Deficits in Social Cognition: An Unveiled Signature of Multiple Sclerosis

Abstract

Background and Objectives: Multiple sclerosis (MS) is a chronic progressive inflammatory disease of the central nervous system, representing the primary cause of non-traumatic disability in young adults. Cognitive dysfunction can affect patients at any time during the disease process and might alter the six core functional domains. Social cognition is a multi-component construct that includes the theory of mind, empathy and social perception of emotions from facial, bodily and vocal cues. Deficits in this cognitive faculty might have a drastic impact on interpersonal relationships and quality of life (QoL). Although exhaustive data exist for non-social cognitive functions in MS, only a little attention has been paid for social cognition. The objectives of the present work are to reappraise the definition and anatomy of social cognition and evaluate the integrity of this domain across MS studies. We will put special emphasis on neuropsychological and neuroimaging studies concerning social cognitive performance in MS. Methods:Studies were selected in conformity with PRISMA guidelines. We looked for computerized databases (PubMed, Medline, and Scopus) that index peer-reviewed journals to identify published reports in English and French languages that mention social cognition and multiple sclerosis, regardless of publication year. We combined keywords as follows: (facial emotion or facial expression or emotional facial expressions or theory of mind or social cognition or empathy or affective prosody) AND multiple sclerosis AND (MRI or functional MRI or positron emission tomography or functional imaging or structural imaging). We also scanned references from articles aiming to get additional relevant studies. Results: In total, 26 studies matched the abovementioned criteria (26 neuropsychological studies including five neuroimaging studies). Available data support the presence of social cognitive deficits even at early stages of MS. The increase in disease burden along with the “multiple disconnection syndrome” resulting from gray and white matters pathology might exceed the “threshold for cerebral tolerance” and can manifest as deficits in social cognition. Admitting the impact of the latter on patients’ social functioning, a thorough screening for such deficits is crucial to improving patients’ QoL. (JINS, 2017, 23, 266–286)

INTRODUCTION

Multiple sclerosis (MS) is a chronic progressive disease of the central nervous system (CNS) representing the primary cause of non-traumatic disability in young adults (Compston & Coles, 2008). Its precise etiology remains unclear and includes a constellation of mechanisms. Its most common type is the relapsing-remitting (RR) which usually shifts to a secondary progressive (SP) fate (Compston & Coles, 2008). Primary progressive (PP) MS is a third form which still does not have approved disease-modifying therapies and is considered to have a poor prognosis (Gajofatto & Benedetti, 2015; Segal & Stüve, 2016). The disease course can be very heterogeneous, through which patients may develop sensorimotor, cerebellar, emotional, and cognitive symptoms (Compston & Coles, 2008).

Cognitive decline occurs in approximately 40–65% of MS patients at some point during their life (Benedict et al., 2006; Rao, Leo, Bernardin, & Unverzagt, 1991; Sanfilipo, Benedict, Weinstock-Guttman, & Bakshi, 2006) and may involve any of the six core functional domains: are perceptual-motor functions, language, learning and memory, executive functions, complex attention, and social cognition (5th ed.; DSM–5; American Psychiatric Association, 2013). Working memory and information processing speed (IPS) are the most frequently impaired areas in MS, followed by learning, memory, and executive functions (Benedict et al., 2006; Rao, Leo, Bernadin, et al., 1991; Sanfilipo et al., 2006).

Although these domains have been well studied in MS (Mohr & Cox, 2001), little attention has been paid for social cognition, which defines the individual’s ability to understand others’ mind and feelings (Sebastian et al., 2012; Uekermann, Channon, Winkel, Schlebusch, & Daum, 2007; Uekermann & Daum, 2008; Uekermann et al., 2010; Vistoli, Brunet-Gouet, Baup-Bobin, Hardy-Bayle, & Passerieux, 2011; Wolkenstein, Schonenberg, Schirm, & Hautzinger, 2011). It is a multi-component construct that includes theory of mind (ToM) (Abdel-Hamid et al., 2009; Koelkebeck, Abdel-Hamid, Ohrmann, & Brune, 2008), empathy (Carr, Iacoboni, Dubeau, Mazziotta, & Lenzi, 2003; Decety & Jackson, 2004; Leslie, Johnson-Frey, & Grafton 2004; Seitz, Nickel, & Azari, 2006; Vollm et al., 2006), and social perception of emotions from prosody, facial expressions, and bodily gestures (Calder & Young, 2005; Ethofer et al., 2006; Heikkinen et al., 2010; Ross, Thompson, & Yenkosky, 1997; Wheaton, Thompson, Syngeniotis, Abbott, & Puce, 2004; Uekermann & Daum, 2008; Uekermann, Abdel-Hamid, Lehmkamper, Vollmoeller, & Daum, 2008).

The integrity of social cognitive functions is crucial for proper retrieval of information from social stimuli, to establish an appropriate social interaction and cope with chronic diseases such as MS (Montel & Bungener, 2007). In this perspective, deficits in social cognition might have a drastic impact on quality of life (QoL) and interpersonal communication. Interestingly, altered social interactions have been frequently reported in MS patients (Buhse, 2008; Kesselring & Klement, 2001; Rao, Leo, Elllington, et al., 1991) and could be reflected by high rates of divorce and unemployment (Julian, Vella, Vollmer, Hadjimichael, & Mohr, 2008; Langdon, 2011; Pfleger, Flachs, & Koch-Henriksen, 2010; Rao, Leo, Elllington, et al., 1991) and increased level of social anxiety (Poder et al., 2009).

The main objective of the present work is to review the available data concerning social cognition in MS. First, we will reappraise terms defining social cognition, particularly social perception of emotions, theory of mind and empathy. This section will also include the neuroanatomy of social cognition in healthy brain. The second section will examine neuropsychological studies regarding social cognition in MS. This will be followed by a third section that puts emphasis on neuroimaging studies of social cognition in this population. Finally, findings will be discussed in the light of the “cognitive reserve” hypothesis. The clinical assessment of social cognition is developed elsewhere (for reviews, see Henry, von Hippel, Molenberghs, Lee, & Sachdev, 2016) and is beyond the scope of this review.

NEUROANATOMICAL CORRELATES OF SOCIAL COGNITION

In the past few years, tremendous advances in neuroimaging have unveiled many cerebral hubs that take part in brain networks dedicated to social cognition. Although social cognitive domains might recruit different cerebral areas, an overlapping seems to occur among their networks.

