Review article
A meta-analytic review of social cognitive function following stroke

https://doi.org/10.1016/j.neubiorev.2019.03.011Get rights and content

Highlights

  • Large deficits in social perception and social behaviour are evident post-stroke.

  • Following stroke, difficulties with theory of mind are of a moderate magnitude.

  • Problems with affective empathy are moderate in magnitude but not significant.

  • Social cognitive deficits generalise across stroke laterality and task features.

  • Social cognitive measures should be routinely included in post-stroke assessment.

Abstract

Although it is now widely recognised that social cognitive difficulties are often evident following stroke, important questions remain about the nature and magnitude of these difficulties, as well as the factors that determine the magnitude of this impairment. A meta-analysis of 58 datasets involving 2567 participants (937 with stroke, 1630 non-clinical controls) was therefore conducted. The results indicated that three of the four core domains of social cognitive function were significantly disrupted in people with stroke. Specifically, while the effect size for affective empathy failed to attain significance (r = −.33), moderate to large deficits were identified for theory of mind (r = −.44), social perception (r = −.55), and social behaviour (r = −.53). These deficits were robust across both left and right lateralized lesions, across social cognitive assessments that differed in their broader cognitive demands, as well as in tasks that varied in their modality of presentation. These data are discussed in the context of broader neuropsychological models of social cognitive function.

Introduction

‘Social cognition’ refers broadly to the cognitive processes used to encode and decode social information, and is critical for people to understand themselves and others, as well as the norms and procedures of the social world (Beer and Ochsner, 2006). Social cognition has been operationalised in a number of different ways. For example, in the schizophrenia literature, social cognition has often been fractionated into four domains: emotion processing, social perception, theory of mind (ToM), and attribution bias (Green et al., 2008, Pinkham et al., 2014). Conversely, in a recent systematic review of meta-analyses examining social cognitive dysfunction across 30 clinical conditions, social cognition was operationalised with just two key domains: ToM and facial emotion recognition (Cotter et al., 2018). However, in the present meta-analysis, social cognition will be defined according to the scheme proposed by Henry et al. (2016) as this scheme was specifically developed with a view to assisting clinical decision making. This scheme identifies four core domains of social cognition: ToM, social perception, affective empathy, and social behaviour.

ToM is defined as the ability to attribute mental states to others, and to recognise that these may differ from one's own (Premack and Woodruff, 1978). A common way to measure this construct is through the use of false-belief tasks which require an individual to inhibit their own perspective and to understand that another person may have different, erroneous beliefs (Baron-Cohen et al., 1985). Social perception refers to the ability to interpret cues to emotional states (Phillips et al., 2002, Mitchell and Phillips, 2015), where these cues may or may not be inherently emotional in and of themselves. Social perception has most commonly been assessed using the Ekman Faces stimuli (Ekman and Friesen, 1976) which are photographs depicting the six basic emotions (disgust, anger, fear, surprise, sadness and happiness) as well as neutral faces. Affective empathy refers to a person's emotional response to the perceived situation of another (Davis, 1994, Shamay-Tsoory et al., 2004). This capacity is commonly assessed using self- or informant-report questionnaires such as the Empathy Quotient (Baron-Cohen and Wheelwright, 2004) or the Empathic Concern subscale of the Interpersonal Reactivity Index (Davis, 1983). Because definitions of cognitive empathy are quite similar to those of ToM (Shamay-Tsoory et al., 2004), focusing on the affective component of empathy alone allows for clearer differentiation between these domains at the behavioural level. Finally, social behaviour refers to how a person conducts themselves during interactions with other people. Social behavioural abnormalities may manifest in many ways such as making inappropriate comments or advances (Grace et al., 1999), and are commonly assessed using informant-rated measures such as the Frontal Systems Behaviour Scale which examines behavioural disturbances such as disinhibition (Grace and Malloy, 2001). Although social behavioural abnormalities can arise because of disturbances in other social cognitive domains (such as poor ToM or problems with social perception, see e.g., Halberstadt et al., 2011, Ruffman et al., 2016), they can also arise as a consequence of broader neurocognitive impairment, such as reduced executive control (Beer et al., 2006, Henry et al., 2009). As a consequence, it is possible to exhibit social behavioural abnormalities while other aspects of social cognitive function are intact.

