

Siéroff, E. (2017). Acquired spatial dyslexia. Annals of Physical and Rehabilitation Medicine , 60 (3), 155–159. https://doi.org/10.1016/j.rehab.2015.07.004

Steward, K. A., & Kretzmer, T. (2022). Anosognosia in moderate-to-severe traumatic brain injury: A review of prevalence, clinical correlates, and diversity considerations. The Clinical Neuropsychologist, 36 (8), 2021–2040. https://doi.org/10.1080/13854046.2021.1967452

Stockbridge, M. D., Sheppard, S. M., Keator, L. M., Murray, L. L., Blake, M. L., & Right Hemisphere Disorders working group, Evidence-Based Clinical Research Committee, Academy of Neurological Communication Disorders and Sciences. (2022). Aprosodia subsequent to right hemisphere brain damage: A systematic review and meta-analysis. Journal of the International Neuropsychological Society , 28 (7), 709–735. https://doi.org/10.1017/S1355617721000825

Tompkins, C. A. (2012). Rehabilitation for cognitive-communication disorders in right hemisphere brain damage. Archives of Physical Medicine and Rehabilitation , 93 (Suppl. 1), S61–S69. https://doi.org/10.1016/j.apmr.2011.10.015

Tompkins, C. A., Fassbinder, W., Blake, M. L., Baumgaertner, A., & Jayaram, N. (2004). Inference generation during text comprehension by adults with right hemisphere brain damage: Activation failure versus multiple activation. Journal of Speech, Language, and Hearing Research , 47 (6), 1380–1395. https://doi.org/10.1044/1092-4388(2004/103)

Tompkins, C. A., Scharp, V. L., Meigh, K. M., & Fassbinder, W. (2008). Coarse coding and discourse comprehension in adults with right hemisphere brain damage. Aphasiology , 22 (2), 204–223. https://doi.org/10.1080/02687030601125019

Ukaegbe, O. C., Holt, B. E., Keator, L. M., Brownell, H., Blake, M. L., & Lundgren, K. (2022). Aprosodia following focal brain damage: What’s right and what’s left? American Journal of Speech-Language Pathology, 31 (5S), 2313–2328. https://doi.org/10.1044/2022_AJSLP-21-00302

World Health Organization. (2001). International Classification of Functioning, Disability and Health.

Wright, H. H., & Capilouto, G. J. (2012). Considering a multi-level approach to understanding maintenance of global coherence in adults with aphasia. Aphasiology , 26 (5), 656–672. https://doi.org/10.1080/02687038.2012.676855

About This Content

Acknowledgments.

Content for ASHA’s Practice Portal is developed through a comprehensive process that includes multiple rounds of subject-matter expert input and review. ASHA extends its gratitude to the following subject-matter experts who were involved in the development of the Right Hemisphere Disorder page:

• Christine R. Baron, MA, CCC-SLP
• Margaret L. Blake, PhD, CCC-SLP
• Perrine Ferré, MA
• Melissa Johnson, MA, CCC-SLP
• Yves Joanette, PhD
• Kristine M. Lundgren, ScD, CCC-SLP
• Jamila M. Minga, PhD, CCC-SLP
• Ilana F. Oliff, MA, CCC-SLP
• Amy D. Rodriguez, PhD, CCC-SLP
• Victoria L. Scharp, PhD, CCC-SLP
• Shannon Sheppard, PhD, CCC-SLP
• Melissa Stockbridge, PhD, CCC-SLP

Citing Practice Portal Pages

The recommended citation for this Practice Portal page is:

American Speech-Language-Hearing Association. (n.d.). Right hemisphere Disorder [Practice portal]. https://www.asha.org/Practice-Portal/Clinical-Topics/Right-Hemisphere-Disorder/

In This Section

• Practice Portal Home
• Clinical Topics
• Professional Issues
• Advertising Disclaimer
• Advertise with us

Evidence Maps

• ASHA Evidence Maps

Peer Connections

• Connect with your colleagues in the ASHA Community
• ASHA Special Interest Groups

ASHA Related Content

• Find related products in ASHA's Store
• Search for articles on ASHAWire
• ASHA Stream

Content Disclaimer: The Practice Portal, ASHA policy documents, and guidelines contain information for use in all settings; however, members must consider all applicable local, state and federal requirements when applying the information in their specific work setting.

ASHA Corporate Partners

• Become A Corporate Partner

The American Speech-Language-Hearing Association (ASHA) is the national professional, scientific, and credentialing association for 234,000 members, certificate holders, and affiliates who are audiologists; speech-language pathologists; speech, language, and hearing scientists; audiology and speech-language pathology assistants; and students.

• All ASHA Websites
• Work at ASHA
• Marketing Solutions

Information For

Get involved.

• ASHA Community
• Become a Mentor
• Become a Volunteer
• Special Interest Groups (SIGs)

Connect With ASHA

American Speech-Language-Hearing Association 2200 Research Blvd., Rockville, MD 20850 Members: 800-498-2071 Non-Member: 800-638-8255

MORE WAYS TO CONNECT

Media Resources

• Press Queries

Site Help | A–Z Topic Index | Privacy Statement | Terms of Use © 1997- American Speech-Language-Hearing Association

• Search Close search
• Find a journal
• Search calls for papers
• Journal Suggester
• Open access publishing

We’re here to help

Find guidance on Author Services

Open access

What are the functional outcomes of right hemisphere stroke patients with or without hemi-inattention complications? A critical narrative review and suggestions for further research

• Cite this article
• https://doi.org/10.3109/09638288.2015.1037865

Introduction

Main results and findings, conclusions.

• Acknowledgements
• Full Article
• Figures & data
• Reprints & Permissions
• View PDF PDF

Implications for Rehabilitation

Findings from this review demonstrate a paucity of evidence to support the presence of hemi-inattention as a key predictor of functional recovery in patients with right hemisphere stroke; as such, practitioners should take this into consideration when planning rehabilitation programmes of their patients.

In the initial months following right hemisphere stroke, there are wide-ranging differences in the rate and amount of functional recovery in patients, with and without hemi-inattention. Practitioners should not limit the aspirations of their patients based on the presence or absence of hemi-inattention.

This review has identified a number of measurement limitations in commonly employed assessment tools for hemi-inattention and overall functional recovery. As such, practitioners should take the limitations of specific measures into account when interpreting the results contextually and with respect to their patients’ situation.

• Functional outcomes
• hemi-inattention
• right hemisphere

Hemi-inattention (HI), commonly referred to as “neglect”, is a complex, heterogeneous and disabling condition which acutely affects up to 80% of patients with right hemisphere stroke (RHS) dysfunction [ Citation 1 , Citation 2 ]. Despite considerable research and advances in the field, HI remains poorly defined as a condition per se . This is supported by the use of multiple descriptors (e.g. unilateral neglect, unilateral inattention) and taxonomies in the literature [ Citation 3–5 ].

Clinically, HI is characterised by reduced attention and/or spatial awareness to details in the environment (commonly towards the left side of the body). HI can affect one or more functional domains (e.g. sensory-motor, visual-spatial) [ Citation 6 , Citation 7 ] and often co-exists with anosognosia and depression [ Citation 8 , Citation 9 ]. HI has been regarded as responsible for delayed and challenging rehabilitation, reduced safety awareness, poor functional outcomes, increase in dependency levels and risk of institution care [ Citation 10–12 ].

Historically, findings from published studies have reported disparity in functional ability scores; with patient groups affected by HI (HI+) underachieving compared to those without (HI−) [ Citation 11 , Citation 13 , Citation 14 ]. Traditionally the cause of this disparity has been largely attributed to the presence of HI, although findings from predictive models have been conflicting and inconclusive [ Citation 10 , Citation 13 , Citation 15–17 ]. This has led to considerable confusion and uncertainty about the clinical importance and significance of differences thought to be associated with HI [ Citation 9 , Citation 18–20 ]. The paucity of relevant evidenced-based reviews has not helped to clarify the predictive role of HI or to promote good rehabilitation practice.

The last systematic review was undertaken by Jehkonen et al. [ Citation 21 ]. The authors focused on the methodological quality of 26 studies published in 1996–2005, which evaluated the impact of Neglect on functional ability in predominantly generic stroke patient samples with mixed lesion sites. Jehkonen et al. [ Citation 21 ] highlighted as an issue considerable differences in patient samples and inconsistencies in results but nonetheless concluded that HI had a significant negative impact on functional outcome, either as an independent predictive factor or in the presence of other variables. Their findings corroborated those of earlier reviews [ Citation 2 , Citation 22 ] which were not specifically focused on the relationship between HI and functional ability. Jehkonen et al. [ Citation 21 ] recommended further research on homogeneous patient groups with respect to right/left hemispheric lesions to improve consistency in the results.

Given the paucity of research in this area, an in-depth evidence-based critical review of relevant studies is offered here with a different approach to that taken by Jehkonen et al. [ Citation 21 ]. However, considering the extent of methodological differences between studies [ Citation 21 ], a narrative review was appropriate. This enabled the inclusion of sufficient, relevant contemporary studies which would have otherwise been excluded by the more stringent selection criteria of a pure systematic review. Narrative reviews “lay out the most recent and best knowledge of various aspects of a problem” [ Citation 23 , p. 427], and are considered appropriate when a diversity of research methods are used in the studies considered as relevant (rather than focusing only on randomised controlled trials), where studies have used different outcome measures and/or non-equivalent samples [ Citation 24 ] and when studies are of relatively poor methodological quality [ Citation 25 ].

The current review examined traditional claims made by previous studies and reviews [ Citation 21 , Citation 22 ] about the negative impact of HI on function; more specifically the strength of the relationship between HI status and functional recovery following RHS. Another aim was to estimate the magnitude of functional differences between HI± patient groups. The current review extended the work carried out by Jehkonen et al. and used a more rigorous and systematic approach to the selection of studies and the review process. Consequently it included fewer ( n  = 12) but more homogenous studies with RHS patient samples. Theoretically similarly designed studies tend to be more comparable than heterogeneous stroke studies.

Both the discrepancy in HI± patient scores and the relationship between HI and functional recovery are of interest to rehabilitation professionals. Together with other indicators (e.g. stroke severity) they may be used to predict likely change in function with time since stroke. This knowledge can guide rehabilitation decisions, e.g. as to which patients are suitable for early supported home discharge schemes. Currently there is an urgent need for reliable predictors and indicators to support the transfer of in-patient rehabilitation services to appropriate stroke survivors in the community. The final aim was to formulate more robust research strategies based on the limitations of studies to date.

A literature search was conducted from 1995 to February 2015 of the databases MEDLINE, AMED, CINAHL, PsycINFO and COCHRANE using several descriptors of neglect subtypes in the literature including HI , spatial, visual, unilateral, personal, extra-personal, motor, sensory, hemi and representational. The words; stroke, CVA, functional* and activities of daily living (ADL) were added to the final search so that studies focused on specific functional activities were included. Children or young adults (≤18 years) and non-human samples were excluded.

The search yielded three Cochrane reviews and 195 publications; AMED (70), CINAHL (86), MEDLINE (102) and PsycINFO (57). In line with the aims of the review, and supported by recommendations from Jehkonen et al. [ Citation 21 ], only studies that compared the homogeneous patient groups with respect to hemispheric lesion site (RHS) and presenting comparisons of patients with or without HI were selected (including intervention designs); all other studies with the heterogeneous patient samples and no HI group comparison were excluded. In addition, functional ability had to be quantifiably measured so that the HI± group differences in scores could be calculated. Two reviewers read the abstracts and, when in doubt, the publication to determine relevance. This process led to the selection of 12 international studies.

Table 1. Critical evaluation of reviewed studies (abbreviations are defined in Appendix 2).

On the GRADE scale; A is high and assigned to well-performed Randomised controlled trials (RCTs) and observational studies with consistent results and/or strong effects. B is moderate – serious flaws in the design in which the estimated effect is likely to be considerably different than the true effect. C is low – studies with serious limitations in which the true effect is likely to be very different than the estimated, e.g. failure to include relevant confounding factors. D is very low – as in C but any estimated effect is very uncertain and highly unlikely to reflect the true effect.

Population studied and demographics

Geographically, studies were undertaken in Canada (1), Italy (3), Israel (2), UK (2), USA (3), USA & Italy (1) with local RHS populations. Two of the studies probably used the same population [ Citation 16 , Citation 27 ].

Age and gender were described consistently; study [ Citation 11 ] made reference to educational background and family burden. The age range varied from 57 (SD 10) [ Citation 10 ] to 60–69 [ Citation 11 , Citation 14 , Citation 16 , Citation 27 , Citation 28 ] and 70–76 years [ Citation 12 , Citation 15 , Citation 29–32 ]. In addition, there was considerable variation in age within specific studies, e.g. 33–88 years [ Citation 11 ] and 40–99 years [ Citation 29 ]. Gender tended to be equally distributed. Morbidity was documented in studies [ Citation 16 , Citation 29 ]; stroke was associated with hypertension, diabetes and heart disease. Stroke severity was not always made clear. It was reported as moderate in two studies [ Citation 15 , Citation 16 ], whilst study [ Citation 30 ] indicated that patients with severe stroke were excluded. Two studies [ Citation 11 , Citation 14 ] recruited only patients with (perceived) good rehabilitation potential. Stroke severity was unreported in seven studies [ Citation 10 , Citation 12 , Citation 27 , Citation 28 , Citation 29 , Citation 31 , Citation 32 ].

Definition of function and “Neglect/HI” syndrome

Conceptually, functional ability/outcome was rarely defined but inferred from ADL measurement scales, mainly the Barthel Index (BI) [ Citation 33 ] and Functional Instrumental Measure (FIM) [ Citation 34 ]. “Neglect/HI” tended to be traditionally defined as a failure to orient, report or respond to stimuli located on the opposite side to the site of the brain lesion which cannot be explained by either primary sensory or motor deficits [ Citation 35 ]. Different studies referred to HI sub-types interchangeably, e.g. visual neglect [ Citation 15 ], unilateral spatial neglect [ Citation 12 ] but effectively measured the same condition because the measurements used cannot differentiate between sub-types of neglect (e.g. visual, spatial and unilateral) [ Citation 3 , Citation 4 , Citation 36 ] but provide an overall measure of the degree or profundity of the condition.

Research settings

Research settings were insufficiently described to allow clear comparison between countries, e.g. termed as a rehabilitation facility or hospital in Israel and a stroke unit in England. They tended to be either acute in-patient hospital and/or community rehabilitation facilities which would suggest research on samples assessed at varying intervals after stroke onset.

Nine studies [ Citation 10 , Citation 14–16 , Citation 27 , Citation 28 , Citation 30–32 ] employed a prospective design, two studies [ Citation 12 , Citation 29 ] employed a retrospective design and one study [ Citation 11 ] employed both. Study [ Citation 11 ] employed a cross-sectional design; most studies [ Citation 5 , Citation 12 , Citation 16 , Citation 28–30 , Citation 32 ] employed a serial design characterised by variable baseline (T0) measures and one follow-up at discharge. Four studies [ Citation 10 , Citation 14 , Citation 27 , Citation 31 ] included up to three follow-up observations. The longest follow-up period was one year since stroke [ Citation 10 ]. All other follow-ups were not fixed in time but varied relative to the discharge point.

Selection criteria

Inclusion criteria tended to be vague; 10/12 studies [ Citation 10–12 , Citation 15 , Citation 16 , Citation 27–29 , Citation 31 , Citation 32 ] included only patients with “good rehabilitation potential” which was not clearly defined. However, by inference it would appear that severely cognitively impaired patients and those with common (age-related) morbidities (e.g. cardio-pulmonary) were automatically excluded early on from most of the studies.

Confirmation of stroke

Stroke was reportedly confirmed by a neurologist in all the studies and by radiological means in 42% [ Citation 10 , Citation 14 , Citation 16 , Citation 28 , Citation 31 ]. Stroke severity was measured in three studies [ Citation 15 , Citation 16 , Citation 27 ] but the scale score was only reported once in [ Citation 16 ] (using the Canadian Neurological Scale – CNS = 7). Aetiologically, infarct was predominant but the majority of studies also included haemorrhage.

Time to first (1st) observation

Baseline measures were taken at variable (non-comparable) times across the studies, making direct comparison difficult. Time to 1st observation was not reported in study [ Citation 29 ] and unclear in [ Citation 32 ]. In hospital settings, initial measurement varied from 7 to 15 days since stroke in studies [ Citation 12 , Citation 15 ], up to 30 days [ Citation 10 , Citation 28 , Citation 30 , Citation 31 ], up to 40 days [ Citation 14 , Citation 16 ] but occurred after 2–6 months in community rehabilitation facilities [ Citation 11 , Citation 27 ].

Sample size

Sample size and composition varied considerably; from 16 participants [ Citation 32 ] to 178 [ Citation 16 ]. Six studies [ Citation 11 , Citation 12 , Citation 15 , Citation 16 , Citation 29 , Citation 30 ] had more than 100 participants; three studies [ Citation 14 , Citation 16 , Citation 28 ] reported between 50 and 100; three studies [ Citation 10 , Citation 31 , Citation 32 ] had less than 50 participants. The proportion of HI+ to HI− patients also varied; 7/12 studies had less than 50% HI+ in the sample, the smallest being 19% [ Citation 10 ] and largest 60% [ Citation 29 ] (the latter being therefore more adequately powered to statistically detect differences between the HI± groups).

Attrition rates

Attrition rates of 1%, 16%, 7.8% and 11% were reported [ Citation 33 , Citation 14 , Citation 16 , Citation 31 ], respectively. Reasons for attrition were due to a combination of factors (incomplete documentation at discharge, loss to community follow-up and mortality).

Assessment of Neglect/HI

In regard to HI, both diagnostic tools and frequency of assessment varied; most studies [ Citation 11 , Citation 12 , Citation 14–16 , Citation 27–30 ] assessed only at baseline (which differed in time across studies), whilst three studies [ Citation 10 , Citation 31 , Citation 32 ] assessed patients at admission and discharge (which also differed in time since stroke). In half of the studies [ Citation 10 , Citation 14 , Citation 28 , Citation 30–32 ], HI was identified and assessed by a validated test battery – the Behaviour Inattention Test (BIT) [ Citation 37 ]. Single letter cancellation and line-bisection (pen and paper tasks) were used in two studies [ Citation 12 , Citation 15 ], respectively, whilst three used various standardised Neglect-specific tests [ Citation 11 , Citation 16 , Citation 27 ]. Study [ Citation 29 ] relied on mention of Neglect in the medical documents.

Other assessments

Functional ability was assessed by the FIM in all but four studies [ Citation 15 , Citation 16 , Citation 27 , Citation 31 ] which used different versions of the BI. Most studies measured or recorded additional factors which ranged from length of stay (LOS), discharge destination outcome, continence status, aspects of cognitive-motor function including perception, muscle strength, balance and tactile sensation. Validated measurement scales were generally used for these purposes.

Statistical data analysis

Data tended to be summarised by group (HI±) scores. Median or mean statistics were frequently reported with standard deviation (SD), and inter-quartiles to a lesser extent. Data distribution was rarely described but inferred from the summary statistic used. Rasch data transformation was undertaken by two studies [ Citation 29 , Citation 30 ] in an attempt to “normalise” a skewed data distribution. Estimation and management of missing data were not specifically reported, with one exception [ Citation 30 ].

Type of data analysis

The type and extent of data analysis varied substantially. For clarity, only general tendencies are described in this section; for specific details refer to Table 1 .

The majority of studies carried out preliminary tests for uni/bivariate associations (e.g. neglect × functional ability) and/or group (HI±) score comparison, e.g. [ Citation 11 , Citation 12 , Citation 30 ]. The correlation coefficients used were Spearman’s rho or Pearson’s r , whereas t -test, Mann–Whitney U and Chi square test were frequently used for group comparisons. In order to minimise type I error (i.e. a false statistically significant result), one study [ Citation 31 ] adjusted for multiple testing of the same participants over time by means of Bonferroni correction. Adjustment for small sample size was reported in two studies [ Citation 15 , Citation 32 ] but the adjustment method (Pillai’s trace) was only described once in study [ Citation 32 ].

Eight studies [ Citation 10–12 , Citation 15 , Citation 16 , Citation 28–30 ] used regression methods to evaluate various relationships between predictor or explanatory variables (IV’s) with one or more dependent variables (DVs), including functional ability. However, the type of model used was not clearly identified (predictive versus associative model). Therefore, it was difficult to assess the suitability of the models employed for the purpose of answering the question posed. For example, Paolucci et al’s study [ Citation 16 ] modelled the impact on later function of a large number and combination of IV’s (admission stroke severity score, gender, type of lesion, hypertension, diabetes, heart disease, unilateral spatial neglect, depression, epileptic seizures post-stroke, family support, education level, discharge destination in various combinations) but the extent of adjustment for confounding factors was variable – stroke severity was inconsistently adjusted for and no adjustment for differences in age was undertaken. Therefore, it was difficult to tease out specific relationships and infer cause from complex regression models. Differences in age, gender and duration of in-patient stay were adjusted in some models evaluated by studies [ Citation 11 , Citation 16 , Citation 30 ]. Furthermore, the rationale behind the choice, order of entry and measurement level (continuous/categorical) of IV’s was rarely stated (e.g. study [ Citation 10 ] used stepwise methods, whereas study [ Citation 16 ] used forward stepwise), which complicated understanding and interpretation of the results.