Social Perception

Social perception of emotions from facial expressions

A chief element in social interaction is the ability to recognize facial expressions and their emotional significance (Brothers, 1990; Van Kleef, 2009). A large-scale network and a complex processing have been suggested by this skill. The first step consists of early visual processing of faces, which entails a relatively shared neural pathway for facial identity discrimination and facial emotion recognition (Calder & Young, 2005; LaBar, Crupain, Voyvodic, & McCarthy, 2003; Palermo & Rhodes, 2007; Vuilleumier & Pourtois, 2007). In the following steps, distinct cortical regions would intervene. For instance, the fusiform face area (FFA) plays a key role in recognizing invariant or neutral facial aspect that defines identity (Kanwisher & Yovel, 2006). Other areas, such as the superior temporal sulcus (STS), are more specialized in changeable facial features (i.e., perception of eyes and mouth movements; Allison, Puce, & McCarthy, 2000). The amygdala is an essential element in automatic attentional capture by emotionally relevant facial expressions (Vuilleumier & Pourtois, 2007). The orbitofrontal cortex (OFC) is crucially involved in processing non-conscious aspects of facial expressions (Adolphs, 2006; Krause et al., 2009). To note, non-conscious perception of emotional stimuli is an intrinsic property of the healthy brain. Through this process, emotionally relevant visual stimuli that are not perceived consciously can induce behavioral responses manifesting as changes in emotional states (Tamietto & de Gelder, 2010).

Of interest, observing facial emotions is known to trigger an affective reaction which subsequently leads to adaptive changes in the observer’s behavior (Van Kleef, 2009). Such a reaction depends on the generation of an “emotional state” and a “motivation state.” The former is mainly created by the anterior insula that integrates environmental cues with viscero-reception of internal body state (Adolphs, 2002). The latter is completed via the action of the anterior cingulate cortex (ACC; Critchley, 2005).

Social perception of vocal cues: Affective prosody

Prosody is an aspect of language represented by acoustic characters such as the pattern of intonation (i.e., timing, pitch, rhythm, stress, and pausing; Heikkinen et al., 2010; Uekermann & Daum, 2008; Uekermann, Abdel-Hamid et al., 2008). Among the subdivisions of prosody, the most relevant here are the linguistic and affective components (Ross et al., 1997; Uekermann & Daum, 2008). While processing linguistic prosody seems to involve left-sided brain regions, perception of affective prosody is a dominant function of the right hemisphere and encompasses many steps (Ethofer et al., 2006; Uekermann & Daum, 2008; Uekermann, Abdel-Hamid, et al., 2008; Wildgruber, Ackermann, Kreifelts, & Ethofer, 2006).

For example, primary and higher order right-hemispheric acoustic areas deal with extracting suprasegmental acoustic information whose meaningful representation mainly involves the right STS. Bilateral inferior frontal regions maintain an explicit assessment of affective prosody. Lastly, the corpus callosum (CC) participates by ensuring the inter-hemispheric integration of language functions (Ross et al., 1997); this seems crucial to understanding emotional prosody, especially when the latter is not in agreement with the linguistic component. In this situation, a proper understanding requires a successful prioritization of the affective aspect over the linguistic one (Uekermann et al., 2010).

Theory of Mind

ToM, also known as “mentalizing,” suggests understanding and predicting mental states of others, based on (i) their emotions and feelings (affective ToM) and/or (ii) their intentions, thoughts, and beliefs (cognitive ToM; Stone, Baron-Cohen, & Knight, 1998; Uekermann et al., 2007; Uekermann, Channon, et al., 2008). ToM is a key aspect of social cognition and constitutes an important prerequisite for adequate social interactions. The two extremes of ToM abnormalities are known as “undermentalizing” (insufficient ToM) and “overmentalizing” (excessive ToM), which, respectively, refer to deficits commonly encountered in patients with autism (Baron-Cohen, 2000) and schizophrenia (Frith, 2004).

ToM recruits a complex neural network which includes the ACC, OFC, amygdala and many areas of the temporal lobe (i.e., posterior STS, temporal pole, and temporoparietal junction [TPJ]; Adolphs et al., 2002; Frith & Frith, 2006; Herold et al., 2009; Kuperberg et al., 2003; Schulte-Rüther et al., 2011; Stone et al., 1998; Uekermann et al., 2007, 2010). Remarkably, available data suggest that ToM subcomponents be modulated by distinct frontal circuits. Saying differently, while the ventromedial prefrontal cortex (VMPFC) appears to be particularly involved in processing affective ToM (Shamay-Tsoory & Aharon-Peretz, 2007); the ventrolateral prefrontal (VLPFC) and dorsolateral prefrontal cortices (DLPFC) seem to be chiefly implicated in mediating cognitive ToM (Shamay-Tsoory & Aharon-Peretz, 2007).

Empathy

Empathy lies in the individual’s ability to reason, predict the consequences of emotions, and have a compassionate response accordingly (Decety & Jackson, 2004; Ruby & Decety, 2004; Uekermann & Daum, 2008; Uekermann, Channon, et al., 2008; Uekermann et al., 2010). Such a skill consists of taking another person’s perspective (other-oriented emotions), which often leads to altruistic helping behavior. In contrast, self-oriented emotions, such as personal distress, primarily focus on the empathizer’s feelings in a way that it might interfere with prosocial behavior and, therefore, are not considered empathy (Davis, 1983; Tangney, Stuewig, and Mashek, 2007). The empathy network includes anterior insula and regions of the prefrontal and frontal cortices (i.e., dorsal and middle parts of the ACC, supplementary motor areas; Decety & Jackson, 2004; Fan, Duncan, de Greck, & Northoff, 2011; Gallese, Keysers, & Rizzolatti, 2004; Seitz et al., 2006; Vollm et al., 2006).

STUDY SELECTION

For the aims of this review, studies were selected in conformity with PRISMA guidelines (Moher et al., 2009). First, we searched for computerized databases that index peer-reviewed journals (PubMed, Medline, and Scopus) to identify published reports, in English and French languages, mentioning social cognition and multiple sclerosis, regardless of publication year. For the section dealing with neuropsychological studies, we combined keywords as follows: (facial emotion or facial expression or emotional facial expressions or theory of mind or social cognition or empathy or affective prosody) AND multiple sclerosis. Second, for the section dedicated to neuroimaging underpinnings of social cognitive performance in MS, our combination consisted of the previous keywords AND [MRI/functional MRI (fMRI)/positron emission tomography (PET)/functional imaging/structural imaging]. In both researches, we scanned references from articles aiming to get additional relevant studies. Twenty-six neuropsychological studies matched these criteria (25 in English, 1 in French), of which five also contained neuroimaging data.

SOCIAL COGNITION ACROSS MULTIPLE SCLEROSIS STUDIES

After defining social cognition and its neuroanatomical substrates in healthy humans, we will continue by reviewing the neuropsychological studies assessing social cognition in MS patients.