Empirical evidence consistently shows that a breakdown in any of these four core social cognitive domains can lead to broader problems in a variety of social function domains (Blonder et al., 2012, Cooper et al., 2014, Couture et al., 2006, Fett et al., 2011, Kalin et al., 2015). Moreover, in addition to being linked to poor social outcomes, social cognitive deficits have been consistently identified as important predictors of broader prognostic outcomes, including mental health and quality of life (Cooper et al., 2014, Maat et al., 2012, Milders et al., 2003, Phillips et al., 2011). Indeed, the critical role of social cognitive function in clinical practice has now been formally recognised, with social cognition included for the first time as one of the six core domains of cognitive function in the most recent version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013). The significance of social cognitive function has also been further highlighted in a recent systematic review of meta-analyses across 30 different clinical conditions. This review concluded that social cognitive impairments are a core feature of many clinical groups, and are equivalent in magnitude to deficits seen across more ‘traditional’ neurocognitive domains such as memory and executive function (Cotter et al., 2018, p. 94).

Stroke is one of the many clinical groups that has been found to be at significantly increased risk of social cognitive impairment (Happé et al., 1999, Yeh and Tsai, 2014, Yuvaraj et al., 2013). A stroke occurs when blood supply to the brain is interrupted. This happens when an artery carrying blood has been blocked (ischaemic stroke) or has ruptured (haemorrhagic stroke). The interruption in blood supply deprives brain cells of the oxygen and nutrients normally carried in the blood, consequently leading to cell death and brain damage (Markus, 2008). Although a stroke can occur at any age, it is most likely to occur in late adulthood. Indeed, age is one of the biggest risk factors for stroke with people aged 75–84 years at 25 times the risk of stroke than those aged 45–54 years (Markus, 2008). As a consequence, with an ageing population worldwide, the number of stroke cases are predicted to substantially increase in coming years (Di Carlo, 2009, United Nations Population Division, 2007). It is therefore of unprecedented importance to gain a more thorough and nuanced understanding of the consequences of stroke so that more effective treatment can occur, potentially improving patients’ quality of life and thus alleviating some of the global disease burden (Feigin et al., 2014). Specifically, identifying the nature and magnitude of any social cognitive impairment associated with stroke not only has important theoretical implications, but from a practical perspective has the potential to inform and direct clinical practice by allowing health practitioners to better anticipate the type(s) of social cognitive problems likely to arise for individual stroke patients.

It is therefore surprising that despite more than 30 years of research focused on social cognitive function following stroke, to date there has been no meta-analytic investigation of this literature. This means that important questions remain about the typical magnitude of stroke-related social cognitive impairment, as well as how robust any impairment is across different types of stroke, different types of social cognitive domain, and across social cognitive tasks that vary in methodological characteristics. The present study therefore used meta-analytic methodology to quantify the magnitude of stroke-related deficits separately for each of the four core social cognitive domains, and where sufficient data was available, across different types of stroke, and task characteristics within each social cognitive domain.

It was anticipated that stroke would be associated with impairment across all four social cognitive domains relative to healthy controls. This was based on the fact that, to date, most empirical studies have identified significant social cognitive impairment following stroke (e.g. Abbott et al., 2014a, Blonder et al., 2012, Braun et al., 2005, Cooper et al., 2014, Hamilton et al., 2017, Happé et al., 1999, Yeh and Tsai, 2014), with relatively few identifying null or opposite effects (e.g. Etcoff et al., 2000, Martin and McDonald, 2006). This finding that social cognitive impairment is consistently identified following stroke makes sense in light of the neural substrates known to subserve social cognitive operations. Specifically, social cognition imposes demands on a large number of different brain areas and their connectivity, including (but not limited to) the fusiform gyrus, superior temporal sulcus, insula, precuneus, temporo-parietal junction, amygdala, orbitofrontal cortex, cingulate, somatosensory cortices and medial prefrontal cortex (Adolphs, 2001, Kennedy and Adolphs, 2012, Molenberghs et al., 2016, Roberts et al., 2018, Schurz et al., 2014, Yang et al., 2015). Even relatively mild dysfunction in one of these structures, diffuse dysfunction that is not neuroanatomically specific, or white-matter damage, has the potential to disrupt social cognitive function. Deficits of social cognition should therefore be understood as a disruption of the interactions within and between large-scale social cognition networks (for a review, see Henry et al., 2016). Thus, although there is considerable clinical heterogeneity in the types of brain damage seen following stroke, social cognitive deficits should be a common presenting symptom.