Three out of four studies with more than one follow-up point evaluated change in functional ability over time by means of ANOVA’s and/or multiple regression analysis [ Citation 10 , Citation 14 , Citation 27 ]; study [ Citation 10 ] specified repeated measures ANOVA. However, both methods have considerable limitations which potentially impact on study power and accuracy of results, especially in serial models with more than one follow-up and several repeated measures on the same individual. To this end, ANOVA requires complete data sets, which is problematic in stroke research due to the likelihood of missing data in the long-term. Ordinary single or multivariate regression analysis does not take into account correlation generated by multiple responses from the same individual on the same assessment measure/s (this violates the statistical assumption of independent observations in regression analysis [ Citation 38 , Citation 39 ]). Consequently, both the validity and accuracy of ordinary regression results are threatened including any inferences based on the results.

A substantial number of findings were reported across studies, only those pertaining to functional outcome and HI are summarised in this section (refer to Table 1 for details by study).

Disparity between the HI± group scores

All the studies found statistically significant disparities in average HI± group scores wherein the HI+ patients scored less that the HI− in overall functional ability and sensory-motor components on the BI, FIM and RMI (Rivermead Mobility Index) [ Citation 40 ], at discharge and up to one year post-stroke onset. Relatively less (statistically non-significant) disparity was found on cognitive FIM sub-scale scores in three studies [ Citation 14 , Citation 29 , Citation 30 ]. Nevertheless, the magnitude of differences reported across studies was considerably variable even when the same measurement scales were used at discharge. Study [ Citation 15 ] reported a difference between the HI± groups of 2 BI units (10%); whilst other studies reported differences of 7 BI units (35%) [ Citation 31 ], 10 FIM units (8%) [ Citation 14 ] and >30 FIM units (24%) [ Citation 30 ]. However, it must be pointed out that time to 1st observation and discharge point were also considerably variable across the studies (see later section), therefore it is difficult to extrapolate further from the findings to isolate specific influences of HI.

Progress rates

In general, similar rates of progress between the HI± groups were found prior to discharge but again these rates varied across studies even though the samples were homogenous with respect to lesion side (RHS). Four studies [ Citation 10 , Citation 14 , Citation 27 , Citation 31 ] followed up patients beyond discharge; study [ Citation 27 ] found that specific HI training improved functional ability of the HI+ group but gains were not maintained by the end of the study (estimated from highly variable published data; recorded about 6–10 months after stroke).

Length of in-patient stay and discharge destination outcome

The HI+ patient group tended to have longer LOS/days but this varied considerably across studies, e.g. HI+ 64, HI− 36 days [ Citation 15 ], HI+ 31, HI− 25 days [ Citation 12 ] and HI+ 79, HI− 52 days [ Citation 31 ]. On average, levels of community support and rates of institutional care were higher in the HI+ patients. However, entry of both the groups to institutional care was variable; ranging from 1/40 (2.5%) [ Citation 10 ], 32/178 (18%) [ Citation 16 ] and 6/28 (22%) [ Citation 31 ].

Modelling results

In regard to multiple-regression modelling results, six out of eight studies reported that HI significantly and negatively predicted outcome in a variety of models of differing complexity and specification [ Citation 10 , Citation 16 ]. However, without a clear approach to modelling, such as identifying the strategy (e.g. step-up/down) used to build the models, the technique used to implement that strategy (forward/backward step-wise) and the decision criteria (e.g. theory driven) used within the technique, it was difficult to follow the modelling process and know the true effect of individual variables (inclusive of HI) as evident from the following examples.

Whilst four studies [ Citation 10–12 , Citation 16 ] reported independent prediction of HI for the measured functional outcomes, they did not clarify what “independent” means (e.g. explains more than 10% of the DV). In contrast, two studies [ Citation 15 , Citation 29 ] did not find a significant relationship between HI status and functional ability in any of the models evaluated – using either FIM or BI (0–20 scale) scores as the DV and HI as a categorical or continuous predictor variable, respectively. Notably the measurement of HI was itself a variable, since different studies used different measurement levels and scales.

In regard to covariates, pre-stroke function was unrelated to functional ability at admission (T0) according to three studies [ Citation 15 , Citation 16 , Citation 30 ]. Baseline function was significantly, positively related to functional outcome at discharge according to three studies [ Citation 12 , Citation 28 , Citation 29 ]. Age was significantly, negatively related to change in functional ability according to studies [ Citation 28 , Citation 29 ] but unrelated to final functional ability in models evaluated by studies [ Citation 15 , Citation 30 ]. Cognitive ability was unrelated to ADL as measured by BI and FIM in studies [ Citation 15 , Citation 30 ], respectively, but positively related in study [ Citation 10 ].

Overall model fit

HI generally explained little of the final variance in functional ability (DV). Study [ Citation 30 ] reported the largest amount of variance explained in the DV (discharge FIM scores) as 49%; 44% of which were explained by admission FIM scores and a further 5% by BIT scores (HI levels). A key limitation of these analyses was that none of the studies undertook basic sensitivity analysis, e.g. how well the models met regression assumptions, such as normal distribution of residuals. In addition, confidence intervals (CIs) around regression coefficients or standard error (SE) sizes were rarely reported which complicated the interpretation of regression coefficients.

Overall quality

Taking everything into account, individual studies were graded as C/D on the GRADE scale. This reflected (but was not limited to) serious limitations in the design and data analysis methods described in previous sections. In particular, the exclusion of RHS patients with higher levels of stroke severity and HI, effectively limited representativeness of the samples and generalisation to the wider clinical population. There was a lack of adjustment of potential confounding factors (such as stroke severity, age and time since stroke) in statistical models estimating specific effects of HI on outcomes; and insufficient data/details to enable the reader to make informed decisions (e.g. omission of confidence intervals around regression coefficient estimates which enable accurate interpretation of the result and lack of sensitivity analysis to support validity and conclusions made from the findings).

The purpose of the review was threefold – to test traditional claims that neglect/HI has a deleterious impact on functional ability after stroke, to assess the extent of differences between RHS patient groups (with and without HI) and clarify the relationship between HI and functional recovery. In addition, suggestions for more rigorous research were formulated based on a critique of these studies.

Based on findings from the 12 reviewed studies, it is apparent that the presence of HI+ is linked to poor functional outcomes in RHS patients when compared to their counterparts without HI impairments (HI−). However, it was not possible to assess the extent of differences between the HI± groups at specific points in time (elapsed after stroke) because of considerable variation and inconsistency in design within and across studies – not just in assessment measures but also in sample mix (age, stroke severity); time to 1st observation as well as follow-up assessments; and methods of data analysis. Neither was it possible to draw firm conclusions on the relationship between HI status and change in basic functional ability over time. In other words, the specific contribution of HI to the functional ability and outcome of RHS patients cannot be known with certainty from the results of the reviewed studies. It is possible that other influential factors, such as stroke severity, age and time since stroke substantially, explain the discrepancy in the HI± patient group scores reported in the literature. To this end, the implications of poor methodology in the field of stroke and HI were also highlighted in previous reports [ Citation 2 , Citation 21 , Citation 22 ]. There is also recent discussion [ Citation 38 ] of the importance of optimising data analysis methods in the field of neglect/HI to enhance the accuracy of predictive models.

Aside from clear methodological limitations, such as exclusion of patients with severe stroke and marked HI+ at baseline, it is likely that unmeasured patient factors also influenced the results and are responsible for a proportion of unexplained variance in functional ability post-stroke. These include pre-stroke educational levels and personality characteristics which would be expected to vary across patients. Furthermore, seven studies [ Citation 12 , Citation 15 , Citation 16 , Citation 28–30 , Citation 32 ] used discharge as the only follow-up point. Although discharge is a key point of change in the stroke recovery pathway, it is subject to natural variation in stroke recovery rate and initial severity. Furthermore, the optimal discharge point may be influenced by differing cultural expectations as to when the recovering patient is ready to be discharged to community care and where this will be undertaken, i.e. home or institution. From the description available in past reviewed studies, it would appear that contextual features, such as stroke unit versus acute general rehabilitation hospital, were not sufficiently comparable in care and rehabilitation provision across different countries especially at the time that they were undertaken [ Citation 10 , Citation 12 , Citation 14–16 ]. This lack of comparability in context of care is in itself a source of important variation when evaluating functional outcomes from different countries. In relation to the issues associated with discharge as an informative follow-up point, these can be minimised by including at least one standardised fixed follow-up point for all the patients, e.g. baseline – discharge – follow-up 6 months after stroke.

Sample representation and size

Although all patients in the studies selected had RHS, there were considerable sample differences in age and recruitment settings which were not directly comparable. Furthermore, time to 1st observation and patient selection criteria were substantially different; and patients with severe cognitive impairment, and probably severe neglect/HI+ levels, tended to be automatically excluded at the recruitment phase. This strongly suggests that the samples were not sufficiently representative of the RHS population which limits the generalisation of any findings. It is imperative that future studies are conducted with representative (severity-inclusive) samples especially in predictive/explanatory relationship studies. This can be achieved by augmenting the design to suit specific requirements of severely cognitively impaired patients and then applying advanced methods of data analysis (e.g. multi-level modelling – MLM). Specific advantages of MLM over traditional methods of regression analysis and ANOVA are highlighted later on.

The importance of a large enough sample size in relation to study power and reducing type I error cannot be overemphasised. For instance, two studies [ Citation 10 , Citation 27 ] followed up patients for longer periods (up to one year after stroke) but with relatively small starting samples (40 and 59, respectively). Although attrition rate was not reported, this can be substantial at one year post-stroke which further reduces study power to detect differences between the HI± groups. Therefore, future studies should ensure sufficient initial sample size, particularly in longer term follow-up designs (lasting more than 3 months after stroke) to account for possible attrition; an allowance of 15–20% is cited in the literature [ Citation 41 , Citation 42 ].

Measurement

The accuracy of results depended heavily on the precision of instruments and consistency in measurement. Only 50% of studies followed evidence-based recommendations in favour of diagnosing neglect/HI with validated test batteries rather than single cancellation task/s; the rationale [ Citation 35 , Citation 43 , Citation 44 ] being that several tests are more likely to detect neglect than a single test. To this end, the proportion of the HI+ patients detected in studies that assessed with the BIT is considerably higher than in those that used single tests. Like other test batteries, the BIT has measurement limitations in that it is restricted to assessment of HI in activities conducted within close proximity (arm’s length) of surrounding body space and possibly mental representation of objects. In addition, similar to other test batteries (e.g. Chaterine Bergago scale [ Citation 45 ]), the BIT cannot distinguish between different sub-types of the disorder [ Citation 3 , Citation 36 ] because all the assessed tasks inherently draw on visual, sensory, motor, spatial, perceptual and cognitive function to varying degrees. These components are intricately inter-twined and dependent on one another which make it virtually impossible to determine the relative contribution of individual components to the results or behaviour observed. Consequently (and of relevance to clinicians and researchers), it is not possible to accurately explain test results in terms of specific sub-types of HI in routine rehabilitation settings. That being said, assessment batteries provide an overall score of HI severity and are indicative of the type of perceptual, cognitive and/or motor impairments present.

Pending the development of more comprehensive and practical assessment tools, future researchers and clinicians should follow evidence-based assessment guidelines, i.e. use of validated assessment batteries for HI. Interpretation of results will also be enhanced by reporting the proportion of patients with mild compared to severe neglect/HI in the sample by categorising the HI+ patients at baseline. This is possible with the BIT [ Citation 36 , Citation 46 ].

Functional ability in the reviewed studies was assessed with at least two versions of the BI (using scales of 0–20 and 0–100) with differing sensitivity and precision [ Citation 47 , Citation 48 ], making comparison difficult. Furthermore, neither version assesses social/communicative components. They are therefore possibly more limited than the FIM in this respect [ Citation 46 ]. In comparison, the Extended Barthel Index version has been used in several stroke studies [ Citation 49 , Citation 50 ]. It is a validated measure and overcomes some of the limitations of other BI versions in that it includes assessment of social and communicative components and accounts for time taken to complete ADL tasks.

Ideally future studies should measure functional ability with comprehensive evidence-based measures; however it must be acknowledged that current choices are limited not only by content in terms of validity but also reliability over time and applicability across acute and community settings.

The position, interval and frequency of observations

In relation to design, the position, interval and frequency of observations is of critical importance in recording important changes in ability at the points when they are likely to occur. Furthermore, the amount and quality of information collected partly determines the type and extent of data analysis that is possible in order to answer specific research questions [ Citation 39 , Citation 51 , Citation 52 ]. This ability was greatly compromised within and across the reviewed studies by the variability in time around the initial measurement point and follow-up, the number and frequency of observations made over time, and the duration of individual studies.

In some studies, the extent of variation around the mean LOS blurred the beginning and end of the study (e.g. study [ Citation 16 ] reported mean LOS (HI+) 117 ± 61 days versus (HI−) 81 ± 38, and study [ Citation 29 ] reported (HI+) 29 versus (HI−) 22, range 3–75 days). This would suggest considerable differences in LOS and patient exposure to care within and across studies which have implications for assessing recovery.

At baseline, the total variation in initial measurement points across studies was 7 days to 6 months post-stroke in studies [ Citation 27 , Citation 32 ], respectively. During this period, both spontaneous and rehabilitation-driven recovery processes are known to actively contribute to outcome [ Citation 53–55 ]. Based on known recovery patterns, the impact of spontaneous recovery processes is expected to confound rehabilitative processes in at least 11/12 studies. Future studies should aim to collect initial data as early as is pragmatically possible, ideally within the 1st week of stroke. This would potentially streamline data collection procedures, enhance comparison of results across studies and minimise the effect of spontaneous recovery processes on outcome so that cause (e.g. influence of HI) can be more confidently inferred from regression results.

Future studies could also consider fixing the last follow-up point whilst still recording HI and functional change at discharge. Better still would be to statistically adjust for the confounding effect of time elapsed post-stroke. An optional growth curve modelling (MLM) approach could be used to model change over time in time-variant factors – this approach has been applied in a predictive stroke patient study [ Citation 56 ]. The advantages afforded by MLM in relation to analysis of stroke data are next summarised in this review.

Multi-level modelling approaches to data analysis

The advantages of multi-level modelling over traditional regression methods of analysis should not be overlooked when modelling dependency in the data due to multiple measurements from the same patient over time and interdependency between potential explanatory factors associated with functional ability (e.g. cognitive-motor processes). MLM can also handle data missing at random and unequal interval observations – all of which were potentially problematic features of the reviewed studies. As a result, MLM regression estimates are highly accurate and relatively stable compared to estimates obtained by ordinary multivariate regression analysis [ Citation 51 , Citation 52 ]. MLM has other useful features, such as estimation of unexplained variance within and across patients over time (in serial studies). It is imperative that future studies consider the advantages afforded by novel advanced statistical techniques in serial designs, which provide rich information but need to be appropriately statistically analysed for optimal results [ Citation 38 , Citation 57 , Citation 58 ].

From the discussion so far it can be deduced that regression estimates in 8/12 studies which modelled outcome are likely to be overinflated with under-estimated standard errors. In turn, this increases the risk of type 1 error which has adverse implications for interpretation of results, not least the predictive importance of HI on functional ability.

Confounding factors in stroke functional outcome studies

Another issue complicating the inference of causality from regression results is the lack of statistical adjustment for the established confounding effect of stroke severity and to a lesser extent chronological age. Besides the natural variation of both factors in different samples, stroke severity and age are both associated with HI [ Citation 21 , Citation 59 , Citation 60 ]. Therefore, it is important that these particular variables are modelled in studies exploring multiple predictor variables and functional outcome [ Citation 39 , Citation 61 ]. As already stated, the use of MLM can greatly help with teasing out complicated relationship dynamics. By the same token, the need for an adequately powered study is emphasised. Considering that only six studies started with approximately 100 patients and some cases [ Citation 10 , Citation 27 ] also analysed predictive multivariate models, it is likely that respective studies were underpowered to detect true effects. This increases the uncertainty of the modelled results and detracts from the validity of their findings.

Strengths and limitations of the review

The current review extended the work carried out by Jehkonen et al. by reviewing a homogeneous (versus mixed) patient sample with respect to hemispheric lesion (RHS). Further, a relatively more rigorous and systematic approach to study selection and the overall review process was undertaken.

The current review focused on group comparisons of functional outcomes of RHS patients with and without HI and the relationship of HI status with functional change with time since stroke, whereas Jehkonen et al. undertook a generalised review of the methodological issues from a wider range of studies which did not always include patient comparisons in the design (HI±). Group comparisons allow for the calculation of mean differences between patient groups and estimation of the modelled relationship between HI (group) status and functional outcome. This resulted in new insights into the data, such as the lack of adjustment for established confounding factors in past studies (e.g. stroke severity, time since stroke and age), leading to a different conclusion, i.e. no inferences could be made from the data available on either the relationship between neglect and functional outcome or the magnitude of difference in measureable scores between the patients with and without HI due to considerable heterogeneity in design and methods used to statistically analyse the data. Currently there is an urgent need for valid predictors and indicators to support the transfer of in-patient stroke rehabilitation services to appropriate stroke survivors living in the community.

Furthermore, due to the structure and layout of the review by Jehkonen et al., it is not possible to tell which studies found what and when in terms of results and time since stroke. The current review is more detailed in this respect. It is also more systematic and transparent both in the selection of studies and their evaluation (a checklist was used to ensure parity and consistency during the review process). The methodological quality of each study was separately graded. All these factors contribute to the rigour of the review and the validity of the findings.

The main limitation of the review is the possibility that relevant studies may have been missed due to indexing and multiple terms used to describe the “neglect” syndrome although a thorough search of the literature has been conducted several times. In addition, only English language publications or translations were considered.

The review does not address HI conditions arising from specific cortical or sub-cortical structures or anterior/posterior circulation division or left hemisphere lesions. Although HI is associated with different brain areas, HI emanating from the right hemisphere is commonly encountered in rehabilitation settings and challenging to treat [ Citation 4–7 , Citation 9 ]. In addition, the interpretation of results from assessments of HI in the left hemisphere lesions is confounded by speech and language impairments which frequently accompany LHS conditions. Non-language based clinical assessment for HI is urgently needed for this purpose. There is a lack of comparative group studies specifically on HI/neglect emanating from sub-cortical structures [ Citation 63 , Citation 64 ] probably because this is harder to identify in routine clinical settings and possibly less commonly encountered.

To conclude, based on findings from this review, the evidence for an important relationship between HI status and functional outcome in RHS patients (with or without consideration of other influential factors) cannot be substantiated from the data available. This is due to significant methodological limitations of published studies highlighted in this review and the relatively few studies published in this field in the last decade [ Citation 29–32 ]. The evidence that differences exist between the HI± patient groups on discharge and in the early months that follow tends to be better substantiated and is of interest to rehabilitation practitioners. However, on its own, this knowledge does not advance practice in the field.

In light of findings from this review, it is recommended that rehabilitation practitioners not only take into account the severity of HI in decision making but also recognise that evidence is weak for it being an independent marker of future functional recovery. Given the extent of differential progress possible in patients with RHS with and without HI, it is recommended that rehabilitation practitioners consider both the rate and amount of functional recovery over time. This is likely to be more informative overall than isolated measurements of HI soon after stroke. In relation to measurement, it is recommended that rehabilitation professionals sufficiently consider the psychometric limitations of commonly employed assessments of HI and overall functional ability (e.g. the BIT, BI and FIM) when interpreting the results.

Further research is warranted to identify the magnitude of the differences within and between the patient HI± groups and importantly the precise relationship between HI status and functional recovery with time since stroke. For this reason, well-designed, longitudinal, serial studies on large RHS patient samples are required which can withstand the scrutiny of future reviews on the subject.

It is recommended that future rehabilitation research in the field takes into account important methodological features to improve the quality and robustness of their findings. These include but are not limited to adequate sample size to detect differences in groups of patients with and without HI, with a full spectrum of stroke severity and HI, sufficient amounts of data by means of standardised methods repeated over time rather than isolated data points, and with sufficient follow-up to allow functional recovery to occur over time since stroke.

Declaration of interest

This work received no external funding. The authors report no conflicts of interest.