Social Perception

Social perception of facial emotions in multiple sclerosis

In the past two decades, there was a growing interest in understanding the abilities of MS patients to recognize emotional facial expressions (EFE). The earliest insight into this topic came from a pioneering study by Beatty and colleagues (1989). Patients with chronic progressive MS and age and education matched healthy controls (HCs) performed the Benton Facial Recognition Test (BFRT) for facial identity discrimination (Benton, Sivan, Hamsher, Varney, & Spreen, 1994) and an affective judgment task that evaluates the ability to recognize the six basic facial emotions (i.e., happiness, sadness, anger, fear, disgust and surprise; Ekman and Friesen, 1976).

Compared to HCs, patients had worse cognitive performance and lower accuracy in both facial identity discrimination and facial emotion recognition. The deficits in emotion recognition were not restricted to a particular emotional state. Furthermore, correlation analysis revealed a positive correlation between scores on BFRT and those on affective judgment task. This made the authors consider the observed deficits in EFE recognition as secondary to those in facial identity discrimination which can somewhat reflect visuoperceptual deficits. Concurrently, the authors included a group of RR MS patients who, unlike their progressive counterparts, had preserved abilities to recognize EFE but were “slightly” impaired on BFRT test. Based on these findings, one would assume that clinical and demographic differences between both patient groups accounted for the observed differences in recognizing EFE. Unfortunately, the RR MS group was not included in the remaining statistical analyses.

Consistent with the first report, Parada-Fernández et al. considered a mixed cohort of RR and progressive MS patients and healthy subjects (2015). The authors used BFRT and Facially Expressed Emotion Labeling task (Kessler, Bayerl, Deighton, & Traue, 2002) which, respectively, evaluate facial identity discrimination and facial emotion recognition. To further eliminate any bias that might result from visual impairment, the authors excluded patients who had visual difficulties which disable them from reading and/or writing. This study showed that patients had difficulties in facial emotion recognition and identity discrimination. Moreover, a stepwise multiple regression analysis revealed that disease type and non-social cognitive abilities were the main contributors to the observed deficits in recognizing EFE. Facial identity discrimination did not seem to contribute to social cognitive deficits in this study.

Similarly, Berneiser et al. applied the facial affect task of Florida Affect Battery (Bowers, Blonder, & Heilman, 1991, 2001) to evaluate EFE recognition abilities in patients with different MS subtypes and HCs (2014). Compared to HCs, patients had worse performance in all subsets of the facial affect task, even after exclusively considering those with intact abilities to discriminate facial identity. This stands with what Parada-Fernández et al. stated (2015) and is against the earlier suggestion by Beatty et al. (1989). Again, Berneiser et al. found more pronounced deficits among SP MS patients compared to those suffering from RR MS. In addition, emotion recognition scores were directly correlated with cognitive performance and indirectly correlated with each of depression and fatigue scores, disease duration, and level of physical disability based on the Expanded Disability Status Scale score (EDSS).

Analogously, in the study by Cecchetto et al., patients had poorer performance than HCs on tasks assessing the recognition of all of the six basic facial emotions but had intact abilities to discriminate facial identity (2014). When patients were subdivided based on physical disability (EDSS scores), only highly disabled ones were impaired in labeling EFE. The latter was further correlated with disease characteristics (i.e., disease duration and EDSS scores) and non-social cognitive performance.

In the same perspective, Phillips et al. have assessed emotions’ recognition skills using static (Ekman & Friesen, 1976) and dynamic measures (videos featuring frustration, excitement, annoyance, and boredom by Sullivan & Ruffman, 2004) (2011). Compared to HCs, patients had worse mood and cognitive scores and showed deficits in recognizing facial emotions without differences in facial identity discrimination. The deficits remained significant even after accounting for depression and cognitive decline. In addition, EFE recognition was associated with social and psychological aspects of QoL (Phillips et al., 2011).

Unlike the above-mentioned studies that brought out social cognitive deficits in all of the six basic facial emotions, others rather found an isolated pattern of impairment in recognizing EFE. For instance, two MS trials documented exclusive deficits in identifying the emotions “fear” and “anger” (Henry et al., 2009, 2011). These results are in line with those of a third study comparing HCs and two groups of MS patients with or without altered abilities to recognize EFE (Krause et al., 2009). Here, compared to HCs and the preserved MS group, affected patients had deficits in recognizing “sadness,” “fear,” and “anger” but were able to discern positive emotions. Moreover, in a fourth study, patients with intact abilities to discriminate facial identity had significant impairment in identifying “fear,” “sadness,” “anger,” and “surprise” (Prochnow et al., 2011). More interestingly, when considering physical disability as a variable, severely disabled patients had worse cognitive performance and displayed an additional deficit in the emotion “disgust.” Thus, higher disability levels seem to contribute to the emergence of other deficits.

The isolated involvement of negative emotions in the latter studies (Henry et al., 2009, 2011; Krause et al., 2009; Prochnow et al., 2011) might be explained as follows: One of the possibilities is that positive emotions might be relatively easier to process than negative ones and could hence be more compensated (Skowronski & Carlston, 1989). This idea is supported by one MS study in which “happiness,” for example, was better recognized than “fear” or “sadness” (Cecchetto et al., 2014). Another reason is that MS patients might express low sensitivity toward aversive stimuli (Di Bitonto et al., 2011). Indeed, these patients were found to have reduced emotional reactivity to negative stimuli (i.e., sounds and pictures) compared to HCs but had normal reactivity to positive ones (Di Bitonto et al., 2011).

Functional neuroimaging data can provide a third explanation. In fact, the normal processing of each emotion seems to induce a selective pattern of brain activation (Jehna, Neuper, et al., 2011). For example, some cerebral areas (i.e., VLPFC, ACC, and superior temporal gyri) are more activated during processing of “sadness,” while others (i.e., DLPFC, cingulate gyrus, inferior temporal gyrus, and cerebellum) appear to be more specific for “happiness” (Habel, Klein, Kellermann, Shah, & Schneider, 2005). This idea can be exemplified by one fMRI study in MS where the selective deficit in recognizing negative emotions was associated with hypoactivation of cortical areas devoted to processing negative emotions (i.e., ACC, fSTS, and VLPFC) (Krause et al., 2009).

