Original Paper Received: June 23, 2015 Accepted: November 11, 2015 Published online: December 12, 2015
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
Assessment of Social Cognition and Theory of Mind: Initial Validation of the Geneva Social Cognition Scale Marie-Dominique Martory a Alan John Pegna a, b Laurent Sheybani c Mélanie Métral a Françoise Bernasconi Pertusio a Jean-Marie Annoni d
a
Neuropsychology Unit and Laboratory of Experimental Neuropsychology, Department of Neurology, Geneva University Hospital, b Faculty of Psychology and Educational Science and c Department of Neuroscience, University Medical Center (CMU), University of Geneva, Geneva, and d Neurology Unit, Department of Medicine, Faculty of Sciences, University and Hospital of Fribourg, Fribourg, Switzerland
Abstract Background: Social cognition is widely studied in neurology. At present, such evaluations are designed for research or for specific diseases and simple general clinical tools are lacking. We propose a clinical evaluation tool for social cognition, the Geneva Social Cognition Scale (GeSoCS). Methods: The GeSoCS is a 100-point scale composed of 6 subtests (theory of mind stories, recognition of social emotions, false beliefs, inferences, absurdity judgement and planning abilities) chosen from different validated tests of social and cognitive evaluation. Eighty-four patients with neurological disorders and 52 controls participated in the study. Evaluation duration lasted 20–60 min. Results: Mean scores were 92.6 ± 4.5 for controls and 76.5 ± 15.3 for patients and differentiate patients and controls in all subtests. With a cut-off score of 84, the scale had a sensitivity of 62% and a specificity of 94%. In our stroke subgroup, right CVAs failed in cartoons, inferences, ‘mind in the eyes’, and in the temporal rule task while
© 2015 S. Karger AG, Basel 0014–3022/15/0746–0288$39.50/0 E-Mail [email protected] www.karger.com/ene
left CVAs were impaired in verbal/discourse tasks (social cognition, inferences, absurd stories, and cartoons. Conclusions: The GeSoCS is a medium duration assessment tool that appears to detect and characterize significant social impairment in neurological patients. © 2015 S. Karger AG, Basel
Introduction
The last decade saw the development of social neuroscience, which integrated social psychology, cognitive neuropsychology and neurosciences and raised the question of whether social cognition exists as a specific domain (for review see Ward [1]). Social cognition refers to the abilities and processes allowing people to interact and to understand the behaviour of others. Among these different processes, the capacity to attribute mental states such as feelings and beliefs and then to predict behaviour is called theory of mind (ToM) and plays a determining role in social adaptation. ToM was first studied in primates [2], and subsequently in developmental psychology [3]. Its first clinical application involved autistic persons [4]. Social cognition has also been addressed extenDr. Marie-Dominique Martory Neuropsychology Unit, Department of Neurology Geneva University Hospital 4, Gabrielle-Perret-Gentil, 1205 Geneva (Switzerland) E-Mail mdmartory @ bluewin.ch
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
Key Words Clinical scale · Neuropsychological evaluation · Brain damage · Social emotions · False beliefs · Inferences · Absurdity judgement · Executive planning
GeSoCS as a ToM Clinical Tool
Table 1. Demographic characteristics of control and patient groups
Age, mean (SD) Gender Male Female Handedness Right Left Level of education, years 12
Control group (n = 52)
Patient group (n = 84)
42.2 (16.4)
53.3 (15.5)
20 32
49 35
51 1
71 13
1 51
11 73
Table 2. Patient group aetiologies Patient group (n = 84) Age, mean (SD) Gender Male Female Handedness Right Left
Stroke Trauma Tumor Dementia (n = 31) (n = 17) (n = 7) (n = 17)
Encephalopathy (n = 11)
58 (12.8) 41 (13)
44.8 (16.6)
51 (14) 66.4 (8.8)
15 16
13 4
4 3
11 6
7 4
28 3
14 3
6 1
14 3
8 3
neurological patients rather than a short screening tool. Deficits identified with this tool would then allow subsequent detailed examination using the existing tests.