• Stone SP, Patel P, Greenwood RJ. The incidence of neglect phenomena and related disorders in patients with an acute right or left hemisphere stroke. Age Ageing 1993;22:46–52   PubMed Web of Science ® Google Scholar
• Bowen A, McKenna K, Tallis CT. Reasons for variability in the reported rate of occurrence of unilateral spatial neglect after stroke. Stroke 1999;30:1196–202   PubMed Web of Science ® Google Scholar
• Plummer P, Morris ME, Dunai J. Assessment of unilateral neglect. Phys Ther 2003;83:732–40   PubMed Web of Science ® Google Scholar
• Singh-Curry V, Husain M. Hemispatial neglect: approaches to rehabilitation. Clin Rehabil 2010;24:675–84   PubMed Web of Science ® Google Scholar
• Karnath H, Rorden C. The anatomy of spatial neglect. Neuropsychologia 2012;50:1010–17   PubMed Web of Science ® Google Scholar
• Mark VW. Acute versus chronic functional aspects of unilateral spatial neglect. Front Biosci 2003;8:172–89   Web of Science ® Google Scholar
• Parton A, Malhotra P, Husain M. Hemispatial neglect. J Neurol Neurosur Ps 2004;75:13–21   PubMed Web of Science ® Google Scholar
• Appelros P, Karlssonb GM, Hennerdal S. Anosognosia versus unilateral neglect. Coexistence and their relations to age, stroke severity, lesion site and cognition. Eur J Neurol 2007;14:54–9   PubMed Web of Science ® Google Scholar
• Kortte K, Hillis AE. Recent advances in the understanding of neglect and anosognosia following right hemisphere stroke. Curr Neurol Neurosci Rep 2009;9:459–65   PubMed Web of Science ® Google Scholar
• Katz N, Hartman-Maeir A, Ring H, et al. Functional disability and rehabilitation outcome in right hemisphere damaged patients with and without unilateral spatial neglect. Arch Phys Med Rehabil 1999;80:379–84   PubMed Web of Science ® Google Scholar
• Buxbaum LJ, Ferraro MK, Veramonti T, et al. Hemispatial neglect: subtypes, neuroanatomy, and disability. Neurology 2004;62:749–56   PubMed Web of Science ® Google Scholar
• Gillen R, Howard Tennen H, Tara McKee T. Unilateral spatial neglect: relation to rehabilitation outcomes in patients with right hemisphere stroke. Arch Phys Med Rehabil 2005;86:763–7   PubMed Web of Science ® Google Scholar
• Pedersen PM, Jørgensen HS, Hirofumi N, et al. Hemineglect in acute stroke – incidence and prognostic implications: the Copenhagen Stroke Study1. Am J Phy Med Rehab 1997;76:122–7   PubMed Web of Science ® Google Scholar
• Cherney L, Halper AS, Kwasnica CM, et al. Recovery of functional status after right hemisphere stroke: relationship with unilateral neglect. Arch Phys Med Rehabil 2001;82:322–8   PubMed Web of Science ® Google Scholar
• Kalra L, Perez I, Gupta S, et al. The Influence of visual neglect on stroke rehabilitation. Stroke 1997;28:1386–91   PubMed Web of Science ® Google Scholar
• Paolucci S, Antonucci G, Grasso MG, et al. The role of unilateral spatial neglect in rehabilitation of right brain–damaged ischemic stroke patients: a matched comparison. Arch Phys Med Rehabil 2001;82:743–9   PubMed Web of Science ® Google Scholar
• Nijboer T, Van de Port I, Schepers V, et al. Predicting functional outcome after stroke: the influence of neglect on basic activities in daily living. Front Hum Neurosci 2013;7:1–6   PubMed Web of Science ® Google Scholar
• Plummer P, Morris ME, Hurworth RE, et al. Characterisation of unilateral neglect by physiotherapists. Disabil Rehabil 2006;28:571–7   PubMed Web of Science ® Google Scholar
• Barrett AM, Buxbaum LJ, Coslett HB, et al. Cognitive rehabilitation interventions for neglect and related disorders: moving from bench to bedside in stroke patients. J Cogn Neurosci 2006;18:1223–36   PubMed Web of Science ® Google Scholar
• Menon-Nair A, Korner-Bitensky N, Wood-Dauphinee S. Assessment of unilateral spatial neglect post stroke in Canadian acute care hospitals: are we neglecting neglect? Clin Rehabil 2006;20:623–34   PubMed Web of Science ® Google Scholar
• Jehkonen M, Laihosalo M, Kettunen JE. Impact of neglect on functional outcome after stroke – a review of methodological issues and recent research findings. Restor Neurol Neurosci 2006;24:209–15   PubMed Web of Science ® Google Scholar
• Lincoln N, Bowen A. The need for randomised treatment studies in neglect research. Restor Neurol Neuros 2006;24:401–8   PubMed Web of Science ® Google Scholar
• Dijkers M. Task force on systematic reviews and guidelines the value of traditional reviews in the era of systematic reviewing. Am J Phys Med Rehabil 2009;88:423–30   PubMed Web of Science ® Google Scholar
• Pai M, McCulloch M, Gorman J, et al. Systematic reviews and meta-analyses: an illustrated, step-by-step guide. Natl Med J India 2004;17:86–95   PubMed Web of Science ® Google Scholar
• Nicolson A, Moir L, Millsteed J. Impact of assistive technology on family caregivers of children with physical disabilities: a systematic review. Disabil Rehabil Assist Technol 2012;7:345–9   PubMed Google Scholar
• Guyatt G, Oxman AD, Vist GE. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336:924–6   PubMed Web of Science ® Google Scholar
• Paolucci S, Antonucci G, Guariglia C, et al. Facilitatory effect of neglect rehabilitation on the recovery of left hemiplegic stroke patients: a cross-over study. J Neurol 1996;243:308–14   PubMed Web of Science ® Google Scholar
• Ring H, Feder M, Schwartz J, et al. Functional measures of first-stroke rehabilitation inpatients: usefulness of the Functional Independence Measure total score with a clinical rationale. Arch Phys Med Rehabil 1997;78:630–5   PubMed Web of Science ® Google Scholar
• Odell K, Wollack J, Flynn M. Functional outcomes in patients with right hemisphere brain damage. Aphasiology 2005;19:807–30   Web of Science ® Google Scholar
• Di Monaco M, Schintu S, Dotta M, et al. Severity of unilateral spatial neglect Is an independent predictor of functional outcome after acute inpatient rehabilitation in individuals with right hemispheric stroke. Arch Phys Med Rehabil 2011;92:1250–6   PubMed Web of Science ® Google Scholar
• Stein MS, Maskill D, Marston L. Impact of visual-spatial neglect of stroke functional outcomes, discharge destination and maintenance of improvement post-discharge. BJOT 2009;72:219–25   Web of Science ® Google Scholar
• Timbeck R, Spaulding SJ, Klinger L, et al. The effect of visuospatial neglect on functional outcome and discharge destination: an exploratory study. Phys Occup Ther Geriatr 2013;31:37–46   Google Scholar
• Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Maryland State Med J 1965;14:61–5   PubMed Google Scholar
• Granger CV, Cotter AC, Hamilton BB, et al. Functional Assessment Scales: a study of persons after stroke. Arch Phys Med Rehabil 1993;74:133–8   PubMed Web of Science ® Google Scholar
• Heilman KM, Valenstein E, Watson RT. The neglect syndrome. In: Vinken PJ, Bruyn GW, Klawans HL, eds. Handbook of clinical neurology 45 . Amsterdam: Elsevier Science Publishers; 1985:152–83   Google Scholar
• Menon A, Korner-Bitensky N. Evaluating unilateral spatial neglect post stroke: working your way through the Maze of Assessment Choices. Top Stroke Rehabil 2004;11:41–66   PubMed Google Scholar
• Wilson B, Cockburn J, Halligan PW. Behavioural inattention test, manual . Titchfield, Hampshire: Thames Valley Test Company; 1987   Google Scholar
• Goedert K, Boston R, Barrett A. Advancing the science of spatial neglect rehabilitation: an improved statistical approach with mixed linear modelling. Front Hum Neurosci 2013;7:1–15   PubMed Web of Science ® Google Scholar
• Shmueli G. To explain or to predict? Stat Sci 2010;25:289–310   Web of Science ® Google Scholar
• Collen C, Wade DT, Robb DT, et al. The Rivermead Mobility Index: a further development of the Rivermead Mobility Assessment. Int Disabil Stud 1991;13:50–4   PubMed Google Scholar
• Snijders AB, Bosker RJ. Standard errors and sample sizes for two-level research. J Educ Stat 1993;18:237–59   Google Scholar
• Walker J, Almond P. Interpreting statistical findings: a guide for health professionals and students . Berkshire: Open University Press; 2010   Google Scholar
• Halligan P, Marshall JC, Wade DT. Visuospatial neglect: underlying factors and test sensitivity. Lancet 1989;ii:908–11   Google Scholar
• Lopes M, Ferreira HP, Carvalho JC, et al. Screening tests are not enough to detect hemineglect. Arq Neuropsiquiatr 2007;65:1192–5   PubMed Web of Science ® Google Scholar
• Azouvi P, Samuel C, Louis-Dreyfus A, et al. Sensitivity of clinical and behavioural tests of spatial neglect after right hemisphere stroke. J Neurol Neurosurg Psy 2002;73:160–6   PubMed Web of Science ® Google Scholar
• Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF Activity. Disabil Rehabil 2005;27:315–40   PubMed Web of Science ® Google Scholar
• McDowell I. Measuring health: a guide to rating scales and questionnaires . 3rd ed. New York: Oxford University Press; 2006   Google Scholar
• Quinn TJ, Langhorne P, Stott DJ. Barthel index for stroke trials: development, properties, and application. Stroke 2011;42:1146–51   PubMed Web of Science ® Google Scholar
• Geshwinder H, Rettkeb H, Van den Heuvelc W, et al. Rehabilitation in acute stroke patients in German-speaking Switzerland. Swiss Med Wkly 2007;137:205–11   PubMed Web of Science ® Google Scholar
• Schuster C, Butler J, Andrews D, et al. Comparison of embedded and added motor imagery training in patients after stroke: results of a randomised controlled pilot trial. Trials 2012;13:11 (1–19)   Google Scholar
• Cheng J, Edwards LJ, Maldonado-Molina MM, et al. Real longitudinal data analysis for real people: building a good enough mixed model. Statist Med 2010;29:504–20   PubMed Web of Science ® Google Scholar
• Snijders AB, Bosker RJ. Multilevel analysis: an introduction to basic and advanced multilevel modeling . 2nd ed. London: Sage Publishers; 2012   Google Scholar
• Grefkes C, Fink GR. Reorganization of cerebral networks after stroke: new insights from neuroimaging with connectivity approaches. Brain 2011;134:1264–76   PubMed Web of Science ® Google Scholar
• Kwakkel G, Kollen BJ. Predicting activities after stroke: what is clinically relevant? Int J Stroke 2013;8:25–32   PubMed Web of Science ® Google Scholar
• Rehme AK, Grefkes C. Cerebral network disorders after stroke: evidence from imaging-based connectivity analyses of active and resting brain states in humans. J Physiol 2013;591:17–31   PubMed Web of Science ® Google Scholar
• Tilling K, Sterne J, Wolfe C. Multilevel growth curve models with covariate effects: application to recovery after stroke. Statist Med 2001;20:685–704   PubMed Web of Science ® Google Scholar
• Royston P, Moons K, Altman DG, et al. Prognosis and prognostic research: developing a prognostic model. BMJ 2009;338:1373–7   Web of Science ® Google Scholar
• Kwakkel G, Kollen B, Twisk J. Impact of time on improvement of outcome after stroke. Stroke 2006;37:2348–53   PubMed Web of Science ® Google Scholar
• Appelros P, Gunnel M, Karlsson A, et al. Prognosis for patients with neglect and anosognosia special reference to cognitive impairment. J Rehabil Med 2003;35:254–8   PubMed Web of Science ® Google Scholar
• Appelros P, Karlsson GM, Hennerdal S. Anosognosia versus unilateral neglect. Coexistence and their relations to age, stroke severity, lesion site and cognition. Eur J Neurol 2007;14:54–9   PubMed Web of Science ® Google Scholar
• Twisk JW. Applied multilevel analysis: a practical guide . New York: Cambridge University Press; 2006   Google Scholar
• Critical Appraisal Skills Programme. Available from: http://www.casp-uk.net/# !casp-tools-checklists/c18f8   Google Scholar
• Punt TD, Riddoch MJ. Motor neglect: implications for movement and rehabilitation following stroke. Disabil Rehabil 2006;28:857–64   PubMed Web of Science ® Google Scholar
• Sampanis DS, Riddoch J. Motor neglect and future directions for research. Front Hum Neurosci 2013;7:1–2   PubMed Web of Science ® Google Scholar

Appendix 1.

Table a1. critical evaluation checklist., appendix 2., table a2. abbreviated terms..

• Back to Top

Related research

People also read lists articles that other readers of this article have read.

Recommended articles lists articles that we recommend and is powered by our AI driven recommendation engine.

Cited by lists all citing articles based on Crossref citations. Articles with the Crossref icon will open in a new tab.

• People also read
• Recommended articles

To cite this article:

Download citation, your download is now in progress and you may close this window.

• Choose new content alerts to be informed about new research of interest to you
• Easy remote access to your institution's subscriptions on any device, from any location
• Save your searches and schedule alerts to send you new results
• Export your search results into a .csv file to support your research

Login or register to access this feature

Register now or learn more

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• My Bibliography
• Collections
• Citation manager

Save citation to file

Email citation, add to collections.

• Create a new collection
• Add to an existing collection

Add to My Bibliography

Your saved search, create a file for external citation management software, your rss feed.

• Search in PubMed
• Search in NLM Catalog
• Add to Search

Impact of unilateral stroke on right hemisphere superiority in executive control

Affiliations.

• 1 Department of Psychology, Queens College, The City University of New York, Queens, NY, USA; Department of Psychology, The Graduate Center, The City University of New York, New York, NY, USA; Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
• 2 Department of Psychology, Queens College, The City University of New York, Queens, NY, USA.
• 3 Department of Psychology, Columbia University in the City of New York, New York, NY, USA.
• 4 Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
• 5 Department of Psychology, Queens College, The City University of New York, Queens, NY, USA. Electronic address: [email protected].
• PMID: 33238172
• DOI: 10.1016/j.neuropsychologia.2020.107693

In our previous study, we have demonstrated a right hemisphere superiority in executive control of attention, with the right hemisphere being more efficient in dealing with conflict for stimuli presented in the left visual field. However, the unique and synergetic contribution of the two hemispheres to this superiority effect is still elusive. Here, using the lateralized attention network test, we compared the flanker conflict effect for stimuli presented in the left and right visual fields in patients with an ischemic stroke in the right or left hemisphere as the unilateral lesion groups and in patients with a transient ischemic attack without an acute infarction as the control group. In contrast to the transient ischemic attack group, which demonstrated a right hemisphere superiority in conflict processing, there was no evidence for such an effect in both unilateral stroke groups. These results can be explained by our model proposing that there is bilateral hemispheric involvement for conflict processing for information received from the left visual field and unilateral hemispheric involvement for conflict processing for information received from the right visual field, resulting in more efficient processing for the left visual field, i.e., the right hemisphere superiority effect. When there is damage to either hemisphere, the responsibility of conflict processing will largely fall on the intact hemisphere, eliminating the right hemisphere superiority effect.

Keywords: Executive control; Lateralization; Right hemisphere superiority; Unilateral stroke.

Copyright © 2020 Elsevier Ltd. All rights reserved.

PubMed Disclaimer

Similar articles

• Modulation of hemispheric asymmetry in executive control of attention in schizophrenia with atypical antipsychotic treatment: Potential benefits of olanzapine. Zhao Y, Li Y, Du J, Fang C, Li W, Lv M, Wu Y, Wang K, Wu T, Tian Y, Zhang J. Zhao Y, et al. Schizophr Res Cogn. 2024 Mar 6;36:100306. doi: 10.1016/j.scog.2024.100306. eCollection 2024 Jun. Schizophr Res Cogn. 2024. PMID: 38469136 Free PMC article.
• Right hemisphere superiority for executive control of attention. Spagna A, Kim TH, Wu T, Fan J. Spagna A, et al. Cortex. 2020 Jan;122:263-276. doi: 10.1016/j.cortex.2018.12.012. Epub 2018 Dec 29. Cortex. 2020. PMID: 30661735
• Visual neglect after left-hemispheric lesions: a voxel-based lesion-symptom mapping study in 121 acute stroke patients. Beume LA, Martin M, Kaller CP, Klöppel S, Schmidt CS, Urbach H, Egger K, Rijntjes M, Weiller C, Umarova RM. Beume LA, et al. Exp Brain Res. 2017 Jan;235(1):83-95. doi: 10.1007/s00221-016-4771-9. Epub 2016 Sep 16. Exp Brain Res. 2017. PMID: 27637595
• [Lateralization and interhemispheric integration of directed attention]. Ishiai S. Ishiai S. Rinsho Shinkeigaku. 2001 Dec;41(12):1128-30. Rinsho Shinkeigaku. 2001. PMID: 12235816 Review. Japanese.
• Consciousness wanted, attention found: Reasons for the advantage of the left visual field in identifying T2 among rapidly presented series. Verleger R, Śmigasiewicz K. Verleger R, et al. Conscious Cogn. 2015 Sep;35:260-73. doi: 10.1016/j.concog.2015.02.013. Epub 2015 Mar 13. Conscious Cogn. 2015. PMID: 25777355 Review.
• Altered hemispheric asymmetry of attentional networks in patients with pituitary adenoma: an event-related potential study. Wang S, Fu Z, Sun Y, Zhang M, Chen A, Cao C, Song J. Wang S, et al. Front Neurosci. 2023 May 11;17:1198409. doi: 10.3389/fnins.2023.1198409. eCollection 2023. Front Neurosci. 2023. PMID: 37250398 Free PMC article.
• Search in MeSH

LinkOut - more resources

Full text sources.

• Elsevier Science

Other Literature Sources

• scite Smart Citations
• MedlinePlus Health Information
• Citation Manager

NCBI Literature Resources

MeSH PMC Bookshelf Disclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

Log in using your username and password

• Search More Search for this keyword Advanced search
• Latest content
• Global health
• BMJ Journals

You are here

• Volume 13, Issue 8
• Clinical course of a 66-year-old man with an acute ischaemic stroke in the setting of a COVID-19 infection
• Article Text
• Article info
• Citation Tools
• Rapid Responses
• Article metrics

• http://orcid.org/0000-0002-7441-6952 Saajan Basi 1 , 2 ,
• Mohammad Hamdan 1 and
• Shuja Punekar 1
• 1 Department of Stroke and Acute Medicine , King's Mill Hospital , Sutton-in-Ashfield , UK
• 2 Department of Acute Medicine , University Hospitals of Derby and Burton , Derby , UK
• Correspondence to Dr Saajan Basi; saajan.basi{at}nhs.net

A 66-year-old man was admitted to hospital with a right frontal cerebral infarct producing left-sided weakness and a deterioration in his speech pattern. The cerebral infarct was confirmed with CT imaging. The only evidence of respiratory symptoms on admission was a 2 L oxygen requirement, maintaining oxygen saturations between 88% and 92%. In a matter of hours this patient developed a greater oxygen requirement, alongside reduced levels of consciousness. A positive COVID-19 throat swab, in addition to bilateral pneumonia on chest X-ray and lymphopaenia in his blood tests, confirmed a diagnosis of COVID-19 pneumonia. A proactive decision was made involving the patients’ family, ward and intensive care healthcare staff, to not escalate care above a ward-based ceiling of care. The patient died 5 days following admission under the palliative care provided by the medical team.

• respiratory medicine
• infectious diseases
• global health

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See:  http://creativecommons.org/licenses/by-nc/4.0/ .

https://doi.org/10.1136/bcr-2020-235920

Statistics from Altmetric.com

Request permissions.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

SARS-CoV-2 (Severe Acute Respiratory Syndrome Coronavirus 2) is a new strain of coronavirus that is thought to have originated in December 2019 in Wuhan, China. In a matter of months, it has erupted from non-existence to perhaps the greatest challenge to healthcare in modern times, grinding most societies globally to a sudden halt. Consequently, the study and research into SARS-CoV-2 is invaluable. Although coronaviruses are common, SARS-CoV-2 appears to be considerably more contagious. The WHO figures into the 2003 SARS-CoV-1 outbreak, from November 2002 to July 2003, indicate a total of 8439 confirmed cases globally. 1 In comparison, during a period of 4 months from December 2019 to July 2020, the number of global cases of COVID-19 reached 10 357 662, increasing exponentially, illustrating how much more contagious SARS-CoV-2 has been. 2

Previous literature has indicated infections, and influenza-like illness have been associated with an overall increase in the odds of stroke development. 3 There appears to be a growing correlation between COVID-19 positive patients presenting to hospital with ischaemic stroke; however, studies investigating this are in progress, with new data emerging daily. This patient report comments on and further characterises the link between COVID-19 pneumonia and the development of ischaemic stroke. At the time of this patients’ admission, there were 95 positive cases from 604 COVID-19 tests conducted in the local community, with a predicted population of 108 000. 4 Only 4 days later, when this patient died, the figure increased to 172 positive cases (81% increase), illustrating the rapid escalation towards the peak of the pandemic, and widespread transmission within the local community ( figure 1 ). As more cases of ischaemic stroke in COVID-19 pneumonia patients arise, the recognition and understanding of its presentation and aetiology can be deciphered. Considering the virulence of SARS-CoV-2 it is crucial as a global healthcare community, we develop this understanding, in order to intervene and reduce significant morbidity and mortality in stroke patients.

• Download figure
• Open in new tab
• Download powerpoint

A graph showing the number of patients with COVID-19 in the hospital and in the community over time.

Case presentation

A 66-year-old man presented to the hospital with signs of left-sided weakness. The patient had a background of chronic obstructive pulmonary disease (COPD), atrial fibrillation and had one previous ischaemic stroke, producing left-sided haemiparesis, which had completely resolved. He was a non-smoker and lived in a house. The patient was found slumped over on the sofa at home on 1 April 2020, by a relative at approximately 01:00, having been seen to have no acute medical illness at 22:00. The patients’ relative initially described disorientation and agitation with weakness noted in the left upper limb and dysarthria. At the time of presentation, neither the patient nor his relative identified any history of fever, cough, shortness of breath, loss of taste, smell or any other symptoms; however, the patient did have a prior admission 9 days earlier with shortness of breath.