Finally, five studies found intact EFE recognition abilities in MS patients (Di Bitonto et al., 2011; Jehna et al., 2010; Jehna, Langkammer, et al., 2011; Passamonti et al., 2009; Pinto et al., 2012). This is not surprising given that four of them recruited exclusively (Passamonti et al., 2009; Jehna, Langkammer, et al., 2011; Di Bitonto et al., 2011) or predominantly (Pinto et al., 2012) RR MS patients. Once more, the cohort of the fifth study consisted mostly of preserved patients with clinically isolated syndrome and RR MS (Jehna et al., 2010). Here, the authors assessed the accuracy and reaction time during EFE recognition task. Although accuracy did not differ between both groups, patients were slower than HCs. The observed slowing might not reflect deficits in emotion recognition, but could rather hint to a general delay in IPS which is frequent in MS (Vázquez-Marrufo et al., 2014) or an age-related slowing (Knight & Mather, 2013) since HCs were significantly younger than patients.

The above-mentioned studies are summarized in Table 1. The differences in their outcomes might be explained by the disparity in clinical and demographic characteristics of their cohorts (e.g., Berneiser et al., 2014; Henry et al., 2009, 2011; Prochnow et al., 2011), differences in adopted assessment tools (dynamic vs. static tasks) and presence of confounding variables such as mood and affective disturbances, cognitive deficits, and MS fatigue, all of which might contribute to deficits in identifying EFE.

Table 1 Studies assessing facial emotion recognition in multiple sclerosis

Note. Demographic and clinical data were expressed as mean unless indicated otherwise. Age and disease duration were expressed in years.

ADS-L=Allgemeine Depressions-Skala; BDI=Beck Depression Inventory; BEAST=Bodily Expressive Action Stimulus Test; BERT=Behavioral Emotion Recognition Test; BFRT=Benton Facial Recognition Test; BRB-N=Brief Repeatable Battery of Neuropsychological Tests; CIS=clinically isolated syndrome; CMDI=Chicago Multiscale Depression Inventory; DD=disease duration; FAB=Florida Affect Battery; FEEL=Facially Expressed Emotion Labeling; FEEST=Facial expressions of emotions, stimuli and tests; FSS=Fatigue Severity Scale; FST=Faces Symbol Test; GDS=Geriatric Depression Scale; HADS=Hospital Anxiety and Depression Scale; HAM-A=Hamilton Rating Scale Anxiety; HCs=healthy controls; IADS=International Affective Digitized Sounds and Picture System; IAPS=International Affective Picture System; IPS=information processing speed; MFIS=Modified Fatigue Impact Scale; MMSE=Mini Mental Status Exam; MS=multiple sclerosis; NP=not provided; PASAT=Paced Auditory Serial Attention Test; PCFAE=Test of Perceptual Competence of Facial Affect Recognition; PP=primary progressive; RAVLT=Rey Auditory-Verbal Learning Test; ROCFT=Rey–Osterrieth Complex Figure Test; RR=relapsing remitting; RT=reaction time; SART=Sustained Attention to Response Task; SD=standard deviation; SDMT=Symbol Digit Modalities Test; SEFCI=Screening Examination for Cognitive Impairment; SP=secondary progressive; TAS- 20=Toronto Alexithymia Scale; TMT=Trail Making Test; WAIS-R=Wechsler Adult Intelligence Scale-Revised; WCST=Wisconsin Card Sorting Test; WHOQoL-BREF=World Health Organization quality of life questionnaire.

Starting with disease characteristics, patients with higher physical disabilities seem to be the most affected on tasks assessing EFE recognition (Cecchetto et al., 2014; Prochnow et al., 2011). In this context, some studies were for a correlation between EDSS scores (Kurtzke, 1983) and deficits in judging EFE (Cecchetto et al., 2014; Berneiser et al., 2014), while others denied it (Henry et al., 2011; Jehna et al., 2010). As for disease subtypes, progressive MS patients seem to suffer from more pronounced deficits compared to RR MS patients (Beatty et al., 1989; Jehna et al., 2010; Jehna, Langkammer, et al., 2011; Parada-Fernández et al., 2015; Passamonti et al., 2009; Pinto et al., 2012). Regarding disease duration, it was found to be associated with deficits in labeling EFE in some (Cecchetto et al., 2014; Berneiser et al., 2014) but not all studies (Henry et al., 2011; Jehna et al., 2010).

Facial identity discrimination remains the main confounding variable in the recognition of EFE (Beatty et al., 1989; Di Bitonto et al., 2011; Pinto et al., 2012) since both processes share early common neural processing pathways. Although some MS studies suggest that deficits in the former be behind those in the latter (Beatty et al., 1989; Di Bitonti et al., 2011), most of the remaining data are not in favor of this assumption (Berneiser et al., 2014; Cecchetto et al., 2014; Krause et al., 2009; Parada-Fernández et al., 2015; Phillips et al., 2011; Prochnow et al., 2011).

MS fatigue is another frequent symptom that can be defined as a reversible alteration of cognitive task performance (Chalah et al., 2015). Up until now, only a few studies evaluated its relationship with EFE recognition. While one study found it to be associated with EFE task performance (Berneiser et al., 2014), others were not able to detect any significant relationship (Cecchetto et al., 2014; Henry et al., 2011).

Importantly, an interaction was previously found among emotions, mood, and cognition (Leppanen, 2006; Pessoa, 2008). In some studies, MS patients with deficits on EFE recognition tasks had also high depression scores (Beatty et al., 1989; Berneiser et al., 2014; Henry et al., 2011; Krause et al., 2009; Parada Fernandez et al., 2015; Phillips et al., 2011; Pinto et al., 2012; Prochnow et al., 2011) and poor cognitive abilities (Beatty et al., 1989; Henry et al., 2009, 2011; Krause et al., 2009; Parada-Fernández et al., 2015; Phillips et al., 2009; Prochnow et al., 2011). However, the correlation of EFE recognition with cognitive and mood scores remains controversial. While some studies are in favor of this relationship (Berneiser et al., 2014; Cecchetto et al., 2014; Henry et al., 2009; Jehna et al., 2010; Parada-Fernández et al., 2015; Pinto et al., 2012), others failed to detect any significant association (Cecchetto et al., 2014; Henry et al., 2011; Jehna et al., 2010; Krause et al., 2009; Prochnow et al., 2011).

Alexithymia is an additional variable that might interfere here (Grynberg et al., 2012). By definition, it is a personality trait characterized by difficulties in emotional identification, understanding, and description (Franz et al., 2008). Alexithymia was tackled in two MS trials evaluating EFE recognition. Although one of them featured higher levels of alexithymia in MS patients compared to HCs (Prochnow et al., 2011), the other did not find any group difference (Cecchetto et al., 2014), and neither of them detected an association between alexithymia and deficits in EFE recognition.