However, while it was anticipated that significant social cognitive impairment would be identified using meta-analytic methods, less clear was the magnitude of any impairment, as well as whether these deficits varied systematically across social cognitive domains. This is because only a handful of studies in the stroke literature have examined multiple social cognitive domains (e.g. Honan et al., 2016, Leigh et al., 2013, Parola et al., 2016, Sedeño et al., 2016, Yeh and Tsai, 2014) and these few studies have not presented results in a way so that they can be compared directly with each other.

With respect to different types of stroke, while many clinical characteristics are potentially important, in the present meta-analysis it was only possible to investigate the potential influence of stroke laterality, as only for this stroke characteristic were there sufficient studies that reported data separately for these stroke sub-types. While a number of studies have reported no difference in level of social cognitive impairment as a function of stroke laterality (Abbott et al., 2014a, Braun et al., 2005, Cheung et al., 2006, Cooper et al., 2014, Pell, 2006, Pell and Baum, 1997a, Twist et al., 1991), others have identified greater impairment for strokes lateralised to the right hemisphere, relative to those lateralised to the left hemisphere (Blonder et al., 1991, Borod et al., 1986, Hamilton et al., 2017, Kucharska-Pietura et al., 2003, Ross and Monnot, 2008, Siegal et al., 1996, Surian and Siegal, 2001, Paradee et al., 2008, Walker et al., 2002). In addition, data from non-clinical social neuroscience studies suggest that social cognitive functions may be slightly lateralised to the right hemisphere. For example, Saxe and Wexler (2005) found that the right temporo-parietal junction may be particularly important for inferring another's beliefs and Adolphs et al. (2000) found that damage to the right somatosensory cortices, right anterior supramarginal gyrus, and right insula was associated with impaired facial emotion recognition. In a similar vein, Ross and Monnot (2008) found that the right posterior–superior temporal region was critical for recognition of emotional prosody in particular. It was therefore anticipated that while stroke patients in general would exhibit social cognitive deficits, these difficulties would be more pronounced for those who had suffered right-hemisphere damage as opposed to left-hemisphere damage.

Finally, in the broader social cognitive literature, important distinctions have been made between tasks that vary in their broader cognitive demands (specifically, first- versus second-order ToM, and emotion labelling versus discrimination), as well as in tasks that vary in their modality of presentation (facial, prosodic, multimodal). The final aim of the present meta-analysis was therefore to investigate whether stroke-related social cognitive impairment varied systematically as a function of specific task features. In particular, it was predicted that stroke would be associated with greater impairment on social cognitive assessments that impose greater demands on secondary cognitive resources that are known often to be disrupted following stroke, such as working memory. By this view, any stroke-related impairment should be greater for more complex second-order relative to first-order ToM tasks (see Champagne-Lavau and Joanette, 2009), and for emotion labelling relative to emotion discrimination (Phillips et al., 2010). Less clear was whether stroke-related impairment would differ as a function of modality of presentation (i.e., whether a task involved facial, prosodic or multimodal cues). However, one possibility is that stroke-related deficits may be greater for prosodic cues as it has previously been argued that recognising emotion from vocal cues is generally more difficult than doing so using facial cues (Adolphs, 2002). Indeed, it has been suggested that this may be due to a temporal dimension as emotion in faces can be detected from a single snapshot whereas emotional prosody requires a dynamic stimulus, thus one must integrate information over time (Schirmer and Adolphs, 2017). The present meta-analysis will allow a preliminary test of this possibility by systematically comparing studies that vary in their modality of presentation.