The vague nature of symptoms, entwined with considerable concern over approaching the hospital, due to the risk of contracting COVID-19, created a delay in the patients’ attendance to the accident and emergency department. His primary survey conducted at 09:20 on 1 April 2020 demonstrated a patent airway, with spontaneous breathing and good perfusion. His Glasgow Coma Scale (GCS) score was 15 (a score of 15 is the highest level of consciousness), his blood glucose was 7.2, and he did not exhibit any signs of trauma. His abbreviated mental test score was 7 out of 10, indicating a degree of altered cognition. An ECG demonstrated atrial fibrillation with a normal heart rate. His admission weight measured 107 kg. At 09:57 the patient required 2 L of nasal cannula oxygen to maintain his oxygen saturations between 88% and 92%. He started to develop agitation associated with an increased respiratory rate at 36 breaths per minute. On auscultation of his chest, he demonstrated widespread coarse crepitation and bilateral wheeze. Throughout he was haemodynamically stable, with a systolic blood pressure between 143 mm Hg and 144 mm Hg and heart rate between 86 beats/min and 95 beats/min. From a neurological standpoint, he had a mild left facial droop, 2/5 power in both lower limbs, 2/5 power in his left upper limb and 5/5 power in his right upper limb. Tone in his left upper limb had increased. This patient was suspected of having COVID-19 pneumonia alongside an ischaemic stroke.

Investigations

A CT of his brain conducted at 11:38 on 1 April 2020 ( figure 2 ) illustrated an ill-defined hypodensity in the right frontal lobe medially, with sulcal effacement and loss of grey-white matter. This was highly likely to represent acute anterior cerebral artery territory infarction. Furthermore an oval low-density area in the right cerebellar hemisphere, that was also suspicious of an acute infarction. These vascular territories did not entirely correlate with his clinical picture, as limb weakness is not as prominent in anterior cerebral artery territory ischaemia. Therefore this left-sided weakness may have been an amalgamation of residual weakness from his previous stroke, in addition to his acute cerebral infarction. An erect AP chest X-ray with portable equipment ( figure 3 ) conducted on the same day demonstrated patchy peripheral consolidation bilaterally, with no evidence of significant pleural effusion. The pattern of lung involvement raised suspicion of COVID-19 infection, which at this stage was thought to have provoked the acute cerebral infarct. Clinically significant blood results from 1 April 2020 demonstrated a raised C-reactive protein (CRP) at 215 mg/L (normal 0–5 mg/L) and lymphopaenia at 0.5×10 9 (normal 1×10 9 to 3×10 9 ). Other routine blood results are provided in table 1 .

CT imaging of this patients’ brain demonstrating a wedge-shaped infarction of the anterior cerebral artery territory.

Chest X-ray demonstrating the bilateral COVID-19 pneumonia of this patient on admission.

• View inline

Clinical biochemistry and haematology blood results of the patient

Interestingly the patient, in this case, was clinically assessed in the accident and emergency department on 23 March 2020, 9 days prior to admission, with symptoms of shortness of breath. His blood results from this day showed a CRP of 22 mg/L and a greater lymphopaenia at 0.3×10 9 . He had a chest X-ray ( figure 4 ), which indicated mild radiopacification in the left mid zone. He was initially treated with intravenous co-amoxiclav and ciprofloxacin. The following day he had minimal symptoms (CURB 65 score 1 for being over 65 years). Given improving blood results (declining CRP), he was discharged home with a course of oral amoxicillin and clarithromycin. As national governmental restrictions due to COVID-19 had not been formally announced until 23 March 2020, and inconsistencies regarding personal protective equipment training and usage existed during the earlier stages of this rapidly evolving pandemic, it is possible that this patient contracted COVID-19 within the local community, or during his prior hospital admission. It could be argued that the patient had early COVID-19 signs and symptoms, having presented with shortness of breath, lymphopaenia, and having had subtle infective chest X-ray changes. The patient explained he developed a stagnant productive cough, which began 5 days prior to his attendance to hospital on 23 March 2020. He responded to antibiotics, making a full recovery following 7 days of treatment. This information does not assimilate with the typical features of a COVID-19 infection. A diagnosis of community-acquired pneumonia or infective exacerbation of COPD seem more likely. However, given the high incidence of COVID-19 infections during this patients’ illness, an exposure and early COVID-19 illness, prior to the 23 March 2020, cannot be completely ruled out.

Chest X-ray conducted on prior admission illustrating mild radiopacification in the left mid zone.

On the current admission, this patient was managed with nasal cannula oxygen at 2 L. By the end of the day, this had progressed to a venturi mask, requiring 8 L of oxygen to maintain oxygen saturation. He had also become increasingly drowsy and confused, his GCS declined from 15 to 12. However, the patient was still haemodynamically stable, as he had been in the morning. An arterial blood gas demonstrated a respiratory alkalosis (pH 7.55, pCO 2 3.1, pO 2 6.7 and HCO 3 24.9, lactate 1.8, base excess 0.5). He was commenced on intravenous co-amoxiclav and ciprofloxacin, to treat a potential exacerbation of COPD. This patient had a COVID-19 throat swab on 1 April 2020. Before the result of this swab, an early discussion was held with the intensive care unit staff, who decided at 17:00 on 1 April 2020 that given the patients presentation, rapid deterioration, comorbidities and likely COVID-19 diagnosis he would not be for escalation to the intensive care unit, and if he were to deteriorate further the end of life pathway would be most appropriate. The discussion was reiterated to the patients’ family, who were in agreement with this. Although he had evidence of an ischaemic stroke on CT of his brain, it was agreed by all clinicians that intervention for this was not as much of a priority as providing optimal palliative care, therefore, a minimally invasive method of treatment was advocated by the stroke team. The patient was given 300 mg of aspirin and was not a candidate for fibrinolysis.

Outcome and follow-up

The following day, before the throat swab result, had appeared the patient deteriorated further, requiring 15 L of oxygen through a non-rebreather face mask at 60% FiO 2 to maintain his oxygen saturation, at a maximum of 88% overnight. At this point, he was unresponsive to voice, with a GCS of 5. Although, he was still haemodynamically stable, with a blood pressure of 126/74 mm Hg and a heart rate of 98 beats/min. His respiratory rate was 30 breaths/min. His worsening respiratory condition, combined with his declining level of consciousness made it impossible to clinically assess progression of the neurological deficit generated by his cerebral infarction. Moreover, the patient was declining sharply while receiving the maximal ward-based treatment available. The senior respiratory physician overseeing the patients’ care decided that a palliative approach was in this his best interest, which was agreed on by all parties. The respiratory team completed the ‘recognising dying’ documentation, which signified that priorities of care had shifted from curative treatment to palliative care. Although the palliative team was not formally involved in the care of the patient, the patient received comfort measures without further attempts at supporting oxygenation, or conduction of regular clinical observations. The COVID-19 throat swab confirmed a positive result on 2 April 2020. The patient was treated by the medical team under jurisdiction of the hospital palliative care team. This included the prescribing of anticipatory medications and a syringe driver, which was established on 3 April 2020. His antibiotic treatment, non-essential medication and intravenous fluid treatment were discontinued. His comatose condition persisted throughout the admission. Once the patients’ GCS was 5, it did not improve. The patient was pronounced dead by doctors at 08:40 on 5 April 2020.

SARS-CoV-2 is a type of coronavirus that was first reported to have caused pneumonia-like infection in humans on 3 December 2019. 5 As a group, coronaviruses are a common cause of upper and lower respiratory tract infections (especially in children) and have been researched extensively since they were first characterised in the 1960s. 6 To date, there are seven coronaviruses that are known to cause infection in humans, including SARS-CoV-1, the first known zoonotic coronavirus outbreak in November 2002. 7 Coronavirus infections pass through communities during the winter months, causing small outbreaks in local communities, that do not cause significant mortality or morbidity.

SARS-CoV-2 strain of coronavirus is classed as a zoonotic coronavirus, meaning the virus pathogen is transmitted from non-humans to cause disease in humans. However the rapid spread of SARS-CoV-2 indicates human to human transmission is present. From previous research on the transmission of coronaviruses and that of SARS-CoV-2 it can be inferred that SARS-CoV-2 spreads via respiratory droplets, either from direct inhalation, or indirectly touching surfaces with the virus and exposing the eyes, nose or mouth. 8 Common signs and symptoms of the COVID-19 infection identified in patients include high fevers, severe fatigue, dry cough, acute breathing difficulties, bilateral pneumonia on radiological imaging and lymphopaenia. 9 Most of these features were identified in this case study. The significance of COVID-19 is illustrated by the speed of its global spread and the potential to cause severe clinical presentations, which as of April 2020 can only be treated symptomatically. In Italy, as of mid-March 2020, it was reported that 12% of the entire COVID-19 positive population and 16% of all hospitalised patients had an admission to the intensive care unit. 10

The patient, in this case, illustrates the clinical relevance of understanding COVID-19, as he presented with an ischaemic stroke underlined by minimal respiratory symptoms, which progressed expeditiously, resulting in acute respiratory distress syndrome and subsequent death.

Our case is an example of a new and ever-evolving clinical correlation, between patients who present with a radiological confirmed ischaemic stroke and severe COVID-19 pneumonia. As of April 2020, no comprehensive data of the relationship between ischaemic stroke and COVID-19 has been published, however early retrospective case series from three hospitals in Wuhan, China have indicated that up to 36% of COVID-19 patients had neurological manifestations, including stroke. 11 These studies have not yet undergone peer review, but they tell us a great deal about the relationship between COVID-19 and ischaemic stroke, and have been used to influence the American Heart Associations ‘Temporary Emergency Guidance to US Stroke Centres During the COVID-19 Pandemic’. 12

The relationship between similar coronaviruses and other viruses, such as influenza in the development of ischaemic stroke has previously been researched and provide a basis for further investigation, into the prominence of COVID-19 and its relation to ischaemic stroke. 3 Studies of SARS-CoV-2 indicate its receptor-binding region for entry into the host cell is the same as ACE2, which is present on endothelial cells throughout the body. It may be the case that SARS-CoV-2 alters the conventional ability of ACE2 to protect endothelial function in blood vessels, promoting atherosclerotic plaque displacement by producing an inflammatory response, thus increasing the risk of ischaemic stroke development. 13

Other hypothesised reasons for stroke development in COVID-19 patients are the development of hypercoagulability, as a result of critical illness or new onset of arrhythmias, caused by severe infection. Some case studies in Wuhan described immense inflammatory responses to COVID-19, including elevated acute phase reactants, such as CRP and D-dimer. Raised D-dimers are a non-specific marker of a prothrombotic state and have been associated with greater morbidity and mortality relating to stroke and other neurological features. 14

Arrhythmias such as atrial fibrillation had been identified in 17% of 138 COVID-19 patients, in a study conducted in Wuhan, China. 15 In this report, the patient was known to have atrial fibrillation and was treated with rivaroxaban. The acute inflammatory state COVID-19 is known to produce had the potential to create a prothrombotic environment, culminating in an ischaemic stroke.

Some early case studies produced in Wuhan describe patients in the sixth decade of life that had not been previously noted to have antiphospholipid antibodies, contain the antibodies in blood results. They are antibodies signify antiphospholipid syndrome; a prothrombotic condition. 16 This raises the hypothesis concerning the ability of COVID-19 to evoke the creation of these antibodies and potentiate thrombotic events, such as ischaemic stroke.

No peer-reviewed studies on the effects of COVID-19 and mechanism of stroke are published as of April 2020; therefore, it is difficult to evidence a specific reason as to why COVID-19 patients are developing neurological signs. It is suspected that a mixture of the factors mentioned above influence the development of ischaemic stroke.

If we delve further into this patients’ comorbid state exclusive to COVID-19 infection, it can be argued that this patient was already at a relatively higher risk of stroke development compared with the general population. The fact this patient had previously had an ischaemic stroke illustrates a prior susceptibility. This patient had a known background of hypertension and atrial fibrillation, which as mentioned previously, can influence blood clot or plaque propagation in the development of an acute ischaemic event. 15 Although the patient was prescribed rivaroxaban as an anticoagulant, true consistent compliance to rivaroxaban or other medications such as amlodipine, clopidogrel, candesartan and atorvastatin cannot be confirmed; all of which can contribute to the reduction of influential factors in the development of ischaemic stroke. Furthermore, the fear of contracting COVID-19, in addition to his vague symptoms, unlike his prior ischaemic stroke, which demonstrated dense left-sided haemiparesis, led to a delay in presentation to hospital. This made treatment options like fibrinolysis unachievable, although it can be argued that if he was already infected with COVID-19, he would have still developed life-threatening COVID-19 pneumonia, regardless of whether he underwent fibrinolysis. It is therefore important to consider that if this patient did not contract COVID-19 pneumonia, he still had many risk factors that made him prone to ischaemic stroke formation. Thus, we must consider whether similar patients would suffer from ischaemic stroke, regardless of COVID-19 infection and whether COVID-19 impacts on the severity of the stroke as an entity.

Having said this, the management of these patients is dependent on the likelihood of a positive outcome from the COVID-19 infection. Establishing the ceiling of care is crucial, as it prevents incredibly unwell or unfit patients’ from going through futile treatments, ensuring respect and dignity in death, if this is the likely outcome. It also allows for the provision of limited or intensive resources, such as intensive care beds or endotracheal intubation during the COVID-19 pandemic, to those who are assessed by the multidisciplinary team to benefit the most from their use. The way to establish this ceiling of care is through an early multidisciplinary discussion. In this case, the patient did not convey his wishes regarding his care to the medical team or his family; therefore it was decided among intensive care specialists, respiratory physicians, stroke physicians and the patients’ relatives. The patient was discussed with the intensive care team, who decided that as the patient sustained two acute life-threatening illnesses simultaneously and had rapidly deteriorated, ward-based care with a view to palliate if the further deterioration was in the patients’ best interests. These decisions were not easy to make, especially as it was on the first day of presentation. This decision was made in the context of the patients’ comorbidities, including COPD, the patients’ age, and the availability of intensive care beds during the steep rise in intensive care admissions, in the midst of the COVID-19 pandemic ( figure 1 ). Furthermore, the patients’ rapid and permanent decline in GCS, entwined with the severe stroke on CT imaging of the brain made it more unlikely that significant and permanent recovery could be achieved from mechanical intubation, especially as the damage caused by the stroke could not be significantly reversed. As hospitals manage patients with COVID-19 in many parts of the world, there may be tension between the need to provide higher levels of care for an individual patient and the need to preserve finite resources to maximise the benefits for most patients. This patient presented during a steep rise in intensive care admissions, which may have influenced the early decision not to treat the patient in an intensive care setting. Retrospective studies from Wuhan investigating mortality in patients with multiple organ failure, in the setting of COVID-19, requiring intubation have demonstrated mortality can be up to 61.5%. 17 The mortality risk is even higher in those over 65 years of age with respiratory comorbidities, indicating why this patient was unlikely to survive an admission to the intensive care unit. 18

Regularly updating the patients’ family ensured cooperation, empathy and sympathy. The patients’ stroke was not seen as a priority given the severity of his COVID-19 pneumonia, therefore the least invasive, but most appropriate treatment was provided for his stroke. The British Association of Stroke Physicians advocate this approach and also request the notification to their organisation of COVID-19-related stroke cases, in the UK. 19

Learning points

SARS-CoV-2 (Severe Acute Respiratory Syndrome Coronavirus 2) is one of seven known coronaviruses that commonly cause upper and lower respiratory tract infections. It is the cause of the 2019–2020 global coronavirus pandemic.

The significance of COVID-19 is illustrated by the rapid speed of its spread globally and the potential to cause severe clinical presentations, such as ischaemic stroke.

Early retrospective data has indicated that up to 36% of COVID-19 patients had neurological manifestations, including stroke.

Potential mechanisms behind stroke in COVID-19 patients include a plethora of hypercoagulability secondary to critical illness and systemic inflammation, the development of arrhythmia, alteration to the vascular endothelium resulting in atherosclerotic plaque displacement and dehydration.

It is vital that effective, open communication between the multidisciplinary team, patient and patients relatives is conducted early in order to firmly establish the most appropriate ceiling of care for the patient.

• Cannine M , et al
• Wunderink RG
• van Doremalen N ,
• Bushmaker T ,
• Morris DH , et al
• Wang X-G , et al
• Grasselli G ,
• Pesenti A ,
• Wang M , et al
• American Stroke Assocation, 2020
• Zhang Y-H ,
• Zhang Y-huan ,
• Dong X-F , et al
• Li X , et al
• Hu C , et al
• Zhang S , et al
• Jiang B , et al
• Xu J , et al
• British Association of Stroke Physicians

Contributors SB was involved in the collecting of information for the case, the initial written draft of the case and researching existing data on acute stroke and COVID-19. He also edited drafts of the report. MH was involved in reviewing and editing drafts of the report and contributing new data. SP oversaw the conduction of the project and contributed addition research papers.

Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Competing interests None declared.

Patient consent for publication Next of kin consent obtained.

Provenance and peer review Not commissioned; externally peer reviewed.

Read the full text or download the PDF:

• Access through  your organization
• Purchase PDF
• Patient Access
• Other access options

Article preview

Introduction, section snippets, references (150), cited by (35).

Journal of Neurolinguistics

The role of the right hemisphere in the recovery of stroke-related aphasia: a systematic review.

• • The right hemisphere contributes to early spontaneous language recovery in aphasia.
• • The right hemisphere facilitates treatment-related recovery of language.
• • In chronic aphasia right hemisphere activity may inhibit further recovery.
• o A limited aphasia, with preserved primary language areas (Broca and Wernicke's area), has a full recovery with minimal involvement of the right hemisphere.
• o A moderate aphasia, with an injury partially compromising the primary language areas, has a partial recovery with an activity increase in the right hemisphere in the subacute phase and a shift to the left hemisphere in the chronic phase.
• o A severe aphasia, with an injury totally affecting the primary language areas, has a weak recovery with significant activation in the right hemisphere in the subacute and chronic phase.

Research design and -quality

The role of the right hemisphere in spontaneous and therapy-related aphasia recovery, bilingual aphasia and language control: a follow-up fmri and intrinsic connectivity study, brain and language, the contribution of the right cerebral hemisphere to the recovery from aphasia: a single longitudinal case study, brain plasticity as a basis for recovery of function in humans, neuropsychologia, bilateral brain reorganization with memantine and constraint-induced aphasia therapy in chronic post-stroke aphasia: an erp study, acute conduction aphasia: an analysis of 20 cases, the role of the right hemisphere in recovery from aphasia. two case studies, behavioral and neurophysiologic response to therapy for chronic aphasia, archives of physical medicine and rehabilitation, different modes of word recognition in the left and right visual fields, neuropsychology of lexical ambiguity resolution: the contribution of divided visual field studies, lexical ambiguity resolution: perspectives from psycholinguistics, neuropsychology, and artificial intelligence, a pet follow-up study of recovery after stroke in acute aphasics, neurobehavioural measures in a person with aphasia before and after neuroplasticity based computerized cognitive training, procedia-social and behavioral sciences, effect of socioeconomic status on aphasia severity and recovery, regional changes in word-production laterality after a naming treatment designed to produce a rightward shift in frontal activity, neural recruitment associated with anomia treatment in aphasia, neural correlates of phonological and semantic-based anomia treatment in aphasia, influence of size and site of cerebral lesions on spontaneous recovery of aphasia and on success of language therapy, formal education, socioeconomic status, and the severity of aphasia after stroke, mechanisms of aphasia recovery after stroke and the role of noninvasive brain stimulation, neural activity associated with semantic versus phonological anomia treatments in aphasia, a proposed regional hierarchy in recovery of post-stroke aphasia, disturbance and recovery of language function: correlates in pet activation studies, the spectrum of aphasia subtypes and etiology in subacute stroke, journal of stroke and cerebrovascular diseases, brain plasticity in poststroke aphasia: what is the contribution of the right hemisphere, computer tomographic localization, lesion size, and prognosis in aphasia and nonverbal impairment, the structural determinants of recovery in wernicke' s aphasia, normalisation and increase of abnormal erp patterns accompany recovery from aphasia in the post-acute stage, illiteracy and brain damage—1. aphasia testing in culturally contrasted populations (control subjects), the prognosis for aphasia in stroke, improving language without words: first evidence from aphasia, therapy-induced neuroplasticity in chronic aphasia, recovering two languages with the right hemisphere, functional re-recruitment of dysfunctional brain areas predicts language recovery in chronic aphasia, recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fmri, preferred reporting items for systematic reviews and meta-analyses: the prisma statement, international journal of surgery, effectiveness of low-frequency rtms and intensive speech therapy in poststroke patients with aphasia: a pilot study based on evaluation by fmri in relation to type of aphasia, european neurology, alterations in the functional anatomy of reading induced by rehabilitation of an alexic patient, cognitive and behavioral neurology.

• Aerts, A., van Mierlo, P., Hartsuiker, R., Santens, P., & De Letter, M. (2014). Neurophysiological alterations during...

The complementary role of the cerebral hemispheres in recovery from aphasia after stroke: A critical review of literature

Brain injury, thresholds in cerebral ischemia - the ischemic penumbra, making a case for acute ischemic stroke, journal of pharmacological practice, mapping the ischaemic penumbra with pet: implications for acute stroke treatment, cerebrovascular diseases, the effects of low frequency repetitive transcranial magnetic stimulation (rtms) and sham condition rtms on behavioural language in chronic non-fluent aphasia: short term outcomes, neurorehabilitation, improved language performance subsequent to low-frequency rtms in patients with chronic non-fluent aphasia post-stroke, european journal of neurology, longitudinal modulation of n400 in chronic non-fluent aphasia using low-frequency rtms: a randomised placebo controlled trial, aphasiology, prognostic factors in aphasia, memantine and constraint-induced aphasia therapy in chronic poststroke aphasia, annals of neurology, drug therapy of post-stroke aphasia: a review of current evidence, neuropsychology review, changes in maps of language function and the integrity of the arcuate fasciculus after therapy for chronic aphasia, changes in language-specific brain activation after therapy for aphasia using magnetoencephalography: a case study, changes in maps of language activity activation following melodic intonation therapy using magnetoencephalography: two case studies, journal of clinical and experimental neuropsychology, efficacy of repetitive transcranial magnetic stimulation in treating stroke aphasia: systematic review and meta-analysis.