All in all, deficits in facial emotion recognition might occur early during MS, and do not seem to be restricted to progressive disease subtypes. However, disease characteristics and concomitant symptoms may contribute to such deficits. Heterogeneity in MS lesions location might be behind the different patterns of EFE recognition deficits encountered in various studies. One might speculate that during disease course, clinical and radiological MS progression can also be mirrored by a shift from an intact abilities to recognize EFE (Jehna et al., 2010, Jehna, Langkammer, et al., 2011; Passamonti et al., 2009; Pinto et al., 2012), to an isolated pattern of deficits (Henry et al., 2009, 2011; Krause et al., 2009; Prochnow et al., 2011), and finally to a global deficit (Beatty et al., 1989; Berneiser et al., 2014; Parada-Fernández et al., 2015).

Social perception of affective prosody in multiple sclerosis

In advanced MS stages, visual deficits can become very pronounced, and patients might depend on the perception of affective prosody for a successful social interaction. Only two MS studies have addressed this issue. In the first one, the authors used the comprehension and discrimination portions of Aprosodia battery (Ross et al., 1997) in a cohort of chronic MS patients (Beatty, Orbelo, Sorocco, & Ross, 2003). Compared to HCs, patients had worse performance on affective prosody, mood, and cognitive scales. Measures of affective prosody were positively correlated with cognitive scores but were not associated with mood disturbance, hearing loss, aphasia, treatment profile, or education. Unfortunately, patients’ clinical characteristics were not provided and their impact on prosody was not assessed.

In contrast with the first study, the second one included patients with early stage of RR MS (Kraemer, Herold, Uekermann, Kis, et al., 2013). Compared to HCs, patients had higher depression scores but did not differ on most of the cognitive scores. They poorly discriminated affective prosody, had lower accuracy in matching affective prosody to the facial expression for “anger,” but were able to recognize “happiness.” This finding is in line with the isolated pattern of deficits seen in some EFE studies (anger and fear in Henry et al., 2009, 2011; anger, sadness, and fear in Krause et al., 2009; anger, sadness, fear, and surprise in Prochnow et al., 2011). The observed deficits were unrelated to mood, cognitive performance, or physical disability. Unlike the cohort examined by Beatty et al. (2003), patients seen here performed better than HCs on matching affective prosody to facial expression for the emotion “fear.” This finding might be due to an increased sensitivity for recognizing “fear” in a population of young patients recently shocked by the diagnosis of a chronic disabling disease such as MS (Kraemer, Herold, Uekermann, Kis, et al., 2013).

ToM in MS

The majority of ToM studies in MS have adopted the Faces test (Baron-Cohen, Jolliffe, Mortimore, & Robertson, 1997; Baron-Cohen, Wheelwright, Hill, Raste, & Plumb, 2001), Reading the Mind in the Eyes test and Faux Pas test (Baron-Cohen, O’Riordan, Stone, Jones, & Plaisted, 1999) (for a summary, see Table 2). Faces test consists of 20 photographs of the same actress portraying different complex mental states (Banati et al., 2010; Baron-Cohen et al., 1997, 2001; Mike et al., 2013). During Reading the Mind in the Eyes test, also simply known as Eyes test, patients are asked to observe and comment on feelings or thoughts expressed in 36 face photographs depicting only the eye region (Banati et al., 2010; Baron-Cohen et al., 1999; Mike et al., 2013). Both are non-verbal tasks that evaluate “affective” ToM based on visual cues. The third one, Faux Pas test, is a verbal task that assesses “cognitive” ToM.

Table 2 Studies reporting deficits in the theory of mind in multiple sclerosis

Note. Demographic and clinical data were expressed as mean unless indicated otherwise. Age and disease duration were expressed in years.

BDI=Beck Depression Inventory; C&I=Conversations and Insinuations video-taped task; DD=disease duration; GDS=Geriatric Depression Scale; HADS=Hospital Anxiety and Depression Scale; HCs=healthy controls; IPS=information processing speed; MASC=Movie for the Assessment of Social Cognition; MFIS=Modified Fatigue Impact Scale; MS=multiple sclerosis; NP=not provided; PASAT=Paced Auditory Serial Attention Test; PP=primary progressive; RR=relapsing remitting; SDMT=Symbol Digit Modalities Test; SEFCI=Screening Examination for Cognitive Impairment; SP=secondary progressive; STAI=Spielberger Trait Anxiety Inventory; TASIT=The Awareness of Social Inference Test; TMT=Trail Making Test; ToM=theory of mind; WAIS=Wechsler Adult Intelligence Scale-Revised; WCST=The Wisconsin Card Sorting Test.

Some authors used exclusively verbal or non-verbal ToM tasks and found pronounced ToM deficits in their MS cohorts (Henry et al., 2009, 2011; Parada-Fernández et al., 2015; Roca et al., 2014). Others combined several tools that assess both ToM aspects (affective and cognitive) and obtained heterogeneous results. For instance, in two studies, patients were evaluated by the means of Reading the Mind in the Eyes test, Faces test, and Faux Pas test (Banati et al., 2010; Mike et al., 2013). Patients had an altered performance on the first (Banati et al., 2010; Mike et al., 2013) and second (Mike et al., 2013) tests, but had normal scores on the third one.

At a first glance, the absence of abnormality on the Faux Pas test appears surprising. However, this test seems to have low sensitivity to detect mentalization deficits as seen in some MS trials (Henry et al., 2011; Mike et al., 2013; Ouellet et al., 2010). For instance, in one study, MS patients had ToM deficits according to the Strange Stories task (Happé, Winner, & Brownell, 1998), yet they had normal performance on the Faux Pas test (Baron-Cohen et al., 1999). Such a discrepancy might be due to the fact that the Strange Stories task assesses a diversity of mental states and, unlike the Faux Past test, is not limited to detecting a “faux pas” in social interaction (Ouellet et al., 2010). Another plausible explanation is that MS patients may be more prone to mentalization deficits that depend on visual information processing than verbal processing (Mike et al., 2013). This might be due to a selective involvement of emotional networks at some point during the disease course. This assumption is supported by data from a fMRI study where verbal and non-verbal social information elicited different patterns of neural activation, respectively, in the precuneus/posterior cingulate cortex (PC/PCC) and amygdala (Kuzmanovic et al., 2012).

Besides classical static ToM tasks used in the aforementioned works, some authors used dynamic videotaped tasks presenting social interactions and obtained similar results (Genova, Cagna, Chiaravalloti, DeLuca, & Lengenfelder, 2016; Kraemer, Herold, Uekermann, Kis Wiltfang, et al., 2013; Ouellet et al., 2010; Pöttgen, Dziobek, Reh, Heesen, & Gold, 2013). Interestingly, in the study by Pöttgen et al., MS patients further exhibited insufficient mentalization abilities (2013), similar to those documented in autism (Baron-Cohen, 2000). Dynamic tests such as the one used here necessitate online complex processing abilities for an adequate interpretation of the exposed scenes. This might make of them better simulator of daily life events compared to the static written tests.