In sum, this meta-analysis will provide important clarification about the nature and magnitude of social cognitive impairment following stroke, as well as whether any impairment varies as a function of social cognitive domain (ToM, social perception, affective empathy and social behaviour), stroke characteristics (right versus left lateralised lesions), as well as across tasks that differ in their broader cognitive demands (first- versus second-order ToM, as well as emotion labelling versus discrimination), and their modality of presentation (facial, prosodic, multimodal). The results therefore will have theoretical implications for understanding the social cognitive sequalae associated with stroke, but also important practical implications. This is because providing clinicians with evidence-based data on the prevalence, nature and magnitude of social cognitive dysfunction directly speaks to the issue of whether social cognitive assessments should be routinely included in the initial clinical workup of stroke referrals, and stroke rehabilitation more broadly.

Section snippets

Literature search and study selection

PRISMA guidelines were adhered to when conducting this meta-analysis (Moher et al., 2009). A systematic review of the literature was conducted, with identification of studies eligible for inclusion achieved by searching the PubMed (United States National Library of Medicine), Scopus (Elsevier), Web of Science (Thomson Reuters), and Psyc INFO (American Psychological Association) databases. Backward citation searches were also undertaken. The search was completed in June 2018. Abstracts were

Included articles

As illustrated in Fig. 2, the initial literature search yielded a total of 27,948 articles. Removing duplicates reduced this to 20,342 articles. The titles and abstracts were screened, and 217 articles were retained and subjected to a full-text review. Ultimately, 58 studies published between 1985 and 2018 met inclusion criteria. This included 24 studies examining ToM, 35 examining social perception, four examining affective empathy, and two examining social behaviour. Five studies contributed

The magnitude and specificity of social cognitive deficits following stroke

By providing the first meta-analytic integration of the literature focused on social cognitive outcomes following stroke, the current study provides clear evidence that social cognitive impairment is a common neurocognitive disturbance following stroke, and that the magnitude of this impairment is typically moderate to large in magnitude. These findings therefore align with the broader literature that has argued social cognitive impairment is a prominent feature of many clinical populations (

Conclusions

It is now widely accepted that social cognitive impairment is a common feature of neurological illness, and a key predictor of broader prognostic outcomes, including social function, mental health, and quality of life in many clinical groups (for a review, see Henry et al., 2016). It is therefore important not only theoretically, but from a practical perspective, to better understand the nature and magnitude of any stroke related impairment in this neurocognitive domain. The present

Conflicts of interest

None declared.

Acknowledgements

This research was supported by the Australian Research Council Discovery Project Grant (DP170101239). P.M. was supported by a Heart Foundation Future Leader Fellowship (1000458). J.D.H. was supported by an Australian Research Council Future Fellowship (FT170100096).

References (160)

  • J.C. Borod et al.

    Parameters of emotional processing in neuropsychiatric disorders: conceptual issues and a battery of tests

    J. Commun. Disord.

    (1990)
  • M. Braun et al.

    Emotion recognition in stroke patients with left and right hemispheric lesion: results with a new instrument – the FEEL Test

    Brain Cognit.

    (2005)
  • A.E. Cancelliere et al.

    Lesion localization in acquired deficits of emotional expression and comprehension

    Brain Cognit.

    (1990)
  • A. Caracuel et al.

    Frontal behavioral and emotional symptoms in Spanish individuals with acquired brain injury and substance use disorders

    Arch. Clin. Neuropsychol.

    (2008)
  • M. Champagne-Lavau et al.

    Pragmatics, theory of mind and executive functions after a right-hemisphere lesion: different patterns of deficits

    J. Neurolinguist.

    (2009)
  • S. Charbonneau et al.

    Perception and production of facial prosodic emotions by chronic CVA patients

    Neuropsychologia

    (2003)
  • C.C.Y. Cheung et al.

    The differential effects of thalamus and basal ganglia on facial emotion recognition

    Brain Cognit.

    (2006)
  • J. Cotter et al.

    Social cognitive dysfunction as a clinical marker: a systematic review of meta-analyses across 30 clinical conditions

    Neurosci. Biobehav. Rev.

    (2018)
  • P.J. Easterbrook et al.

    Publication bias in clinical research

    Lancet

    (1991)
  • Y. Fan et al.

    Is there a core neural network in empathy? An fMRI based quantitative meta-analysis

    Neurosci. Biobehav. Rev.

    (2011)
  • V.L. Feigin et al.

    Global and regional burden of stroke during 1990–2010: findings from the Global Burden of Disease Study 2010

    Lancet

    (2014)
  • A.K.J. Fett et al.