Consequently, suppression of these abnormally activated right hemisphere regions may allow the spared left hemisphere regions to contribute to more robust language recovery. Yet another approach proposes that activation of certain right hemisphere regions can be compensatory, suggesting that excitatory stimulation to the preserved right hemisphere regions can facilitate recovery (Cocquyt et al., 2017; Turkeltaub, 2015). Reflecting these different models of neuroplasticity, investigators have applied excitatory rTMS to the left inferior frontal and temporal peri-stroke areas and inhibitory stimulation to the homotopic right hemisphere regions (Barwood et al., 2013; Harvey et al., 2017; Hu et al., 2018; Khedr et al., 2014; Ren et al., 2019; Rubi-Fessen et al., 2015; Seniów et al., 2013; Szaflarski et al., 2018; Thiel et al., 2013; Wang et al., 2014; Yoon et al., 2015).

Left frontotemporal effective connectivity during semantic feature judgments in patients with chronic aphasia and age-matched healthy controls

For example, Heiss and Thiel (2006) proposed a three-tiered framework of chronic aphasia in which optimal, satisfactory, and poor language recovery were associated with, respectively, preservation or reactivation of primary language areas; damage to primary language cortex but spared and functional secondary left hemisphere language areas; and extensive left hemisphere damage with reliance on the contralesional (but possibly ill-suited) right hemisphere to mediate language. However, the past 25 years of collective fMRI and PET research in aphasia has resulted in conflicting evidence for these patterns, particularly the contested role of the right hemisphere in aphasia recovery (see Cappa, 2011; Cocquyt, De Ley, Santens, Van Borsel, & De Letter, 2017; Crosson et al., 2007; Kiran, 2012; Meinzer, Harnish, Conway, & Crosson, 2011; Price & Crinion, 2005; Saur & Hartwigsen, 2012; Thompson & den Ouden, 2008; Zahn, Schwarz, & Huber, 2006 for reviews). Imaging findings accord with the general assertion that left hemisphere activation—and perilesional activity in particular—results in better language recovery in PWA (Fridriksson, Richardson, Fillmore, & Cai, 2012; Heiss, Kessler, Thiel, Ghaemi, & Karbe, 1999; Léger et al., 2002; Marcotte & Ansaldo, 2010; Meinzer & Breitenstein, 2008; Meinzer et al., 2008; Menke et al., 2009; Rosen et al., 2000; Szaflarski, Allendorfer, Banks, Vannest, & Holland, 2013; Vitali et al., 2007; Warburton, Price, Swinburn, & Wise, 1999; Winhuisen et al., 2007; van Hees, McMahon, Angwin, de Zubicaray, & Copland, 2014).

Neuroplasticity in Post-Stroke Aphasia: A Systematic Review and Meta-Analysis of Functional Imaging Studies of Reorganization of Language Processing

A review on treatment-related brain changes in aphasia, change in right inferior longitudinal fasciculus integrity is associated with naming recovery in subacute poststroke aphasia, transient perturbation of the left temporal cortex evokes plasticity-related reconfiguration of the lexical network.

Delusions and the Right Hemisphere: A Review of the Case for the Right Hemisphere as a Mediator of Reality-Based Belief

Information & authors, metrics & citations, view options.

Some faces that [the patient] sees with their normal features, the memory of which is not altered in any way, are nevertheless no longer accompanied by this feeling of exclusive familiarity which determines…immediate recognition… The patient, whilst picking up on a very narrow resemblance between two images, ceases to identify them because of the different emotions they elicit. (Capgras & Carrette 1924) 5

The Right Hemisphere and Delusions: A Brief History

Pragmatic communication, perceptual integration, attentional surveillance, self-monitoring, and novelty/anomaly detection, belief updating, the right hemisphere at the interface of self, environment, and reality, information, published in.

• Hemispheric Asymmetries and Lateralization
• Organic Mental Disorders
• Stroke and Other Cerebral Vascular Disease (Neuropsychiatric Aspects)
• Traumatic Brain Injury

Competing Interests

Export citations.

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu .

View options

Login options.

Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.

Purchase Options

Purchase this article to access the full text.

PPV Articles - Journal of Neuropsychiatry and Clinical Neurosciences

Not a subscriber?

Subscribe Now / Learn More

PsychiatryOnline subscription options offer access to the DSM-5-TR ® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.

Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).

Share article link

Copying failed.

PREVIOUS ARTICLE

Next article, request username.

Can't sign in? Forgot your username? Enter your email address below and we will send you your username

If the address matches an existing account you will receive an email with instructions to retrieve your username

Create a new account

Change password, password changed successfully.

Your password has been changed

Reset password

Can't sign in? Forgot your password?

Enter your email address below and we will send you the reset instructions

If the address matches an existing account you will receive an email with instructions to reset your password.

Your Phone has been verified

As described within the American Psychiatric Association (APA)'s Privacy Policy and Terms of Use , this website utilizes cookies, including for the purpose of offering an optimal online experience and services tailored to your preferences. Please read the entire Privacy Policy and Terms of Use. By closing this message, browsing this website, continuing the navigation, or otherwise continuing to use the APA's websites, you confirm that you understand and accept the terms of the Privacy Policy and Terms of Use, including the utilization of cookies.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 25 September 2024

Human hippocampal and entorhinal neurons encode the temporal structure of experience

• Pawel Tacikowski   ORCID: orcid.org/0000-0003-2434-2112 1 , 2 , 3 ,
• Güldamla Kalender 1   nAff6 ,
• Davide Ciliberti   ORCID: orcid.org/0000-0003-1229-642X 1   nAff7 &
• Itzhak Fried   ORCID: orcid.org/0000-0002-5962-2678 1 , 4 , 5  

Nature ( 2024 ) Cite this article

Metrics details

• Cognitive neuroscience
• Learning and memory

Extracting the underlying temporal structure of experience is a fundamental aspect of learning and memory that allows us to predict what is likely to happen next. Current knowledge about the neural underpinnings of this cognitive process in humans stems from functional neuroimaging research 1 , 2 , 3 , 4 , 5 . As these methods lack direct access to the neuronal level, it remains unknown how this process is computed by neurons in the human brain. Here we record from single neurons in individuals who have been implanted with intracranial electrodes for clinical reasons, and show that human hippocampal and entorhinal neurons gradually modify their activity to encode the temporal structure of a complex image presentation sequence. This representation was formed rapidly, without providing specific instructions to the participants, and persisted when the prescribed experience was no longer present. Furthermore, the structure recovered from the population activity of hippocampal–entorhinal neurons closely resembled the structural graph defining the sequence, but at the same time, also reflected the probability of upcoming stimuli. Finally, learning of the sequence graph was related to spontaneous, time-compressed replay of individual neurons’ activity corresponding to previously experienced graph trajectories. These findings demonstrate that neurons in the hippocampus and entorhinal cortex integrate the ‘what’ and ‘when’ information to extract durable and predictive representations of the temporal structure of human experience.

Extracting temporal patterns of recurring events is fundamentally important for organizing information in memory, predicting the future and guiding flexible behaviours 6 , 7 . How this process is carried out by neurons in the human brain remains unknown. Studies on spatial navigation provide some important clues, as moving through space essentially corresponds to a sequence of visiting locations characterized by specific neuronal signatures. A ‘cognitive map’ of the spatial environment is encoded by a range of interacting neuron types, including hippocampal ‘place cells’ that fire when the animal is at a specific location 8 , 9 and entorhinal ‘grid cells’ that provide a metric of spatial distance 10 , 11 . Remarkably, the brain uses similar neural principles to represent non-spatial features, such as sound frequency 12 , object characteristics 13 , abstract space 14 and time 15 , 16 . This cognitive map is predictive, in that it informs about future states that the agent is likely to experience 17 , 18 , 19 , 20 . The fact that hippocampal–entorhinal neurons represent relations between features of information and encode time makes this brain circuit an ideal candidate system to extract the temporal structure of experience. Functional neuroimaging research in humans generally supports this view 1 , 2 , 3 , 4 , 5 , but how such extraction is achieved by hippocampal–entorhinal neurons remains unknown.

Here we recorded extracellular spiking activity from 17 patients with epilepsy who were implanted with intracranial depth electrodes for clinical reasons 21 (Fig. 1a and Supplementary Table 1 ; 21 recording sessions). Our experimental paradigm capitalized on the fact that the human medial temporal lobe (MTL) contains neurons that respond selectively to particular people 22 , 23 . For each participant, we selected six images that were associated with preferential neuronal responses in the preceding screening experiment. Each image was then arbitrarily assigned to a different location on a pyramid graph (Fig. 1b ). There were three main study phases: pre-exposure (PRE), exposure, and post-exposure (POST) (Fig. 1c ). During PRE (baseline), images were displayed in pseudo random order (60 direct and 60 indirect graph-transitions). During the subsequent six exposure phases (E1–E6), the order of image presentations was determined by the pyramid graph, so that only images directly linked on the graph were displayed immediately one after another (Fig. 1c ). Finally, POST (read-out) was identical to PRE; during this phase, there was no pyramid rule in the sequence of image presentations. During every phase, the participants performed behavioural tasks that were unrelated to the temporal pyramid rule (Fig. 1c ). We hypothesized that hippocampal–entorhinal neurons gradually represent the temporal pyramid structure by responding in an increasingly similar manner to stimuli directly linked on the graph. Note that the configuration of directly and indirectly connected nodes is different, depending on whether the seed is an inner or an outer node (Fig. 1d,f ).

a , Top, extracellular spiking activity was recorded from eight microwires extending from the tip of each depth macro-electrode. There were 7–12 macro-electrodes per patient. Raw local field potential signal was high-pass filtered and thresholded to detect spiking activity. Bottom, spike waveforms from one whole recording session, grouped into two clusters (two putative neurons) based on the waveforms’ amplitude and shape. b , The sequence of stimuli presentation (bottom) corresponded to a ‘random walk’ on a pyramid graph (top) so that only images directly linked on the graph were displayed immediately after another. c , The participant’s task was either to determine whether each displayed image shows a male or a female (gender task; PRE and POST) or whether the image is the same or mirrored when compared to PRE (E1–E6). d , A schematic representation of the hypothesis. Circles represent ‘place fields’ of selective neurons in abstract space. Before exposure, each neuron responds preferentially to a different image, and the arrangement of place fields is largely random. After exposure to the pyramid structure, the green neuron should respond more strongly to images directly linked on the pyramid to its preferred stimulus (magenta) than to images linked indirectly (blue). The same logic applies to all nodes, regardless of whether the ‘seed’ is at an inner or an outer node (see f ). e , Neuronal activity was recorded from multiple brain regions, including the hippocampal–entorhinal system and amygdalae (shaded area). Dots represent localizations of microwires where putative neurons were detected. These localizations are overlaid on the 152-MNI-T1 3D template brain rendered by MRIcroGL software. f , A significant proportion of selective neurons was found in the hippocampus (H), entorhinal cortex (EC) and parahippocampal gyrus (PH). Each row of the heat maps shows the mean spiking activity of one neuron during PRE ( z -scored and baseline-corrected; −0.5 to 0 s). The plot on the right shows mean responses ± s.e.m. from all selective neurons. Note that owing to copyright issues, all original images used in the study were replaced in this and all subsequent figures by comparable free stock photos. The original images are available from the corresponding authors.

Individual neurons

Altogether, we identified 1,456 single- and multi-units (hereafter called ‘neurons’) across multiple brain regions (Fig. 1e and Supplementary Table 2 ). The unit yield was generally high and comparable across the participants (minimum = 27, maximum = 118, average = 69; Supplementary Table 2 ). We first identified selective neurons that responded significantly more strongly to one stimulus than to all other stimuli during PRE ( Methods ). Note that selectivity was defined in a narrow sense, only relative to other images used in the current study. We found a significant proportion of selective neurons in the hippocampus, entorhinal cortex and parahippocampal gyrus (Fig. 1f , Extended Data Fig. 2 and Supplementary Tables 3 and 4 ; n  = 152, n  = 111, and n  = 33, respectively; 45%, 53%, and 56% of all identified neurons from those regions, respectively; P  < 0.001 above chance level for all three regions). Depending on the position of the preferred stimulus on the pyramid graph, we classified the remaining stimuli as ‘direct’ or ‘indirect’ for a given neuron and used these labels consistently to analyse the subsequent study phases. On average, each node on the graph was associated with preferential responses of 49 selective neurons (minimum = 33, maximum = 64; across all recording sessions). Behavioural data showed that the participants generally followed the instructions and completed the experimental tasks successfully (Extended Data Fig. 1a ). Furthermore, stimuli transitions during POST that violated the sequence rules from exposure phases were related to increased response latencies, which suggests that the patients extracted the pyramid graph and used it to guide their behaviour, despite the lack of specific instructions to do so and the task-irrelevant nature of the pyramid (Extended Data Fig. 1b ). At the same time, when asked, “Have you noticed any pattern in the stimulus sequence?” none of the patients reported noticing a graph-like organization of the states. A separate behavioural study conducted on twenty-five healthy controls further supported the lack of detailed explicit knowledge of the pyramid structure by the participants after completing the same version of the task as the patients (Extended Data Fig. 1d ). Together, the above results validate our methodological approach and show that learning the pyramid was largely implicit.

Moving on to the main analysis, we identified temporal ‘relational neurons’ that increased their responses to direct stimuli throughout the study ( Methods ). We found a significant proportion of such neurons specifically in the entorhinal cortex and hippocampus (Extended Data Fig. 2 and Supplementary Tables 3 and 5 ; n  = 42 and n  = 55, respectively; 20% and 16% of all identified neurons in those regions, respectively; P  = 0.024 and P  = 0.012 above the chance level, respectively). Figure 2a shows two representative relational neurons from the right hippocampus (see also Extended Data Fig. 3a ). Of note, these two cells continued to respond more strongly to direct stimuli even during POST, when the order of image presentations no longer followed the pyramid rule and when the behavioural task had changed (Fig. 2b ). Responses of all hippocampal–entorhinal relational neurons to direct stimuli were significantly stronger during late experiment phases (E5 and E6) than during PRE, and significantly stronger during POST than during PRE (Fig. 2c and Extended Data Fig. 4 ; P  = 3.56 × 10 −5 and P  = 0.018, respectively; two-sided Wilcoxon signed-rank test, false discovery rate (FDR)-corrected; note that data from PRE and POST were not used in the statistical selection of these neurons—thus, the above results are not biased by the selection criterion;   Methods ). Notably, relational neurons also gradually decreased their responses to preferred stimuli (Fig. 2c,d and Extended Data Fig. 4 ; comparisons versus PRE; E1 and E2: P  = 1.98 × 10 −6 ; E3 and E4: P  = 2.26 × 10 −7 ; E5 and E6: P  = 9.69 × 10 −9 ; POST: P  = 1.07 × 10 −9 ; two-sided Wilcoxon signed-rank tests, FDR-corrected; for general results on neurons that gradually decreased their selectivity, see Extended Data Figs. 2 and 3b and Supplementary Tables 3 and 6 ). The above results support our hypothesis by demonstrating that hippocampal–entorhinal neurons that initially responded preferentially to one image gradually embedded the pyramid graph, by showing diminished selectivity to that image and increased responses to adjacent stimuli.

a , Two representative hippocampal neurons that responded preferentially to the image of the policeman (left) during PRE. With exposure to the pyramid rule, they began to respond more strongly to images directly linked to their preferred stimulus on the graph (direct) than to images linked indirectly (indirect). b , These two neurons continued to show the same pattern of responses during POST, when the pyramid rule had stopped and the behavioural task had changed. c , Average responses (±s.e.m.) of all relational neurons in the hippocampal–entorhinal region ( n  = 97). Apart from showing increasingly stronger responses to direct images, these neurons showed gradually diminishing selectivity for their preferred stimulus. Each neuron’s responses were z -scored and baseline-corrected (−0.5 to 0 s). d , Two representative neurons showing diminishing selectivity (the bottom panel shows the same neuron as a , right). Raster plots in a , b , d show individual spikes during each stimulus presentation. Line plots in a , b show the mean number of spikes ± s.e.m. Neurons’ identifiers are provided in round brackets.

Population code

Next, we tested whether the pyramid representation was robust enough to shift the activity pattern of the entire hippocampal–entorhinal neuronal population. To this end, we used the Bayesian naive classifier to decode stimulus identity during each image presentation ( Methods ). Instead of simply checking whether decoding was correct, we analysed posterior probabilities that the decoder assigned to the image actually presented (actual), images directly linked to that stimulus on the graph (direct) and images linked indirectly (indirect) (Fig. 3a and  Methods ). This analysis was performed for each recording session separately because of the different stimuli used, but the resulting probability distributions were combined across all sessions. The classifier was trained on data from PRE and tested on all subsequent study phases (for testing in PRE, we used the ‘leave-one-out’ cross-validation). The analysis was performed on all identified hippocampal and entorhinal neurons, regardless of their selectivity ( n  = 546). We found that the data from PRE contained enough information to decode stimulus identity significantly above the chance level; this important prerequisite makes the analysis of subsequent phases meaningful (Fig. 3b and Extended Data Fig. 5 ). Over the course of the study, the classifier assigned progressively lower probabilities to the images actually presented (Fig. 3c ; comparisons versus PRE; E1 and E2: P  = 0.035; E3 and E4: P  = 1.14 × 10 −8 ; E5 and E6: P  = 1.49 × 10 −11 ; POST: P  = 2.02 × 10 −16 ). By contrast, the classifier assigned increasingly higher probabilities to stimuli that were directly linked to the actual stimuli on the pyramid graph (Fig. 3c ; comparisons versus PRE; E1 and E2: P  = 0.313; E3 and E4: P  = 0.022; E5 and E6: P  = 0.002; POST: P  = 1.74 × 10 −4 ). Probability distributions for indirectly linked stimuli did not change significantly over the course of the study (Fig. 3c ; comparisons versus PRE; E1 and E2: P  = 0.722; E3 and E4: P  = 0.518; E5 and E6: P  = 0.442; POST: P  = 0.114). The difference between distributions for direct and indirect stimuli was significant even during POST, when the order of image presentations did not follow the pyramid structure and the behavioural task had changed (Fig. 3d ; POST-direct versus POST-indirect: P  = 8.61 × 10 −5 ). For all the above comparisons, we used Kolmogorov–Smirnov tests (one-sided). It is noteworthy that an analogous control analysis performed on all neurons outside of the hippocampal–entorhinal system did not reveal any consistent evidence of the pyramid representation (Extended Data Fig. 6a ; n  = 910). The above findings validate and go beyond the results from individual neurons, by showing that the pyramid graph representation affected the activity of the entire neuronal population in the hippocampal–entorhinal complex.

a , The logic behind the population decoding analysis. b , Neuronal responses contained enough information to successfully decode the stimulus identity during PRE (chance level ≈ 17%; data from all identified neurons; n  = 1,456). The plot shows mean decoding accuracy (±s.e.m.) from 100-ms bins averaged across all recording sessions ( n  = 21; time zero is the stimulus onset). The shaded grey area marks the time window used for further analyses. c , Results from the hippocampal–entorhinal neurons ( n  = 546). P values obtained from the Kolmogorov–Smirnov tests between cumulative distribution functions (CDFs) of posterior probabilities assigned by the decoder during PRE versus subsequent study phases (one-sided). d , The difference between CDFs for direct and indirect stimuli remained significant during POST (Kolmogorov–Smirnov test; one-sided). e , Top row, combined data from trials where the actually presented stimulus was at an outer node of the pyramid. P values represented by dotted or solid lines of different widths were obtained from Kolmogorov–Smirnov tests between each pair of nodes (one-sided; FDR-corrected). Colour intensities correspond to distances (Kolmogorov–Smirnov z -statistic) between the respective CDFs. The seed node is marked in orange. Bottom row, analogous results for trials where the stimulus actually presented was at an inner node. NS, not significant. f , Distance matrixes and graphs corresponding to the geodesic, Euclidean and successor templates. Each graph shows the most faithful 2D representation of the respective distance matrix using the multidimensional scaling analysis. Note that the matrix and graph obtained from the neuronal data (right) closely resemble the successor template (546 hippocampal–entorhinal neurons; E4–E6 data combined for illustration purposes). g , The degree of similarity between data and each template throughout the study. Spearman’s correlation coefficients (Fisher-transformed) between each template and neural data from respective phases (changes from PRE).