Last but not least, cognitive and affective ToM deficits in pediatric-onset MS patients have been documented by Charvet et al. (2014). The observed deficits were correlated with visuospatial attention and IPS scores (Charvet et al., 2014) and remained significant after accounting for cognitive functions.

As seen in EFE section, ToM studies enclosed several confounding factors such as MS fatigue (Henry et al., 2011), low intelligence quotient (Pöttgen et al., 2013), high mood scores (Banati et al., 2010; Henry et al., 2011; Kraemer, Herold, Uekermann, Kis, Wiltfang, et al., 2013; Mike et al., 2013; Parada-Fernández et al., 2015), and cognitive deficits (Banati et al., 2010; Charvet et al., 2014; Genova et al., 2016; Henry et al., 2009, 2011; Kraemer, Herold, Uekermann, Kis, Wiltfang, et al., 2013; Mike et al., 2013; Parada-Fernández et al., 2015; Roca et al., 2014).

While ToM scores were significantly associated with non-social cognitive performance (Charvet et al., 2014; Genova et al., 2016; Henry et al., 2009; Kraemer, Herold, Uekermann, Kis, Wiltfang, et al., 2013; Ouellet et al., 2010; Pöttgen et al., 2013; Roca et al., 2014) and clinical characteristics in some studies (EDSS scores in Pöttgen et al., 2013; progression rate in Banati et al., 2010); other studies did not detect any significant association between ToM performance and each of demographic or clinical characteristics (Charvet et al., 2014; Henry et al., 2011), cognitive profiles (Henry et al., 2011; Roca et al., 2014), and mood scores (Kraemer, Herold, Uekermann, Kis, Wiltfang, et al., 2013; Pöttgen et al., 2013; Roca et al., 2014), or MS fatigue (Roca et al., 2014).

Empathy in MS

Two studies reported low levels of empathy in MS patients (Gleichgerrcht, Tomashitis, & Sinay, 2015; Kraemer, Herold, Uekermann, Kis, Wiltfang, et al., 2013). In the first one, patients also had mood disturbance and cognitive decline (Kraemer, Herold, Uekermann, Kis, Wiltfang, et al., 2013). In the second, they had high levels of alexithymia and altered moral judgment (Gleichgerrcht et al., 2015). It is noteworthy that alexithymia could modulate empathy (Bird et al., 2010). Hence, the observed low levels of empathy and high levels of alexithymia might have contributed to an altered moral judgment (Gleichgerrcht et al., 2015).

Differently, other studies have documented high levels of empathy among MS patients. For instance, Benedict, Priore, Miller, Munschauer, and Jacobs detected a discrepancy between the levels of empathy as reported by patients and their informants (family members or friends) (2001). While informants reported low levels of empathy among patients, patients themselves generated high self-reporting. These results were substantiated by another report by Banati et al. who found higher empathy levels in patients with greater physical disability and shorter disease duration, both of which characterize a rapid disease progression (2010).

The findings of both studies might be explained by different views. For instance, severely impaired patients may exhibit more profound emotional misjudgment compared to relatively preserved patients, which can appear as high levels of empathy. Another possibility is that MS-related emotional stress might lead to a more focused emotional processing which can emerge as a higher estimate of empathy. The phenomenon of benefit finding can also account for the observed results (Pakenham & Cox, 2009). It provides an explanation on how constant challenges, such as those seen during MS course, might lead to positive learning and experiencing psychological growth (Pakenham & Cox, 2009).

Of interest, one study included a cohort of pediatric-onset MS patients and HCs but did not detect any significant group differences based on empathy questionnaires filled by parents (Charvet et al., 2014). However, one should keep in mind that parents-filled questionnaires do not necessarily reflect the patients’ impression.

The papers mentioned above are summarized in Table 3.

Table 3 Studies reporting alteration of empathy in multiple sclerosis

Note. Demographic and clinical data were expressed as mean unless indicated otherwise. Age and disease duration were expressed in years.

BDI=Beck Depression Inventory; DD=disease duration; HCs=healthy controls; HES=Hogan Empathy Scale; MS=multiple sclerosis; NP=not provided; PASAT=Paced Auditory Serial Attention Test; RR=relapsing remitting; SDMT=Symbol Digit Modalities Test; SP=secondary progressive; STAI=Spielberger Trait Anxiety Inventory; TAS-20=Toronto Alexithymia Scale; TMT=Trail Making Test; WAIS=Wechsler Adult Intelligence Scale; WCST=The Wisconsin Card Sorting Test.

NEURAL UNDERPINNINGS OF SOCIAL COGNITIVE DEFICITS IN MULTIPLE SCLEROSIS

Five MRI studies (Table 4) investigated the neural basis of social cognitive deficits in MS (Beatty et al., 2003; Jehna, Langkammer, et al., 2011; Krause et al., 2009; Mike et al., 2013; Passamonti et al., 2009).

Table 4 MRI Studies evaluating social cognition in multiple sclerosis

Note. Demographic and clinical data were expressed as mean unless indicated otherwise. Age and disease duration were expressed in years.

CC=corpus callosum; DD=disease duration; DSM-IV=Diagnostic and Statistical Manual for Mental Disorders, 4th edition; FFA=fusiform facial area; fSTS=facial area of the superior temporal sulcus; GCC=genu of corpus callosum; GM=gray matter; LL=lesion load; LV=lesion volume; MS=multiple sclerosis; NP=not provided; PP=primary progressive; RR=relapsing remitting, SCC=splenium of corpus callosum; SP=secondary progressive; PCC=posterior cingulate cortex; UF=uncinated fasciculus; VLPFC=ventrolateral prefrontal cortex; WM=white matter.

STRUCTURAL NEUROIMAGING DATA

Concerning facial emotion perception, Krause et al. performed voxel-based lesion symptom mapping in MS patients with or without deficits in EFE recognition (Krause et al., 2009). Although lesion volume did not statistically differ between both patient groups, poor performance on facial affect task was correlated with lesions in left temporal WM (Figure 1), an area containing several connections between the OFC and the STS (Cavada, Company, Tejedor, Cruz-Rizzolo, & Reinoso-Suarez, 2000). Therefore, the observed impairment in EFE recognition might be due to interruption of the fibers responsible for visual processing of emotionally relevant stimuli. To note, this study also contained fMRI data that will be analyzed in the following section.

Fig. 1 (a, b) Axial and (c) sagittal brain views illustrating the structural correlates of social cognitive deficits in multiple sclerosis. FFA: left fusiform facial area; R: right; SCC: splenium of the corpus callosum; TP: left temporal pole; *: left temporal white matters lesions; left uncinated fasciculus not shown.