    The relationship between neurocognition and social cognition with functional outcomes in schizophrenia: a meta-analysis

    Neurosci. Biobehav. Rev.

    (2011)
  • F. Happé et al.

    Acquired “theory of mind” impairments following stroke

    Cognition

    (1999)
  • J.D. Henry et al.

    Recognition of disgust is selectively preserved in Alzheimer's disease

    Neuropsychologia

    (2008)
  • W. Hirst et al.

    Constraints on the processing of indirect speech acts: evidence from aphasiology

    Brain Lang.

    (1984)
  • M. Kalin et al.

    Social cognition, social competence, negative symptoms and social outcomes: inter-relationships in people with schizophrenia

    J. Psychiatr. Res.

    (2015)
  • P. Kanske et al.

    Dissecting the social brain: introducing the EmpaToM to reveal distinct neural networks and brain-behavior relations for empathy and theory of mind

    NeuroImage

    (2015)
  • J.A. Kaplan et al.

    The effects of right hemisphere damage on the pragmatic interpretation of conversational remarks

    Brain Lang.

    (1990)
  • D.P. Kennedy et al.

    The social brain in psychiatric and neurological disorders

    Trends Cognit. Sci.

    (2012)
  • K. Kucharska-Pietura et al.

    Perception of emotions from faces and voices following unilateral brain damage

    Neuropsychologia

    (2003)
  • A. Liberati et al.

    The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration

    J. Clin. Epidemiol.

    (2009)
  • A. Maat et al.

    Social cognition and quality of life in schizophrenia

    Schizophr. Res.

    (2012)
  • H. Markus

    Stroke: causes and clinical features

    Medicine

    (2008)
  • J.F. Martín-Rodríguez et al.

    Theory of mind deficits in patients with acquired brain injury: a quantitative review

    Neuropsychologia

    (2010)
  • R.L.C. Mitchell et al.

    The overlapping relationship between emotion perception and theory of mind

    Neuropsychologia

    (2015)
  • P. Molenberghs et al.

    Understanding the minds of others: a neuroimaging meta-analysis

    Neurosci. Biobehav. Rev.

    (2016)
  • L. Monetta et al.

    Irony comprehension and theory of mind deficits in patients with Parkinson's disease

    Cortex

    (2009)
  • F. Muller et al.

    Exploring theory of mind after severe traumatic brain injury

    Cortex

    (2010)
  • J.D. Abbott et al.

    The influence of left and right hemisphere brain damage on configural and featural processing of affective faces

    Laterality

    (2014)
  • M. Adamaszek et al.

    Impairment of emotional facial expression and prosody discrimination due to ischemic cerebellar lesions

    Cerebellum

    (2014)
  • R. Adolphs et al.

    A role for somatosensory cortices in the visual recognition of emotion as revealed by three-dimensional lesion mapping

    J. Neurosci.

    (2000)
  • American Psychiatric Association

    Diagnostic and Statistical Manual of Mental Disorders: DSM-5

    (2013)
  • P.E. Bailey et al.

    Electromyographic evidence for age-related differences in the mimicry of anger

    Psychol. Aging

    (2009)
  • N. Balaban et al.

    The effect of theory of mind impairment on language: referring after right-hemisphere damage

    Aphasiology

    (2016)
  • N. Balaban et al.

    Theory of mind impairment after right-hemisphere damage

    Aphasiology

    (2016)
  • S. Baron-Cohen et al.

    The empathy quotient: an investigation of adults with Asperger syndrome or high functioning autism, and normal sex differences

    J. Autism Dev. Disord.

    (2004)
  • J.S. Beer et al.

    Orbitofrontal cortex and social behavior: integrating self-monitoring and emotion-cognition interactions

    J. Cognit. Neurosci.

    (2006)
  • S. Besharati et al.

    Mentalizing the body: spatial and social cognition in anosognosia for hemiplegia

    Brain

    (2016)
  • A. Biervoye et al.

    Assessing the integrity of the cognitive processes involved in belief reasoning by means of two nonverbal tasks: rationale, normative data collection and illustration with brain-damaged patients

    PLoS One

    (2018)
  • L.X. Blonder et al.

    The role of the right-hemisphere in emotional communication

    Brain

    (1991)
  • Cited by (0)

    View full text