Next, we tested whether the neuronal representation of the pyramid graph followed geodesic geometry—that is, whether distances between neuronal responses to different nodes were equivalent to the minimum number of edges connecting these nodes (that is, the ‘shortest path’ distance). If that were the case, there should be: (1) no prominent differences when all direct nodes are compared to each other; and (2) no prominent differences when all indirect nodes are contrasted with each other. Because the pyramid graph is symmetrical, we grouped together all trials where the actually displayed image (seed) was located at one of the three outer nodes of the pyramid and calculated pairwise distances between posterior probability distributions (see previous paragraph) assigned to the seed versus the remaining nodes. This analysis was performed on all hippocampal–entorhinal neurons ( n  = 546), first for each recording session separately and then combined across all sessions. As expected, during PRE, the seed node differed significantly from all other nodes and the other nodes did not differ significantly between each other or differed only marginally (Fig. 3e ). During POST, the seed still differed significantly from the remaining nodes, but now, both direct nodes also significantly differed from the indirect-outer nodes (Fig. 3e ). Notably, there was no significant difference between either of the direct nodes and the indirect-inner node. Thus, not all indirect nodes changed their representations to a similar degree, suggesting that the neural encoding of the pyramid was not strictly geodesic (see earlier). An analogous analysis for all inner seed nodes combined revealed generally similar results (Fig. 3e ; bottom). In line with this, a single-neuron analysis showed that there was a small but significant proportion of hippocampal relational neurons that during the late study phases responded significantly more strongly to indirect-inner than to indirect-outer nodes ( n  = 12; 4% of all hippocampal neurons; P  = 0.024). A population decoding analysis analogous to Fig. 3c further supported that the decoding probability of indirect-inner nodes changed throughout the study in a similar manner as the decoding probability of direct nodes (Extended Data Fig. 6c ). Together, these findings indicate that the population of hippocampal–entorhinal neurons accurately encoded the general layout of the pyramid graph, but this mapping was not strictly geodesic.

Recovering the entire graph

Next, we tested whether it was possible to reconstruct the entire pyramid structure from the population activity of hippocampal and entorhinal neurons and if so, what geometry that representation followed. To this end, we calculated Euclidean distances between neurons’ responses to each image versus all other images (this was done for all subsequent study phases), and then, compared the resulting distance matrixes to three templates (Fig. 3f ). In the ‘geodesic template’, distances between each pair of nodes corresponded to the shortest path (see ‘Population code’). In the ‘Euclidean template’, distances between relevant nodes (1–5, 2–6 and 4–3) were calculated from the Pythagorean theorem. The ‘successor template’ assumed that the pyramid representation is predictive. This idea has been previously formalized as the ‘successor representation’, which informs how often an agent will experience a particular destination state after starting in the initial state 17 , 18 , 19 , 20 . In the present study, temporal predictions can be based on the structure of the pyramid itself. Specifically, the length of all possible paths between the different inner nodes is generally shorter than the length of all paths connecting the outer nodes. Thus, during a random walk, the inner nodes are likely to occur closer in time. If the neural representation is predictive, the above regularities should significantly distort the graph’s representation by shortening distances between the inner nodes (Fig. 3f and   Methods ).

We found that over the course of the study, all templates improved their fit to the neural data (Fig. 3g ; geodesic: E1 and E2: P  = 0.0231; E3 and E4: P  = 0.009; E5 and E6: P  = 0.009; POST: P  = 0.2335; Euclidean: E1 and E2: P  = 0.0035; E3 and E4: P  = 0.0008; E5 and E6: P  = 0.0008; POST: P  = 0.1397; successor: E1 and E2: P  = 0.0001; E3 and E4: P  < 0.0001; E5 and E6: P  < 0.0001; POST: P  = 0.0434; differences from PRE; 10,000 permutations; FDR-corrected;   Methods ). However, the successor template significantly outperformed the other templates (Fig. 3g ; geodesic: E1 and E2: P  < 0.0001; E3 and E4: P  < 0.0001; E5 and E6: P  = 0.0001; POST: P  = 0.0295; Euclidean: E1 and E2: P  = 0.0094; E3 and E4: P  = 0.0094; E5 and E6: P  = 0.034; POST: P  = 0.0825; differences from PRE; 10,000 permutations; FDR-corrected;   Methods ). Remarkably, the patients who developed a robust hippocampal–entorhinal successor representation showed longer reaction times during trials in POST that violated the pyramid rules from exposure phases (see ‘Individual neurons’), which suggests that this representation was used to guide behaviour (Extended Data Fig. 1c ). A control analysis performed on neurons outside of the hippocampal–entorhinal complex did not show any significant evidence that the pyramid representation was present during POST, either when compared to the geodesic, Euclidean or successor templates (Extended Data Fig. 6b ). Together, these findings demonstrate that the coordinated activity of multiple hippocampal–entorhinal neurons progressively represented a detailed structure of the entire temporal structure and that this representation was predictive in nature.

Apart from affecting the pyramid encoding at the population level (see above), the successor representation should modulate the activity of individual neurons. For example, during spatial navigation, the successor model accounts for the warping of place cells’ receptive fields around environmental barriers 17 . If the pyramid representation involved similar mechanisms, the receptive fields of neurons representing the pyramid’s outer nodes should elongate throughout the study because, from these nodes, the ‘agent’ can only proceed in one general direction (that is, back). Conversely, receptive fields of neurons representing the inner nodes of the pyramid should be more symmetric, as from these nodes, the agent can move in three directions. To complement our neuronal population results that support the above hypotheses (Fig. 3e ), we measured the distance between individual neurons’ responses to different stimuli. By analogy with place cells, we analysed selective hippocampal–entorhinal neurons grouped by their preferred node (inner: n  = 144; outer: n  = 119). We found that neurons selective to an outer node responded significantly differently to indirect-inner versus indirect-outer nodes, which is consistent with the elongation of their receptive fields (Fig. 4a ; E1 and E2: P  = 0.0151; E3 and E4: P  = 0.0151; E5 and E6: P  = 0.0298; POST: P  = 0.1869; two-sided Wilcoxon rank-sum test; FDR-corrected). By contrast, neurons preferring an inner node did not respond significantly differently to various outer nodes, which suggests that their receptive fields remained symmetric (Fig. 4a ; P  = 0.9825 in all phases; two-sided Wilcoxon rank-sum test, FDR-corrected). Additionally, we found that the ‘inner-to-inner distances’ became shorter than the ‘outer-to-inner distances’, which is also in line with the successor representation (Extended Data Fig. 7a ; E1 and E2: P  = 0.005; E3 and E4: P  = 0.0116; E5 and E6: P  = 0.0116; POST: P  = 0.299; two-sided Wilcoxon rank-sum test; FDR-corrected). The above results closely resemble functional properties of place cells during spatial navigation and reveal single-neuron mechanisms of predictive representations of temporal structures.

a , Top, selective hippocampal–entorhinal neurons that preferred a stimulus at an outer node ( n  = 119) responded significantly differently to stimuli from indirect-inner versus indirect-outer nodes, suggesting that these neurons’ receptive fields progressively elongated. Bottom, there was no such effect for neurons that preferred a stimulus at an inner node ( n  = 144), which suggests that their receptive fields were rather symmetrical. Plots show the mean Euclidean distance (±s.e.m.) between responses to respective stimuli (data centred on PRE and z -scored per neuron). P values from Wilcoxon rank-sum tests (two-sided, FDR-corrected). Orange circles indicate the locations of preferred stimuli. Orange areas illustrate the hypothesized shapes of receptive fields. b , The successor representation in the hippocampus was more impaired by growing proportions of artificially removed neurons than in the entorhinal cortex. Similarity to the successor template is plotted as a function of the percentage of removed neurons (relative to 1% of neurons removed; for each 1% step, we randomly selected a given proportion of neurons 10,000 times). The actual difference between the third quartiles was compared with the same difference in 1,000 permutations of the region labels. c , The replay analysis focused on three-element graph trajectories consisting of one seed node, a direct node and an indirect node. We analysed triplets of selective hippocampal–entorhinal neurons (recorded in the same session) whose preferred stimuli mapped onto those trajectories. Only spiking activity during breaks between phases was analysed (B1–B7). d , Examples of pyramid-congruent replays detected for triplets of selective hippocampal–entorhinal neurons. Coloured circles indicate the graph location of each neuron’s preferred stimulus during PRE. Raster plots show the spiking activity of ‘direct’ and ‘indirect’ neurons during each spontaneous repetition of a given replay. The bottom plot shows combined spiking activity across all repetitions and the mean spikes’ latencies (±s.e.m.). Plots are time-locked to the seed neuron’s relevant spikes. The probability of pyramid-congruent replays increased throughout the study and in B2–B7 was significantly higher than that of incongruent replays (1,000 random permutations of ‘direct’ and ‘indirect’ spike labels). P values in b , d were calculated as the number of permutations with a higher difference than the one actually detected, divided by the total number of permutations. If in none of the permutations the difference was above the actual one, the P  < 0.001 range is reported. No adjustment for multiple comparisons was applied in d .

Hippocampal versus entorhinal codes

Next, we tested whether the neuronal pyramid representation differed between the hippocampus and the entorhinal cortex. We found that during exposure phases (E1–E6), hippocampal neurons represented the pyramid more accurately than entorhinal neurons (successor: P  = 0.0429; Euclidean: P  = 0.0055; geodesic: P  = 0.0042; H minus EC difference between Spearman correlation coefficients for each template; P values based on 10,000 permutations of the brain region labels). The above result is not simply due to a different number of hippocampal and entorhinal neurons that we detected in this study, as the above analysis balanced this aspect (10,000 random selections of subsets of hippocampal neurons to match the number of entorhinal neurons). Interestingly, during POST, the successor representation was more preserved in the entorhinal cortex than in the hippocampus, suggesting that the former utilizes a more stable neuronal code than the latter (successor: P  = 0.037; Euclidean: P  = 0.5963; geodesic: P  = 0.5484; EC minus H difference between Spearman correlation coefficients for each template; P values based on 10,000 permutations of region labels; the number of neurons was balanced; see above). We also tested the robustness of hippocampal versus entorhinal representations against removing growing proportions of neurons from each region. The presumably more structural (‘pure-position’) neural code in the entorhinal cortex should be less affected by such removals than the relational (object-based) code in the hippocampus 2 , 7 . Indeed, as we removed more neurons from the analysis, similarity to the successor template diminished more rapidly in the hippocampus than in the entorhinal cortex (Fig. 4b ; P  = 0.007; difference between third quartiles; P value from 1,000 permutations of region labels; number of neurons balanced; combined data from E1–E6). Analogous differences were not significant for the Euclidean and geodesic templates ( P  = 0.622 and P  = 0.296, respectively). The above findings suggest that the hippocampus contains a more dynamic object-related representation of temporal sequences, whereas the entorhinal cortex uses a more stable structural code.

Neuronal replay

Neuronal representation of the pyramid was likely to rely on mechanisms of synaptic plasticity, where the ordering of spikes from the pre-and post-synaptic cells determines whether long-term potentiation or depression occurs 24 . But how can relations between stimuli that occurred seconds apart rely on synaptic phenomena that have a time window of approximately 30ms? One possible explanation is neuronal replay, which refers to a time-compressed reactivation of experienced place cell sequences happening during rest or sleep 25 , 26 , 27 . Whether an analogous single-neuron mechanism exists in humans during the encoding of non-spatial relations remains largely unknown. We looked for triplets of selective hippocampal–entorhinal neurons whose preferred stimuli mapped onto three-node trajectories experienced during exposure phases (Fig. 4c ). Each triplet consisted of a neuron selective to an image (‘seed neuron’), a second neuron selective to a directly linked image (‘direct neuron’), and a third neuron that was selective to an indirectly linked image (‘indirect neuron’). Putative replays were defined as consistent firing of the direct and indirect neurons within 30 ms after the seed neuron’s spike. In pyramid-congruent replays, the direct neuron should fire before the indirect one. By contrast, during incongruent replays, which we used as a control condition, the indirect neuron would fire first (Fig. 4c ). Importantly, this analysis used only data recorded during breaks (B1, a break after PRE; B2–B7, breaks after each exposure phase). We found that the proportion of congruent replays significantly increased during the course of learning, whereas the proportion of incongruent replays did not change significantly (Fig. 4d ; congruent: P  < 0.001; incongruent: P  = 0.195; 1,000 permutations of the ‘direct’ and ‘indirect’ spike labels). The above findings bridge the gap between behavioural and synaptic timescales and demonstrate that the neural representations of spatial and temporal structures rely on similar neurophysiological mechanisms.

The human experience is the integration of events characterized by objects with spatial and temporal coordinates—the ‘what’, ‘where’ and ‘when’ of information processing performed by the brain. In the present study, we examined the neural integration of the ‘what’ and ‘when’ of human experience to encode the underlying temporal structure of events. We find that such integration is a process explicitly expressed in the activity of neurons in the hippocampal–entorhinal system, albeit largely implicitly by participant’s awareness. Responses of these neurons scaled with distances between respective nodes of the spatiotemporal graph, thus reflecting the relational contingencies between events characterizing the experience and enabling the predictive representation of expected future states. This neuronal ensemble developed relatively rapidly during the study and remained even when the temporal structure was no longer present. The pyramid graph was extracted directly from experience, without explicitly instructing the participants, and it was abstracted away from the specifics of the task, such as image orientation or behavioural responses.

Our findings provide important insights into the fundamental question of how the human brain forms temporal associations, a critical component in the encoding of episodic memories. Only recently, studies have begun to reveal how this process is implemented by individual neurons in the human MTL. It was demonstrated that cells initially responding only to the picture of a given person started firing to the picture of a given place as a result of the experimental simultaneous pairing of the ‘what’ and ‘where’ 28 . It was also shown that the degree of subjectively reported association between two objects could be successfully predicted from the neurons’ responses 29 . The above evidence, combined with results from animal studies 30 , suggests that the MTL has a critical role in the encoding of relational knowledge 31 . The present study extends this view by demonstrating that hippocampal–entorhinal neurons dynamically embed a complex matrix of ‘what’ and ‘when’ contingencies, by precisely scaling their firing rates to the temporal distance between events during sequential experience.

The present study is also in line with the idea that the hippocampal–entorhinal system is critically involved in the abstraction of knowledge. Such abstraction has been described as a cognitive map in the context of spatial navigation 6 , 7 and ‘schemas’ or ‘learning sets’ in the context of human behaviour and memory research 32 , 33 . Recent computational research suggests that the brain implements similar neural mechanisms to extract the underlying structure of spatial as well as non-spatial problems and that the integration of ‘what’, ‘where’ and ‘when’ is essential for this process 7 , 34 , 35 . The temporal relational neurons that we identify here in human participants during a non-spatial task, have important implications for the hippocampal–entorhinal system as a neural substrate of the cognitive map.

Arguably the main purpose of extracting the underlying structure of temporal sequences is to predict what is likely to happen next in order to choose appropriate actions and maximize reward 17 , 18 , 19 , 20 . A recent computational study showed that neuronal firing patterns that are classically attributed to the encoding of space, such as place cells and grid cells, can be modelled using a predictive successor representation of likely future states, which accounted for a range of empirical findings that cannot be explained by purely Euclidean or geodesic representations 17 . Furthermore, the successor representation can be simulated with neural phenomena that are known to exist in the hippocampal–entorhinal formation, such as the theta phase precession and spike-timing dependent plasticity 20 . Our finding that the neuronal representation of the pyramid graph resembled the successor representation provides the human single-neuron evidence supporting the predictive nature of the hippocampal–entorhinal system function.

The human single-neuron methodology implemented in this study provided a unique window into the possible mechanisms by which the neuronal reorganization occurred during a temporally structured experience. One such mechanism that we demonstrate here is the experience-dependent replay of neuronal firing of specific hippocampal–entorhinal cells taking place between experiment phases. These findings extend previous evidence from rodent studies by showing that encoding of temporal relations between abstract objects in humans engages mechanisms similar to the encoding of spatial trajectories 25 , 26 , 27 , 36 . These results also expand existing evidence from human studies in which replay has been tested more indirectly, by comparing general patterns of neural activity during and after a given experience 37 , 38 , 39 , 40 , 41 or by detecting ‘sharp-wave ripples’ that in rodents often co-occur with replay of individual neurons 42 .

In this study, the neural pyramid topology developed spontaneously from the mere observation of a temporal sequence, without the participants’ detailed explicit knowledge of existent regularity. This finding is consistent with a growing body of evidence that the MTL has a key role in the implicit learning of statistical patterns which does not require deliberate intention or cognitive effort 1 , 43 , 44 , 45 . For example, a recent study 45 using human intracranial electroencephalography found that early cortical processing tracked individual syllables, whereas the hippocampus encoded the ordinal position and identity of pseudowords. The present study demonstrates how individual neurons in the human hippocampal–entorhinal system may encode such implicit structure of temporal associations between serial elements of information.

The probabilities of inner-inner and outer-inner node transitions did not differ significantly, so there is no reason to assume that the transition rates determined the strength of respective associations (Extended Data Fig. 8a ). However, the inner nodes were presented more frequently during exposure phases than the outer nodes (Extended Data Fig. 8b ). This is a natural consequence of the pyramid structure combined with a random walk policy, which happens to mimic many real-life situations (for example, central hubs of a metro system are visited more often than peripheral ones) and experimental setups (for example, a T-maze). However, one could argue that some neurons gradually increased or decreased their firing rate simply owing to stimulus familiarity, which would affect the neural distances between respective nodes. We found that neither relational, selective nor all detected hippocampal–entorhinal neurons responded significantly differently to the inner versus outer nodes (Bayes factors supported the null hypotheses; Extended Data Fig. 8c ). In fact, the proportion of hippocampal–entorhinal neurons that significantly increased or decreased their responses to the inner or outer nodes did not significantly differ from chance level (that is, we analysed responses of each hippocampal–entorhinal neuron to all inner or all outer nodes in E5 and E6 versus E1 and E2; n  = 10 and n  = 6, respectively; 2% and 1% of all hippocampal–entorhinal neurons, respectively; P  > 0.99 and P  = 0.967, respectively; analysis analogous to the Extended Data Fig. 2 ). Furthermore, we replicated all principal findings of this study when the analysis included only inner or only outer nodes (Extended Data Fig. 8d–g ). Thus, stimulus familiarity did not drive our main results. Future studies focusing on how different transition strategies affect the geometry of neuronal representations could manipulate this aspect by using a random walk versus Hamiltonian cycles or other policies.

One might ask whether the current design allows us to disambiguate between distance-dependent scaling and the formation of simple pairwise associations, since every pair of nodes that was not a direct link on the pyramid automatically was two links apart. However, if multiple respective links were not scaled according to a common metric (distance), it would not be possible to recover the entire pyramid graph from the neuronal population activity, especially not the successor representation where various direct and indirect links have different lengths (Extended Data Fig. 9 ). Such a reconstruction was possible in the present study (Fig. 3f ). To further address this point, we collected data from five additional patients (7 sessions; 221 neurons) with a diamond-shaped graph where links to indirect stimuli varied between two and three edges (D2 and D3, respectively). We found that, during late-exposure phases, hippocampal relational neurons responded more strongly to images located two edges away from their preferred stimulus than to images located three edges away. We also replicated population decoding results from the main study and showed that the representational overlap was greater for D2 stimuli than for D3 (Extended Data Fig. 10 ). The above evidence supports distance-dependent scaling in the encoding of the temporal structure of the sequence.

Together, the findings of this study reveal multiple similarities between the neurophysiological properties of individual cells representing locations in physical space and neurons encoding abstract objects in a temporal sequence structure; these parallels include reorganization and functional overlap of representations of adjacent states, experience-dependent and predictive modulation of receptive fields, as well as offline replay of individual neurons’ activity congruent with past experience. Thus, the human brain appears to be using analogous mechanisms to represent seemingly very different types of information: relations in space and time. The remarkable entorhinal–hippocampal neuronal machinery likely evolved to form scalable and partly non-Euclidean (‘warped’) representations of space-time trajectories to enable learning and prediction, necessary for the organism’s survival. Here, keeping space constant, we demonstrate at the neuronal level how such representations of object trajectories in time are incorporated by the human entorhinal–hippocampal system.

Participants

The participants were 17 patients with intractable epilepsy who were implanted with depth electrodes to delineate a potentially surgically treatable epileptogenetic zone. Demographics information and neuropsychological scores are presented in Supplementary Table 1 . Electrode placements were determined solely on the basis of clinical treatment criteria. The follow-up studies (Extended Data Figs. 1 and 10 ) included 33 healthy controls (26 female participants; mean age: 31 ± 7 years old) and 5 additional patients with epilepsy (2 female participants; mean age: 38 ± 12 years old). All participants volunteered for the study by providing informed consent according to a protocol approved by the UCLA Medical Institutional Review Board (IRB).

Neural recordings

Patients were stereotactically implanted with 7–12 Behnke-Fried electrodes with 40-µm diameter microwire extensions (eight high-impedance recording wires and one low-impedance reference wire per depth electrode) that capture local field potentials and extracellular spike waveforms 46 . Microwire electrophysiology data were amplified and recorded at 30 kHz on a Blackrock Microsystems recording system or at 32 kHz on a Neuralynx recording system (Cheetah 5.0).

Microelectrode localizations

Prior to data collection, each microelectrode location was confirmed by an expert neurosurgeon (I.F.) based on the patient’s postoperative computed tomography (CT) scan with visible electrode artifacts overlaid on a co-registered preoperative T1 structural MRI (BrainLab software). For descriptive purposes (Fig. 1d ), we additionally used the following procedure to transform locations from each participant’s ‘native brain space’ to the standard Montreal Neurological Institute (MNI) space. First, each participant’s MRI and CT images were co-registered using the FSL ‘flirt’ function. Second, the MRI image was: (1) segmented into the grey matter, white matter, and cerebrospinal fluid probability maps; (2) resampled (1 × 1 × 1 mm voxel size); and (3) normalized to the 152 T1-weighted MNI template using the nonlinear transformation algorithm implemented in the Statistical Parametric Mapping toolbox (SPM12, Wellcome Department of Cognitive Neurology, London, UK). Third, the same transformation parameters were applied to the participant’s CT image. MNI coordinates for each microelectrode were extracted manually from the normalized CT overlaid on the normalized MRI from a given participant using the FSLeyes software.