Our insight into ToM in MS arises from the study by Mike et al. who compared structural MRI data between MS patients and healthy controls (Mike et al., 2013). In addition to the observed social cognitive deficits in the patients’ group, inverse correlations were found between each of the Faces and Reading the Mind in the Eyes tests and total T1 lesion volume (tT1LV). More interestingly, patients’ performance on the Faces test was inversely associated with regional T1 lesion volume (rT1LV) of CC (genu and splenium) and several fasciculi (bilateral uncinated fasciculus, right inferior longitudinal and fronto-occipital fasciculi); with regional T2 lesion volume of CC (genu) and left fornix, and with cortical thinning of many areas (i.e., bilateral FFA, right entorhinal cortex).

Second, performance on Reading the Mind in the Eyes test was inversely correlated with rT1LV of the CC (splenium) and cortical thinning of left anterior inferior temporal gyrus (temporal pole), left FFA and right caudal middle frontal gyrus (right premotor frontal eye field, FEF). However, performance on the Faux Pas test did not correlate with any of the studied parameters. The multiple regression analysis also revealed several issues. For instance, rT1LV of left uncinated fasciculus was an independent predictor of the Faces test performance. Besides, performance on Reading the Mind in the Eyes test was predicted by rT1LV of the splenium of CC and cortical thickness of left FFA and left temporal pole (Figure 1).

Of interest, all of these structures are neural nodes which take parts of social cognitive networks. For instance, the genu and splenium of CC links, respectively, identical anterior (prefrontal and premotor) and posterior cortical areas (occipital, parietal, and temporal lobes) involved in emotional, cognitive, and visual processing (Park et al., 2008). The role of FFA has been already seen in facial identity discrimination and emotion recognition (Haxby, Hoffman, & Gobbini, 2000, 2002; Zaki, Hennigan, Weber, & Ochsner, 2010). The temporal pole enables the confrontation of perceived social and emotional cues (visual information) with stored general knowledge (contextual information) (Frith & Frith, 2006).

As for affective prosody, the available data are derived from only one study in which the comprehension of affective prosody did not correlate with any studied parameters, namely the CC size and the extent of right or left hemispheric lesions (Beatty et al., 2003). The absence of correlations might be due to the use of basic MRI measures which could have been different with the adoption of non-conventional MRI techniques (Rovaris, Comi, & Filippi, 2001).

Functional Neuroimaging Data

The available fMRI studies in MS patients focused on EFE recognition. The first one included early stage RR MS patients with intact social cognitive abilities and healthy controls (Passamonti et al., 2009). The imaging acquisition took place during the execution of an active task that consisted of processing facial emotions relative to neutral stimuli (geometric shapes such as circles, or horizontal and vertical ellipses) (Passamonti et al., 2009). Compared to their healthy counterparts, patients exhibited a hyperactivation within bilateral prefrontal areas (VLPFC) and left posterior cortices (PC, superior parietal cortex) (Passamonti et al., 2009). They also displayed a reduced pattern of functional connectivity between prefrontal cortices (ventrolateral and medial parts) and left amygdala (Figure 2).

Fig. 2 (a, b) Axial and (c) sagittal brain views illustrating the functional changes during social cognitive performance in multiple sclerosis. AMY: left amygdala; mPFC: medial prefrontal cortex; R: right; SPC: superior parietal cortex; vlPFC: ventrolateral prefrontal cortex; upward arrows: hyperactivation pattern seen in multiple sclerosis patients with preserved social cognitive abilities compared to healthy controls; downward arrows: hypoactivation pattern seen in multiple sclerosis patients with social cognitive deficits compared to those with intact abilities; dashed lines: reduced functional connectivity in the tagged networks; left superior parietal cortex not shown.

It is noteworthy that a lateralization pattern of amygdalar activation exists in the normal human brain during emotional processing, with the left amygdala being more activated than the right one (Baas, Aleman, & Kahn, 2004). In fact, by communicating with posterior brain regions that are involved in visual processing, the amygdala has a pivotal role in decoding emotionally significant sensory stimuli and by doing so, it participates in the formation of emotional memory. Also, the dialogue between the amygdala and prefrontal cortex is crucial in the processing of emotional information (Ghashghaei, Hilgetag, & Barbas, 2007).

The findings of this study were supported soon after by another one in which early stages RR MS patients had normal performance on cognitive and facial affect recognition tasks compared to healthy controls (Jehna, Langkammer, et al., 2011), yet they exhibited a hyperactivation within fusiform gyri and other right cortical areas (i.e., fontral pole, ACC, and paracingulate cortex) during the performance of neutral faces (facial identity); and hyperactivation of PC and PCC during the performance of “anger” (left activation) and “disgust” (right activation) contrasted to neutral faces (Figure 2).

In addition to the above-stated role of the amygdala, PCC is implicated in mediating the interactions between emotional and memory-related processes (Maddock, Garrett, & Buonocore, 2003). More interestingly, The PC seems to be divided into two parts, an anterior one dealing with self-centered mental imagery strategies and a posterior one in charge of episodic memory retrieval (Cavanna & Trimble, 2006).

A third study by Krause et al. provides additional evidence. MS patients with or without deficits in facial affect recognition underwent functional imaging (Krause et al., 2009). Compared to the preserved MS group, the impaired group showed a hypoactivation in the facial area of STS, left VLPFC and insula, all of which are normally implicated in the social perception of EFE (Figure 2). In the whole patients group (impaired and preserved), the accuracy on facial affect task was correlated with the increased activation within the left anterior insula and left VLPFC.

To sum up, all of the three studies, featured a hyperactivation pattern in MS patients with preserved social cognitive abilities (Jehna, Langkammer, et al., 2011; Krause et al., 2009; Passamonti et al., 2009). To explain these findings, one can speculate that compensatory processes occur early in the disease course to restrain the social cognitive deficits that might arise from MS-related gray (GM) and white (WM) matter pathologies (Mainero et al., 2004; Mainero, Pantano, Caramia, & Pozzilli, 2006; Pantano et al., 2002; Rocca et al., 2009; Sumowski, Wylie, Deluca, & Chiaravalloti, 2009; Staffen et al., 2002; Sweet, Rao, Primeau, Durgerian, & Cohen, 2006; Wegner et al., 2008). These mechanisms could radiologically manifest as increased regional activation patterns (Jehna, Langkammer, et al., 2011; Passamonti et al., 2009) or reduced functional connectivity of some brain networks (Passamonti et al., 2009). Saying so, the increase in lesions load and subsequent diffuse neural disorganization might lead to reduced or maladaptive plasticity processes (Citri & Malenka, 2008; Morgen et al., 2004). This would cause poor social cognitive performance and lead to regional hypoactivation on fMRI as seen with the impaired MS group of the third study (Krause et al., 2009).