General procedure

Before the main experiment (typically 1–2 days prior), a screening experiment was conducted to find 6 stimuli (images of people) associated with robust and preferential responses of single neurons in the MTL. These six images were then used during the main experimental task (Fig. 1b ), which was introduced to the patients as a follow-up of the screening study without mentioning that the stimuli would be presented in a specific order. At the end of the main experiment, we asked the participants to answer the following questions: “Have you noticed any pattern in the sequence of images shown in any of the phases? If yes, what was it?”; “Have you had any special strategy during this study?”. None of the participants reported noticing any pattern that was relevant to the experimental manipulation (Fig. 1b ). We used the Psychophysics Toolbox to control the timings of stimuli presentation and register behavioural responses 47 .

Screening session

During screening, approximately 120 images were repeatedly shown to the patients on a laptop computer (taking around 40 min). These images showed people, animals, objects and landmarks that were partly selected based on the participant’s preferences (for example, favourite actors, musicians, places, etc.). The experiment consisted of eight blocks, each with a different instruction (for example, block 1: “Determine whether each image shows a person or not”; block 2: “Determine whether each image shows a plant or not”; etc.). Each image was presented exactly once during each block, for the duration of 1 s, against a black background. The order of stimuli presentation was random. Participants indicated their responses using two assigned keys on a hand-held game pad.

Experimental task

The main study consisted of three parts: pre-exposure (PRE), exposure (E1–E6), and post-exposure (POST; Fig. 1c ). During PRE (121 stimuli presented), all images were displayed in a pseudo random sequence (60 direct and 60 indirect graph-transitions; on average, each direct transition was presented 7 times and each indirect transition 9 times). The task was to determine whether each image showed a male or female (gender task). The participants used the right and left arrow keys on a laptop keyboard to indicate their responses. During the six subsequent exposure phases (121 trials in each phase), the order of stimuli was still randomized but restricted by the topological structure of the pyramid graph (Fig. 1c ) so that only images directly linked on the graph were shown immediately after another. The starting location was selected randomly in each experiment phase. The behavioural task during all exposure phases was to determine whether a given image was mirrored or not when compared to PRE (Fig. 1c ; the participants used the right and left arrow keys on a laptop keyboard to indicate their responses). During each phase, 61 images were ‘normal’ and 60 were ‘mirrored’. The order of mirrored and normal images was random. The POST phase was the same as PRE (all stimuli presented in a pseudo random sequence, without the ‘pyramid rule’; on average, each direct transition was presented 7 times and each indirect transition 8 times). Behavioural instructions displayed in the beginning of each phase emphasized that the participants should try to respond as quickly and accurately as possible. The first trial in each phase (that is, the beginning of a sequence) was discarded from the analyses, so effectively each phase consisted of 120 trials. The experiment in all phases was self-paced, that is: (1) a given image was displayed for as long as it took the participant to respond; and (2) the participants could have had breaks between phases for as long as they needed. All stimuli were displayed against a grey background. During a randomized inter-trial interval (1-3 s), a black ‘fixation’ circle was displayed in the middle of the screen. After each stimulus presentation, the participants received feedback (“correct!” or “incorrect” in relation to the currently performed task) displayed for 500 ms. All trials (correct and incorrect) were included in the analysis of electrophysiological data, as the behavioural tasks were unrelated to the main research question. Behavioural accuracy of responses during PRE and POST was near-perfect indicating that the gender task was easy for all the participants (Extended Data Fig. 1a ). Accuracy in the ‘mirror task’ was lower but improved over the course of the study (Extended Data Fig. 1a ; this task was supposed to be more challenging to maintain the participants’ attention).

Spike sorting

Automated spike detection and sorting were performed using the WaveClus3 software package in MATLAB 48 . We then manually reviewed each unit for inclusion by evaluating the waveform’s shape, amplitude, inter-spike intervals, and firing consistency across study phases. We rejected units that were likely contaminated by artifacts, in keeping with field-standard spike evaluation criteria 49 . For electrodes with multiple putative units that passed this inclusion check, we merged units whose waveform features could not be well-separated in principal components space, retaining for analysis a combination of single- and multi-units.

Single-neuron analyses

For each neuron and each stimulus presentation, we selected a time window around the stimulus onset (from −1 to +2 s). Then we calculated the number of spikes in 0.1 s time bins, smoothed (moving sum: ± 0.25 s) and baseline-corrected the data (subtracted the mean activity in the −0.5 to 0 s time window). The ‘response window’ was defined from 0.1 to 1.2 s after the stimulus onset. For a given neuron, the ‘preferred stimulus’ was the image associated with the strongest mean response in the response window during PRE. Depending on the position of the preferred stimulus on the pyramid, the remaining images were labelled as ‘direct’ or ‘indirect’ (Fig. 1b ). This assignment was used across all study phases. ‘Selective neurons’ were defined as cells that during PRE: (1) responded significantly stronger to the preferred stimulus in the response window versus baseline; and (2) responded significantly stronger to the preferred stimulus than to the remaining stimuli combined. ‘Relational neurons’ were defined as cells that: (1) were selective (see above); (2) responded significantly stronger to the direct than indirect stimuli during E5 and E6; and (3) responded significantly stronger to direct stimuli during E5 and E6 than during E1 and E2. The ‘diminishing selectivity neurons’ were defined as cells that: (1) were selective; and (2) responded significantly weaker to the preferred stimulus in E5 and E6 than in E1 and E2. All the above criteria were tested with the Wilcoxon signed-rank tests (one-sided) with a P value threshold of 0.05. The above procedure was repeated 1,000 times, with random permutations of the stimulus or phase labels, depending on which criterion was tested. These permutations informed how many neurons of a given type are expected in a given brain region by chance. The empirical P value was calculated as the number of permutations with more neurons of a given type than the number of neurons actually detected, divided by the total number of permutations. If this value was less than 0.05, we concluded that a given brain region contained a significant proportion of a given neuron type (Extended Data Fig. 2 and Supplementary Table 3 ). To analyse combined responses of all relational neurons (Extended Data Fig. 4 ), we calculated the difference between each neuron’s mean responses to direct minus indirect stimuli and preferred minus non-preferred stimuli. This was done for each study phase separately. Then, we peak-normalized and baseline-corrected (−0.5 to 0 s) those differences and extracted the mean from the 0.1 to 1 s time window. For line plots showing the mean responses of individual neurons (Fig. 2a,b and Extended Data Figs. 3a and 9b ), we used 0.01 s bins and the ±0.25 s moving sum. For plots showing multiple neurons (Figs. 1f and 2c ), we z -scored and baseline-corrected (−0.5 to 0 s) the data from each neuron (for illustration purposes, we used the ± 0.2 s moving sum and heat maps were additionally smoothed with ± 0.1 s moving average).

Neural population analyses

To decode stimulus identity during each image presentation, we used the Poisson naive Bayes classifier, as implemented in the Neural Decoding Toolbox 50 . The spiking activity of each neuron was extracted from the −1 to +2 s time window relative to the stimulus onset. Data was binned (0.1 s) and smoothed (moving sum: ±0.25 s). The decoder was run on the summed spiking activity in the 0.1 to 1 s time window (Extended Data Fig. 5 ). The main analysis focused on posterior probabilities assigned by the decoder to the image actually presented (actual), images directly linked to that stimulus on the graph (direct), and images linked indirectly (indirect). The analysis was performed for each recording session separately (different stimuli), but the resulting probability distributions were combined across all sessions and image presentations. The classifier was trained on the data from PRE and tested on all subsequent phases. For testing in PRE, we used the ‘leave-one-trial-out’ cross-validation. Kolmogorov–Smirnov tests were used to compare cumulative distribution functions (CDFs) of posterior probabilities (one-sided). To reconstruct the entire pyramid graph (Fig. 3f,g and Extended Data Figs. 6b and 7b ), we calculated Euclidean distances between mean responses of each neuron to each pair of images across all relevant neurons (neurons that stopped firing during the late study phases were excluded; distances were z -scored; bin-size: 0.1 s; baseline-correction: −0.5 to 0 s; moving sum: ± 0.15 s; time window: 0.1 to 1 s). Then, we compared the resulting neural distance matrixes to three templates (Fig. 3f ). In the geodesic template, distances between each pair of nodes corresponded to the number of edges of the shortest path connecting the nodes. In the Euclidean template, distances between nodes 1–5, 4–3 and 6–2 were calculated from the Pythagorean theorem (right triangles: 1–5–6, 4–3–1, 6–2–1). The remaining distances corresponded to the shortest path (see above). In line with the previous literature 2 , 51 , the successor template (ST) was calculated as the negative of the matrix exponential of the adjacency matrix A :

The above metric provided a slightly better fit to the data than a related index that defines the relationships between states (Extended Data Fig. 7b ):

Here, entries a ij for each A n correspond to the number of possible paths of length n between objects i and j and a discount factor is 0 <  γ  < 1 (refs. 2 , 19 ). To illustrate most faithful 2D representations of the respective distance matrixes, we used the multidimensional scaling analysis (MDS; ‘mdscale’ function in MATLAB; criterion: ‘sammon’). Because MDS can only be performed on matrices with positive entries, we normalized the matrixes by adding the absolute value of the matrix’s minimum plus a constant of 0.1. The similarity between neural distance matrixes and each template was calculated as the Spearman correlation (Fisher-transformed). Because the aim was to test how this similarity changes over the course of the study (unconfounded by any potential pre-existing similarity), for each phase, we subtracted the degree of similarity in PRE (Fig. 3g and Extended Data Figs. 6b and 7b ). To obtain null distributions of correlation coefficients, the above procedure was repeated 10,000 times with random permutations of the nodes’ positions. P values were calculated as the number of permutations with higher correlation coefficients than the one actually detected, divided by the total number of permutations. If in none of the permutations the correlation was above the actual value, the P  < 0.0001 range is reported.

Replay analysis

We analysed sessions that contained at least three selective hippocampal–entorhinal neurons from the same hemisphere, whose preferred stimuli from PRE mapped onto three-element pyramid trajectories (Fig. 4c , one seed neuron, one direct neuron and one indirect neuron forming a connected path). For each spike of the seed neuron, we checked whether the direct and indirect neurons fired at least once in the 0 to 30 ms time window. The above situation had to occur at least five times to be included in the analysis (that is, n  < 5 was considered insufficient for robust statistical inference). There were 536 such putative replays in B1 (break after PRE) and 1,012 in B2–B7 (breaks after exposure phases). If the direct neuron fired significantly earlier than the indirect neuron, the replay was labelled ‘congruent’ (Fig. 4c ; we analysed latencies of the first ‘direct spikes’ versus latencies of the first ‘indirect spikes’ across all repetitions of a given replay 52 ; Wilcoxon signed-rank test, one-sided, with a P value threshold of 0.05). If the opposite was true, a replay was labelled ‘incongruent’. To obtain P values for the comparisons between proportions of congruent and incongruent replays throughout the study, we randomly shuffled spikes from the direct and indirect neurons (1,000 permutations) and used the resulting null distribution as reference.

Reporting summary

Further information on research design is available in the  Nature Portfolio Reporting Summary linked to this article.

Data availability

Owing to ethical considerations and protection of patients’ confidentiality, data supporting the results of this study are available from the corresponding authors upon a reasonable request—that is, for collaborative research by researchers, adhering to protocols approved by the Institutional Review Board.

Code availability

Owing to ethical considerations and protection of patients’ confidentiality, code supporting the results of this study is available from the corresponding authors upon a reasonable request—that is, for collaborative research by researchers, adhering to protocols approved by the Institutional Review Board.

Schapiro, A. C., Rogers, T. T., Cordova, N. I., Turk-Browne, N. B. & Botvinick, M. M. Neural representations of events arise from temporal community structure. Nat. Neurosci. 16 , 486–492 (2013).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Garvert, M. M., Dolan, R. J. & Behrens, T. E. J. A map of abstract relational knowledge in the human hippocampal–entorhinal cortex. eLife 6 , e17086 (2017).

Article   PubMed   PubMed Central   Google Scholar  

Bellmund, J. L., Deuker, L. & Doeller, C. F. Mapping sequence structure in the human lateral entorhinal cortex. eLife 8 , e45333 (2019).

Baram, A. B. et al. Entorhinal and ventromedial prefrontal cortices abstract and generalize the structure of reinforcement learning problems. Neuron 109 , 713–723.e7 (2021).

Garvert, M. M., Saanum, T., Schulz, E., Schuck, N. W. & Doeller, C. F. Hippocampal spatio-predictive cognitive maps adaptively guide reward generalization. Nat. Neurosci. 26 , 615–626 (2023).

Tolman, E. C. Cognitive maps in rats and men. Psychol. Rev. 55 , 189–208 (1948).

Article   CAS   PubMed   Google Scholar  

Whittington, J. C. R., Mccaffary, D., Bakermans, J. J. W. & Behrens, T. E. J. How to build a cognitive map. Nat. Neurosci. 25 , 1257–1272 (2022).

O’Keefe, J. & Dostrovsky, J. The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Res. 34 , 171–175 (1971).

Article   PubMed   Google Scholar  

Ekstrom, A. D. et al. Cellular networks underlying human spatial navigation. Nature 425 , 184–188 (2003).

Article   ADS   CAS   PubMed   Google Scholar  

Hafting, T., Fyhn, M., Molden, S., Moser, M.-B. & Moser, E. I. Microstructure of a spatial map in the entorhinal cortex. Nature 436 , 801–806 (2005).

Jacobs, J. et al. Direct recordings of grid-like neuronal activity in human spatial navigation. Nat. Neurosci. 16 , 1188–1190 (2013).

Aronov, D., Nevers, R. & Tank, D. W. Mapping of a non-spatial dimension by the hippocampal–entorhinal circuit. Nature 543 , 719–722 (2017).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Fried, I., Macdonald, K. A. & Wilson, C. L. Single neuron activity in human hippocampus and amygdala during recognition of faces and objects. Neuron 18 , 753–765 (1997).

Constantinescu, A. O., O’Reilly, J. X. & Behrens, T. E. J. Organizing conceptual knowledge in humans with a gridlike code. Science 352 , 1464–1468 (2016).

Pastalkova, E., Itskov, V., Amarasingham, A. & Buzsáki, G. Internally generated cell assembly sequences in the rat hippocampus. Science 321 , 1322–1327 (2008).

MacDonald, C. J., Lepage, K. Q., Eden, U. T. & Eichenbaum, H. Hippocampal “time cells” bridge the gap in memory for discontiguous events. Neuron 71 , 737–749 (2011).

Stachenfeld, K. L., Botvinick, M. M. & Gershman, S. J. The hippocampus as a predictive map. Nat. Neurosci. 20 , 1643–1653 (2017).

Gershman, S. J. The successor representation: its computational logic and neural substrates. J. Neurosci. 38 , 7193–7200 (2018).

Momennejad, I. Learning structures: predictive representations, replay, and generalization. Curr. Opin. Behav. Sci. 32 , 155–166 (2020).

George, T. M., de Cothi, W., Stachenfeld, K. & Barry, C. Rapid learning of predictive maps with STDP and theta phase precession. eLife 12 , e80663 (2023).

Nagahama, Y. et al. Outcome of stereo-electroencephalography with single-unit recording in drug-refractory epilepsy. J. Neurosurg. 139 , 1588–1597 (2023).

Quian Quiroga, R., Reddy, L., Kreiman, G. & Koch, C. Invariant visual representation by single neurons in the human brain. Nature 435 , 1102–1107 (2005).

Article   ADS   PubMed   Google Scholar  

Suthana, N. & Fried, I. Percepts to recollections: insights from single neuron recordings in the human brain. Trends Cogn. Sci. 16 , 427–436 (2012).

Bi, G.-Q. & Poo, M.-M. Synaptic modifications in cultured hippocampal neurons: Dependence on spike timing, synaptic strength, and postsynaptic cell type. J. Neurosci. 18 , 10464–10472 (1998).

Skaggs, W. E. & McNaughton, B. L. Replay of neuronal firing sequences in rat hippocampus during sleep following spatial experience. Science 271 , 1870–1873 (1996).

Wilson, M. A. & McNaughton, B. L. Reactivation of hippocampal ensemble memories during sleep. Science 265 , 676–679 (1994).

Foster, D. J. & Wilson, M. A. Reverse replay of behavioural sequences in hippocampal place cells during the awake state. Nature 440 , 680–683 (2006).

Ison, M. J., Quian Quiroga, R. & Fried, I. Rapid encoding of new memories by individual neurons in the human brain. Neuron 87 , 220–230 (2015).

De Falco, E., Ison, M. J., Fried, I. & Quian Quiroga, R. Long-term coding of personal and universal associations underlying the memory web in the human brain. Nat. Commun. 7 , 13408 (2016).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Dusek, J. A. & Eichenbaum, H. The hippocampus and memory for orderly stimulus relations. Proc. Natl Acad. Sci. USA 94 , 7109–7114 (1997).

Eichenbaum, H. & Cohen, N. J. Can we reconcile the declarative memory and spatial navigation views on hippocampal function? Neuron 83 , 764–770 (2014).

Bartlett, F. C. Remembering: A Study in Experimental and Social Psychology (Cambridge Univ. Press, 1932).

Harlow, H. F. The formation of learning sets. Psychol. Rev. 56 , 51–65 (1949).

Behrens, T. E. J. et al. What is a cognitive map? Organizing knowledge for flexible behavior. Neuron 100 , 490–509 (2018).

Whittington, J. C. R. et al. The Tolman-Eichenbaum machine: unifying space and relational memory through generalization in the hippocampal formation. Cell 183 , 1249–1263.e23 (2020).

Ólafsdóttir, H. F., Bush, D. & Barry, C. The role of hippocampal replay in memory and planning. Curr. Biol. 28 , R37–R50 (2017).

Article   Google Scholar  

Eichenlaub, J.-B. et al. Replay of learned neural firing sequences during rest in human motor cortex. Cell Rep. 31 , 107581 (2020).

Liu, Y., Dolan, R. J., Kurth-Nelson, Z. & Behrens, T. E. J. Human replay spontaneously reorganizes experience. Cell 178 , 640–652 (2019).

Schwartenbeck, P. et al. Generative replay underlies compositional inference in the hippocampal-prefrontal circuit. Cell 186 , 4885–4897 (2023).

Schapiro, A. C., Mcdevitt, E. A., Rogers, T. T., Mednick, S. C. & Norman, K. A. Human hippocampal replay during rest prioritizes weakly learned information and predicts memory performance. Nat. Commun. 9 , 3920 (2018).

Vaz, A. P., Wittig, J. H., Inati, S. K. & Zaghloul, K. A. Replay of cortical spiking sequences during human memory retrieval. Science 367 , 1131–1134 (2020).

Norman, Y. et al. Hippocampal sharp-wave ripples linked to visual episodic recollection in humans. Science 365 , 1–14 (2019).

Schapiro, A. C., Kustner, L. V. & Turk-Browne, N. B. Shaping of object representations in the human medial temporal lobe based on temporal regularities. Curr. Biol. 22 , 1622–1627 (2012).

Schapiro, A. C., Gregory, E., Landau, B., McCloskey, M. & Turk-Browne, N. B. The necessity of the medial temporal lobe for statistical learning. J. Cogn. Neurosci. 26 , 1736–1747 (2017).

Henin, S. et al. Learning hierarchical sequence representations across human cortex and hippocampus. Sci. Adv. 7 , eabc4530 (2021). 19.

Fried, I. et al. Cerebral microdialysis combined with single-neuron and electroencephalographic recording in neurosurgical patients. J. Neurosurg. 91 , 697–705 (1999).

Brainard, D. H. The psychophysics toolbox. Spat. Vision 10 , 433–436 (1997).

Article   CAS   Google Scholar  

Quiroga, R. Q., Nadasdy, Z. & Ben-Shaul, Y. Unsupervised spike detection and sorting with wavelets and superparamagnetic clustering. Neural Comput. 16 , 1661–1687 (2004).

Hill, D. N., Mehta, S. B. & Kleinfeld, D. Quality metrics to accompany spike sorting of extracellular signals. J. Neurosci. 31 , 8699–8705 (2011).

Meyers, E. M. The neural decoding toolbox. Front. Neuroinform. 7 , 8 (2013).

Estrada, E. & Hatano, N. Communicability in complex networks. Phys. Rev. E 77 , 036111 (2008).

Article   ADS   MathSciNet   Google Scholar  

Guyonneau, R., VanRullen, R. & Thorpe, S. J. Neurons tune to the earliest spikes through STDP. Neural Comput. 17 , 859–879 (2005).

Download references

Acknowledgements

The authors thank the participants for their involvement; Z. Aghajan, E. Mankin and S. Nichterwitz for their insightful comments and practical help; and N. Cherry, C. Dao, A. Hampton and A. Rangel for assisting in data collection and preprocessing. This work was supported by the Marie Sklodowska-Curie fellowship 750955 (P.T.), National Institute of Neurological Disorders and Stroke grant (R01NS084017, U01 grants NS108930 and NS123128) (I.F.), and the Foundation for Science and Technology grant 2022.03499.CEECIND (P.T.).