SOCIAL COGNITIVE DEFICITS IN MULTIPLE SCLEROSIS: A PRIMARY OR SECONDARY SIGNATURE

Several cognitive domains could be altered in MS patients, and this population commonly suffers from mood disturbances, fatigue, alexithymia, and sleep problems. Hence, one might ask whether social cognitive deficits in MS constitute a primary phenomenon or rather result from the previously described confounders. Although this issue is still a matter of debate, the influence of these variables on social cognition was addressed in some studies and deserves to be mentioned here.

For instance, despite the high prevalence of alexithymia in MS patients (Bodini et al., 2008; Chahraoui et al., 2008, Chahraoui, Duchene, Rollot, Bonin, & Moreau, 2014; Gay, Vrignaud, Garitte, & Meunier, 2010), only few studies controlled for this factor (Cecchetto et al., 2014; Gleichgerrcht et al., 2015; Prochnow et al., 2011). It is worth noting that patients with alexithymia were found to have social cognitive deficits (Grynberg et al., 2012) and display abnormal pattern of brain activation during EFE processing (Kano et al., 2003).

Moreover, alexithymia was associated with decreased GM volume in regions such as the ACC, amygdala, and insula (Ihme et al., 2013), which had abnormal activation pattern in fMRI studies assessing social cognition in MS (Jehna, Langkammer, et al., 2011; Krause et al., 2009; Passamonti et al., 2009). These facts altogether should prompt screening for alexithymia in future assessment of social cognition.

Depression also appears to be a frequent symptom in MS patients (Feinstein, 2011) and is linked to pathological changes in bilateral frontal regions which are key components in social cognitive processing (Gobbi, Rocca, Riccitelli, et al., 2014). Admitting the influence of mood on social cognition in MS (Berneiser et al., 2014; Pinto et al., 2012; Parada-Fernández et al., 2015) and other clinical settings (Asthana, Mandal, Khurana, & Haque-Nizamie, 1998; Leppanen, 2006; Persad & Polivy, 1993; Suslow et al., 2004), an optimal evaluation of social cognition should account for this variable.

As for MS fatigue per se, its underlying pathophysiology lies in the so-called “cortico-striato-thalamo-cortical loop” (for reviews, see Chalah et al., 2015), which includes pathological alterations of many cerebral tracts such as UF, CC, and IFOF (Bisecco et al., 2016; Gobbi, Rocca, Pagani, et al., 2014). Importantly, abnormalities in these brain structures are also documented in social cognition studies and were inversely correlated with MS patients’ performance on ToM tasks (Mike et al., 2013). The fact that both fatigue and social cognitive deficits in MS share several anatomical pathologies should pave the way for a better control of MS fatigue in upcoming trials.

Furthermore, MS patients commonly suffer from cognitive symptoms (Ayache et al., 2015; Kesselring & Klement, 2001; Vázquez-Marrufo et al., 2014) and significant correlations were found in MS patients between social cognitive performance and several non-social cognitive abilities, such as attention, processing speed, working memory, learning, and executive functions (Benedict et al., 2001; Berneiser et al., 2014; Cecchetto et al., 2014; Charvet et al., 2014; Genova et al., 2016; Henry et al., 2009; Jehna et al., 2010; Kraemer, Herold, Uekermann, Kis, et al., 2013b; Ouellet et al., 2010; Pinto et al., 2012; Pöttgen et al., 2013; Roca et al., 2014). For these reasons, evaluating non-social cognitive abilities in forthcoming works might help better understand their relationship with social cognition.

Nevertheless, altered moral judgment could also co-occur with social cognitive deficits and has been related to pathological changes within the TPJ, the latter region being an important component of the ToM circuit (Samson, Apperly, Chiavarino, & Humphreys, 2004; Young, Camprodon, Hauser, Pascual-Leone, & Saxe, 2010). Lastly, sleep disorders, frequently encountered in MS, might as well influence social cognition and deserve to be taken into consideration (Beattie, Kyle, Espie, & Biello, 2015).

CONCLUSION

Taken together, these data provide convergent evidence on the occurrence of social cognitive deficits even at early stages of MS. Deficits in recognizing negative emotions seem to be more pronounced that those of positive ones among MS patients.

Here, two questions might arise: (i) how individuals with MS could preserve their social cognitive performance early in the disease process despite the continuous accumulation of brain lesions and then, at a certain point in their life, start experiencing deficits; and (ii) why an inhomogeneity in social cognitive performance was observed across MS studies. The hypothesis of “functional brain reorganization” could answer the first question. In fact, in front of the neural damage encountered in MS, compensatory neuroplasticity mechanisms and functional reorganization would take place in an attempt to limit subsequent behavioral deficits that might arise from MS-related pathologies. Later on, the increase in disease burden may exhaust the adaptive mechanisms and functional reserves (Cader, Cifelli, Abu-Omar, Palace, & Matthews, 2006; Pantano et al., 2005) leading to poor social cognitive performance.

The second question could be addressed in light of “cognitive reserve hypothesis.” Cognitive reserve is thought to be a moderator between the amount of brain damage and the extent of clinical outcome (Stern, 2012). This could apply to MS patients in a way that those with higher cognitive reserve might experience less social cognitive deficits than others with similar extent of brain lesions (Sumowski et al., 2009; Sumowski & Leavitt, 2013).

Other important issues remain unresolved and need a careful assessment in future studies. First, the prevalence of social cognitive deficits in MS is still undetermined. In fact, a large number of reported trials dealt with heterogeneous MS cohorts with different disease subtypes, wide ranges of physical disability and advanced stages, which make them more prone to social and non-social cognitive deficits.

Second, whether the social cognitive deficits constitute a primary impairment, or they result from cognitive deficits, or other MS-related symptoms is still a matter of debate. Henceforth, future in-depth assessment of social cognition should focus on confounding factors and the onset of these deficits.

Third, neural components of social cognition need further deciphering. Thus, coupling non-conventional neuroimaging and neurophysiological modalities with more detailed neuropsychological testing could be of particular help.

Fourth, the evaluation of social cognition might benefit from combining static, and dynamic assessment tools since videotaped tasks seem to have better accuracy than classical static tests in evaluating social cognition (Dziobek et al., 2006).

In summary, these considerations would shed the light on the social cognitive deficits in MS and may open a venue for an optimal multidisciplinary approach in MS patient care. By doing so, affected patients will be able to overcome their interpersonal difficulties and improve their QoL.