Author information

Güldamla Kalender

Present address: Department of Psychiatry and Behavioral Sciences, University of Minnesota, Minneapolis, MN, USA

Davide Ciliberti

Present address: Institute of Cognitive Neurology and Dementia Research, Otto-von-Guericke University Magdeburg, Magdeburg, Germany

Authors and Affiliations

Department of Neurosurgery, University of California, Los Angeles, Los Angeles, CA, USA

Pawel Tacikowski, Güldamla Kalender, Davide Ciliberti & Itzhak Fried

Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden

Pawel Tacikowski

Coimbra Institute for Biomedical Imaging and Translational Research, University of Coimbra, Coimbra, Portugal

Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, Los Angeles, CA, USA

Itzhak Fried

Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

You can also search for this author in PubMed   Google Scholar

Contributions

Conceptualization: P.T. and I.F. Formal analysis: P.T. and G.K. Funding acquisition: I.F. and P.T. Surgeries: I.F. Data acquisition: P.T., G.K. and I.F. Methodology: I.F., P.T. and D.C. Supervision: I.F. Visualization: P.T. Manuscript writing: P.T. and I.F. All authors provided critical review and commented on the manuscript.

Corresponding authors

Correspondence to Pawel Tacikowski or Itzhak Fried .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature thanks Mona Garvert and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables

Extended data fig. 1 behavioural performance..

( a ) The participants responded more slowly and committed more errors in the ‘mirror task’ than in the ‘gender task’ (combined data from E1-E6 versus combined data from PRE and POST; P  = 5.96 × 10 −5 and P  = 5.95 × 10 −5 , respectively; n  = 21 sessions; Wilcoxon signed-rank tests; two-sided). ( b ) During exposure, the outer nodes could never immediately follow one another. If the participants extracted this rule, outer-after-outer trials (OaO) during POST should be unexpected and thus related to longer reaction times (RTs) than outer-after-inner trials (OaI). Indeed, we found that the difference between RTs in OaO minus OaI trials increased in POST compared to PRE ( P  = 0.0445; 10,000 permutations of PRE and POST labels). The plot shows means ± s.e.m. Circles correspond to datapoints from individual sessions ( n = 21). The P -value was calculated as a number of permutations with a higher difference than the one actually detected, divided by the total number of permutations. ( c ) The above-mentioned behavioural conflict was especially pronounced among the patients who developed a robust hippocampal-entorhinal representation of the pyramid, calculated per participant as similarity between neuronal population responses and the successor template in POST (see Fig. 3g ; ρ 19  = 0.53; P  = 0.0077; Spearman correlation with 10,000 permutations of the session order). Another index of behavioural conflict—RTs in trials after ‘indirect’ transitions—was also positively correlated with the strength of the hippocampal-entorhinal successor representation (indirect trials: POST minus PRE; ρ 19  = 0.39; P  = 0.04; Spearman correlation with 10,000 permutations of the session order). In contrast, RTs in trials without conflict (i.e., ‘direct’ and OaI transitions; POST minus PRE) did not significantly correlate with the strength of the hippocampal-entorhinal successor representation (ρ 19  = 0.32; P  = 0.0814 and ρ 19  = 0.21; P  = 0.173, respectively). All RT analyses were performed on correct trials only. Very short (<200 ms) and very long (> 5000 ms) RTs were discarded. The above P values were calculated as a number of permutations with a higher correlation coefficient than the one actually detected, divided by the total number of permutations. ( d ) Twenty-five healthy controls (see  Methods ) completed the same behavioural procedure as the patients. They were then asked an open question: ‘Have you noticed any pattern in the sequence of images?’ None of the participants reported noticing a graph-like organization of the sequence. Then, we informed them about the underlying structure and asked them to assign each image to a specific node (the ‘positions’ task). The pyramid has six variants (three rotations and two flips). The ‘positions accuracy’ was calculated as the maximum number of hits. For example, if someone’s highest score was three out of six for one variant and less than three hits for other variants, this person’s accuracy score was 50%. To calculate the ‘links accuracy,’ we checked whether each pair of images was linked directly or indirectly on the graph provided by each participant and compared it to the actual pyramid. Similarly, we calculated Spearman correlation coefficients (Fisher-transformed) between pairwise distances provided by each participant and the actual pairwise distances (‘distance similarity’ index). Finally, we checked how often the participants assigned the correct images to the inner versus outer nodes (‘inner-outer accuracy’). For each index, the chance level was estimated as the mean performance of 10,000 randomly generated ‘participants’ (red dashed line). To establish the ‘explicit benchmark’ (blue dashed line), we tested another eight control participants (see  Methods ). From the beginning, we informed them that the sequence of images during exposure phases will follow the pyramid graph, but we did not explain which image is located where on the graph. All other aspects of the procedure and analysis were the same as explained above. The mean performance of this additional group served as the explicit benchmark. We found that ‘positions accuracy’ did not significantly differ from chance level and was significantly below the explicit benchmark ( P  = 0.1584 and P  = 8.54 × 10 −6 , respectively). Other indexes, arguably referring to less detailed knowledge of the graph, were significantly above chance level, but still below the explicit benchmark (‘links accuracy’ and ‘distance similarity’ versus chance: P  = 0.007; ‘links accuracy’ and ‘distance similarity’ versus explicit: P  = 7.04 × 10 −6 ; ‘inner-outer accuracy’ versus chance: P  = 0.0186; ‘inner-outer accuracy’ versus explicit: P  = 0.0004). All the above P values are from the Wilcoxon signed-rank tests (two-sided). Together, these results suggest that the healthy control participants (and patients) did not have detailed explicit knowledge of the pyramid. The central marks of the box plots (panels a and d) indicate the medians. The bottom and top edges of the boxes indicate the 25th (Q1) and 75th (Q3) percentiles, respectively. The whiskers extend to the most extreme data points not considered outliers (1.5 × interquartile range above Q3 or below Q1).

Extended Data Fig. 2 Proportion of selective, relational, and diminishing selectivity neurons across the brain regions.

The number of ‘selective neurons’ was significantly above chance level in the hippocampus (H), entorhinal cortex (E), and parahippocampal gyrus (PH). The number of ‘relational neurons’ was significantly above chance only in the hippocampus (H) and entorhinal cortex (E). The number of ‘diminishing selectivity neurons’ was significantly above chance level in the hippocampus (H), entorhinal cortex (E), insula (I), and orbitofrontal cortex (OF). Histograms show the number of units of a given type obtained from 1,000 random permutations of stimuli or phase labels (see  Methods ). P values were calculated as the number of permutations with more neurons of a given type than the number of neurons actually detected, divided by the total number of permutations. If in none of the permutations the number of neurons was above the actual count, the P  < 0.001 range is reported. No adjustment for multiple comparisons was applied. Dashed lines represent the actual number of neurons detected. Abbreviations: A – amygdala; E – entorhinal cortex; H – hippocampus; I – insula and operculum; IT – inferior temporal cortex; LT – lateral temporal cortex; OF – orbitofrontal cortex and anterior cingulate cortex; O – occipital cortex; P&C – parietal cortex & middle/posterior cingulate cortex; PH – parahippocampal gyrus; SMA – Supplementary Motor Area.

Extended Data Fig. 3 Other examples of relational and diminishing selectivity neurons.

( a ) Throughout the study, these hippocampal relational neurons showed increasingly more robust responses to stimuli that were directly linked to their preferred image on the pyramid graph (magenta) than to images linked indirectly (blue). ( b ) During the study, these diminishing selectivity neurons decreased responses to their preferred stimulus. Raster plots show individual spikes in each trial. Line plots show the mean number of spikes ± s.e.m. Top and middle panels show hippocampal neurons. The bottom panel shows a neuron from the entorhinal cortex.

Extended Data Fig. 4 Average responses of all hippocampal-entorhinal relational neurons.

These neurons showed significantly stronger responses to direct than indirect images in the late-exposure phases (E5&6) and in POST (left panel; Wilcoxon signed-rank tests against PRE; two-sided; FDR-corrected). These neurons also gradually diminished their selectivity to the preferred image (right panel; Wilcoxon signed-rank tests against PRE; two-sided; FDR-corrected). It is highly unlikely that this diminished selectivity was due to ‘regression toward the mean,’ as responses to preferred stimuli continued to decrease in the subsequent phases (E3&4, E5&6, POST versus E1&2; P  = 0.0346, P  = 0.0315, and P  = 0.0271, respectively; Wilcoxon signed-rank tests; one-sided; FDR-corrected). Each circle corresponds to one neuron. For the definition of box plots, please see the legend of the Extended Data Fig. 1 . Solid lines represent means.

Extended Data Fig. 5 Population activity during PRE contained enough information to decode stimulus identity above the chance level.

We used the Poisson Naive Bayes classifier to decode stimulus identity during PRE (‘leave-one-out’ cross-validation). This analysis was performed on data from all 1456 neurons. The highest decoding accuracy was obtained for the moving sum of ± 0.25 s and the 0.1 – 1 s time window. These parameters were used in the remaining analyses.

Extended Data Fig. 6 Additional neuronal population results.

( a ) Population decoding results from all neurons outside of the hippocampal-entorhinal system ( n  = 910; 21 sessions). Plots show CDFs of posterior probabilities assigned to the actual, direct, and indirect images in the subsequent study phases. P values obtained from the Kolmogorov-Smirnov tests between CDFs of subsequent study phases versus PRE (one-sided). Exact P values for ‘actual’ E1&2: 2.68 × 10 −11 ; E3&4: 1.16 × 10 −20 ; E5&6: 1.28 × 10 −21 ; POST: 3.59 × 10 −18 . ( b ) Reconstruction of the pyramid structure from the activity of all non-hippocampal-entorhinal neurons ( n  = 910; 21 sessions). During exposure, but not during POST, all templates fit the data significantly above the chance level (geodesic: E1&2: P  = 0.0286; E3&4: P  = 0.0286; E5&6: P  = 0.0399; POST: P  = 0.7053; Euclidean: E1&2: P  = 0.005; E3&4: P  = 0.005; E5&6: P  = 0.0064; POST: P  = 0.6505; successor: E1&2: P  < 0.0001; E3&4: P  = 0.0003; E5&6: P  = 0.0003; POST: P  = 0.7169; change from PRE; 10,000 permutations; FDR-corrected). The successor template significantly outperformed the geodesic template during exposure phases but not during POST (successor versus geodesic: E1&2: P  = 0.0008; E3&4: P  = 0.0128; E5&6: P  = 0.0022; POST: P  = 0.5216; change from PRE; 10,000 permutations; FDR-corrected). The difference between successor and Euclidean templates was non-significant or marginally significant (successor versus Euclidean: E1&2: P  = 0.0612; E3&4: P  = 0.1653; E5&6: P  = 0.0878; POST: P  = 0.6476; change from PRE; 10,000 permutations; FDR-corrected). The plot shows Spearman’s correlation coefficients (Fisher-transformed) between each template and the neural data from respective phases. P values were calculated as the number of permutations with a higher correlation coefficient than the one actually detected, divided by the total number of permutations. If in none of the permutations the correlation was above the actual coefficient, the P  < 0.0001 range is reported. ( c ) Population decoding results for trials in which the actually presented images were from an outer node of the pyramid (all 546 hippocampal-entorhinal neurons). Plots show CDFs of posterior probabilities assigned to the actual, direct, indirect-inner, and indirect-outer nodes during the subsequent study phases. P values obtained from Kolmogorov-Smirnov tests between CDFs in the subsequent study phases versus PRE (one-sided). Exact P values for ‘actual’ E1&2: 0.2718; E3&4: 3.59 × 10 −7 ; E5&6: 1.98 × 10 −6 ; POST: 4.56 × 10 −13 . Exact P -value for ‘direct’ POST: 1.53 × 10 −5 .

Extended Data Fig. 7 Additional support for the successor model and a comparison of different successor representation metrics.

( a ) An important prediction of the successor representation, but not the geodesic representation, is that direct links between inner nodes (inner-inner) should be shorter than direct links between each outer node and the adjacent inner nodes (outer-inner). We calculated pairwise distances between spiking responses of all selective hippocampal-entorhinal neurons grouped by their preferred node. Then, we compared the mean distances of inner-inner links (2-3, 3-5, 5-2) to the mean distance of outer-inner links (4-2, 4-5, 6-5, 6-3, 1-2, 1-3). As predicted, the former were significantly shorter than the latter ( P values from Wilcoxon rank-sum tests; two-sided; FDR-corrected). The plot shows mean distances from all 263 (outer: n  = 119; inner: n  = 144) selective hippocampal-entorhinal neurons ± s.e.m. ( b ) There were no significant differences between the two measures of the successor representation in terms of similarity to the neuronal data. The plot shows distance matrixes and graphs corresponding to the two measures (see  Methods ). Each graph shows the most faithful 2D representations of the respective distance matrix obtained from the multidimensional scaling analysis. The right panel shows the degree of similarity between data (546 hippocampal-entorhinal neurons) and each template throughout the study (Spearman’s correlation coefficients; Fisher-transformed; change from PRE). P values (FDR-corrected) were calculated as the number of permutations with a higher difference between the templates than the one actually detected, divided by the total number of permutations (10,000).

Extended Data Fig. 8 Stimulus familiarity.

( a ) The mean number of outer-inner or inner-outer (I-O) transitions did not differ significantly from the mean number of inner-inner (I-I) transitions ( P values from the Wilcoxon signed-rank tests; two-sided; FDR-corrected). ( b ) Images from the inner nodes were displayed more frequently during exposure phases than images from the outer nodes ( P values from the Wilcoxon signed-rank tests; two-sided; FDR-corrected; exact P values equal to 5.88 × 10 −5 for all comparisons). ( c ) Hippocampal-entorhinal neurons did not respond significantly differently to images from the inner versus outer nodes (Wilcoxon signed-rank tests; two-sided; FDR-corrected). Bayes factors supported the null hypotheses (BF 01 ) and are provided in the brackets (paired t-tests; two-sided; Cauchy prior; 0.7071). ( d ) Relational neurons responded more strongly to directly linked stimuli, regardless of whether their preferred stimulus was at an inner or an outer node (see Extended Data Fig. 4 ; Wilcoxon signed-rank tests against PRE; two-sided; FDR-corrected). ( e ) These neurons gradually diminished their selectivity, regardless of whether their preferred stimulus was at an inner or an outer node (see Extended Data Fig. 4 ; Wilcoxon signed-rank tests against PRE; two-sided; FDR-corrected; exact P values for ‘outer’ (from the bottom to the top): P  = 0.004; P  = 0.016; P  = 2.74 × 10 −5 ; P  = 1.73 × 10 −5 ; exact P values for ‘inner’(from the bottom to the top): P  = 0.00014; P  = 6.56 × 10 −5 ; P  = 9.43 × 10 −5 ; P  = 5.65 × 10 −5 ). There were no significant differences between the slopes of diminishing selectivity for relational neurons preferring the inner versus outer nodes (outer versus inner; PRE: P  = 0.566; E1&2: P  = 0.637; E3&4: P  = 0.514; E5&6: P  = 0.831; POST: P  = 0.86; Wilcoxon signed-rank tests, two-sided). ( f ) We replicated our decoding results (see Fig. 3c ) regardless of whether the ‘actual’ image was at an inner or an outer node. Plots show data from all hippocampal-entorhinal neurons ( n  = 546). P values from Kolmogorov-Smirnov tests between CDFs of respective posterior probabilities (PRE versus subsequent study phases; one-sided). Please note that data used for training (PRE) contained the same number of repetitions of each image. Exact P values for ‘outer-actual’ E3&4: P  = 1.56 × 10 −8 ; E5&6: P  = 4.82 × 10 −8 ; POST: P  = 4.82 × 10 −13 . Exact P -value for ‘outer-direct’ POST: P  = 6.08 × 10 −6 . Exact P values for ‘inner-actual’ E5&6: P  = 2.72 × 10 −6 ; POST: P  = 1.58 × 10 −4 . ( g ) We also analysed responses of relational neurons when direct and indirect stimuli were repeated a similar number of times (i.e., direct-inner versus indirect-inner; direct-outer versus indirect-outer; all possible combinations). We found the same pattern of results as before ( P values from Wilcoxon signed-rank tests against PRE; one-sided). The plot shows the mean ‘direct minus indirect’ difference ( ±s.e.m.) in the 0.1 to 1 s time window (peak-normalized and baseline-corrected). For the definition of box plots, please see the legend of Extended Data Fig. 1 .

Extended Data Fig. 9 A simulation study supports that a reconstruction of a successor-like pyramid representation requires distance-dependent scaling of neural responses rather than simple pairwise associations.

Please note the overall similarity between panel e above and the pyramid reconstruction from the neuronal data (Fig. 3f ). For each of the following scenarios (a-f), we generated 6,000 artificial neurons. Each neuron could respond preferentially to any of the six stimuli (random assignment) and this response could vary between 1 to 15 Hz (random assignment). Then, we calculated Euclidean distances between responses to each pair of stimuli across all neurons and used multidimensional scaling to provide the most faithful reconstruction of the respective matrix in 2D (see  Methods ). ( a ) Neurons respond to the preferred stimulus and only one directly linked stimulus on the pyramid; additionally, responses to ‘direct’ are not scaled (i.e., any value between 1 and 15 Hz). ( b ) Neurons respond to the preferred stimulus and only one directly linked stimulus; responses to ‘direct’ are scaled (i.e., 50% of the response to the preferred stimulus). ( c ) Neurons respond to the preferred stimulus and all directly linked stimuli; responses to ‘direct’ are not scaled (see earlier). ( d ) Neurons respond to the preferred stimulus and all directly linked stimuli; responses to ‘direct’ are scaled (i.e., 50% of the response to preferred stimulus). ( e ) Neurons respond to the preferred stimulus and all directly linked stimuli; responses are scaled differently for outer and inner nodes (percentages relative to responses to the preferred stimulus; outer-seeds: direct – 40%; indirect-inner – 20%; indirect-outer – 0%; inner-seeds: direct-inner – 80%; direct-outer – 60%; indirect – 60%). ( f ) Neurons respond to the preferred stimulus and only one stimulus of each type (outer-seeds: direct – 40%; indirect-inner – 20%; indirect-outer – 0%; inner: direct-inner – 80%; direct-outer – 60%; indirect – 60%).

Extended Data Fig. 10 Hippocampal neurons encoded another complex temporal structure with longer paths.

( a ) In a separate study, we tested five additional patients (seven recording sessions; see  Methods ). The procedure was the same as before, but we removed two edges from the pyramid graph during exposure. The resulting ‘diamond’ structure had the ‘shortest path’ distances of length one (D1; corresponding to the ‘direct’ category in the main study) or lengths two and three (D2 and D3, respectively, corresponding to the ‘indirect’ category in the main study). We recorded 221 neurons, of which 55 were located in the hippocampus (we did not have any recording sites in the entorhinal cortex in this additional study). Six hippocampal neurons responded preferentially during PRE to stimuli at the most distant nodes of the diamond (black circles) and were relational neurons according to the criteria described earlier (see  Methods ). ( b ) A relational neuron from the left hippocampus showing progressive tuning to the graph’s distances. Raster plots show individual spikes during each stimulus presentation in E1&2, E3&4, and E5&6. Line plots show the mean number of spikes ± s.e.m. (PRE included for reference). The left panel shows the stimuli and their locations on the graph. ( c ) During late-exposure phases, these hippocampal relational neurons responded more strongly to images located two edges away from their preferred stimulus than to images located three edges away (Wilcoxon signed-rank tests against E1&2 or zero; one-sided). Each circle corresponds to one neuron (the average D2 minus D3 difference in the 0.1 to 1.3 s time window, peak-normalized and baseline-corrected). For the definition of box plots, please see the legend of the Extended Data Fig. 1 . ( d ) We replicated the main population decoding finding that neuronal representations of adjacent nodes progressively overlapped (see Fig. 3c ). That is, the probability of decoding the actual stimulus as ‘actual’ was gradually decreasing, the probability of decoding the direct stimuli as ‘actual’ was increasing, and the probability of decoding indirect stimuli as ‘actual’ did not change significantly. This analysis was conducted on all hippocampal neurons from this additional study ( n  = 55) during all stimuli presentations. Please note that the indirect category combines D2 and D3. P values obtained from Kolmogorov-Smirnov tests comparing CDFs in PRE versus the subsequent study phases (one-sided). Exact P values for ‘actual’ E1&2&3: P  = 1.06 × 10 −7 ; E4&5&6: P  = 1.94 × 10 −7 . ( e ) Next, we analysed CDFs for nodes separated by two versus three links away from the actual stimulus. This analysis was conducted only for trials where the actual stimulus was located at one of the ‘black nodes’ of the diamond (see panel a). We found that the probability of decoding D2 stimuli as ‘actual’ gradually increased during the study, while the probability of decoding D3 stimuli as ‘actual’ gradually decreased. P values obtained from Kolmogorov-Smirnov tests (two-sided). Exact P -value for ‘distance 3’ E4&5&6: P  = 3.69 × 10 −6 .

Supplementary information

Supplementary tables.

This file contains Supplementary Tables 1–6, presenting additional information and results.

Reporting Summary

Rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Tacikowski, P., Kalender, G., Ciliberti, D. et al. Human hippocampal and entorhinal neurons encode the temporal structure of experience. Nature (2024). https://doi.org/10.1038/s41586-024-07973-1

Download citation

Received : 31 May 2023

Accepted : 20 August 2024

Published : 25 September 2024

DOI : https://doi.org/10.1038/s41586-024-07973-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.