Method Participants All subjects gave written informed consent for their participation in accordance with the Declaration of Helsinki. The study was approved by the local Ethics Committee, (Commission d’Ethique de la Recherche Clinique, Geneva University Hospital, Switzerland; CER 10-073). The population was composed of 52 healthy French-speaking, volunteers, with no known neurological or psychiatric disorders. Demographic data are summarized in table 1. Eighty-four brain-damaged French-speaking patients were examined consecutively during their hospitalization in the neurological clinic. The patient population presented various aetiologies (stroke, epilepsy, trauma, tumour, dementia; table 2). Participants with CVA were assessed in the post-acute stage (about 10 days to
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
289
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
sively using brain imaging techniques. Along these lines, fMRI studies, for example, have shown that the temporal lobe, the temporal-parietal-junction, the medial prefrontal cortex and the posterior cingulate cortex are specifically involved in ToM [5–7]. In the field of clinical neuroscience, investigations of brain-injured patients with acute or neurodegenerative disease have also revealed different types of impairments in social cognition and the understanding of people’s behaviour [8]. Indeed, ToM constitutes a complex domain that involves numerous cognitive processes, including executive functions [9, 10] and, in certain conditions, normal language development [11]. Thus, the evaluation of social cognition and ToM in adults relies on a wide range of tasks of varying complexity. The most common involves false beliefs and can be used with young children. In particular, first-order beliefs necessitate understanding another person’s mental state, while second-order beliefs, which are more complex, predict what a person thinks about a third person’s feelings [3, 4]. More complex tasks used to assess the mental state of others require that participants identify a person’s emotional state on the basis of the expression in their eyes (the ‘mind in the eyes’ test) [12, 13], detect a social ‘faux pas’ [14], or understand mental inferences or indirect speech as in metaphors and jokes. Children are able to resolve these tasks at different ages (age 4 for the first order false belief and 9 or 11 for the faux pas). In brain-damaged patients, a number of authors have proposed that patients with brain injuries would process simple false beliefs stories correctly but would fail in more complex tasks [14]. Despite the great number of studies in social neuroscience investigating patients with acute acquired or neurodegenerative brain damage, research in the field has not yet provided tests that are broad enough for use in the clinic and have focused either on specific pathologies, such as multiple sclerosis [15] or neurodegenerative groups (for review see Henry et al. [16]), or alternately, have addressed specifics such as false beliefs [17], or faux pas recognition [14]. Additional global testing, as described earlier, is often too long for clinical practice. To our knowledge, there is yet no general test that allows neuropsychologists to assess ToM and social cognition function more globally in everyday clinical practice. The aim of this study is thus to offer a short assessment tool that is easy to use in everyday neuropsychological practice, which has been developed from existing tests described in the literature such as the mind in the eyes test or false beliefs. This scale was intended as a clinical tool for potential impairment in social cognition found in
Assessment of ToM and Social Cognition The test was developed internally on the basis of the literature reported above and spanned aspects most frequently explored in social cognition. Each of them was evaluated through specific tasks as described in the literature and combined in order to form a scale with a total score of 100 points (online suppl. appendix, see www. karger.com/doi/10.1159/000442412). Tasks 1, 2, 3 and 4 were specific to social cognition and comprehension of inferences. For each of them, the patients were asked to read a story and to answer specific questions involving ToM, but also to respond to simple control questions in order to test linguistic comprehension. The lengths of the stories were similar. In addition, precautions were taken to ensure that the questions and the syntactic structures of the sentences were equivalent and simple enough to be understood by most patients. Task 1 was composed of 5 written ToM and social cognition stories: one ‘faux pas’ situation, one ‘normal’ situation (‘Helen’ and ‘Jim’ stories [14]), one ToM story (‘burglar’ story [18]), and one in-house ToM story involving a self-perspective1. Task 2 involved 10 items chosen from the ‘reading the mind in the eyes’ test [12, 13]. The original test is composed of 36 pairs of eyes each evoking an emotion. The subject has to choose between 4 written propositions the one that best describes the emotion. The items selected (1-4-5-8-14-16-18-30-33-35) were retained as being the most easily identifiable on the basis of a pilot experiment with 25 participants. Task 3 was made up of 5 ToM picture stories [19] representing first- and second-order false beliefs (e.g. A man is sitting on a bench with his dog beside him and he is preparing to eat a sandwich; while he is throwing the paper packaging, a bird steals the sandwich; on seeing the sandwich missing, the man scolds the dog. Question: why does the man scold his dog?) and one story not requiring inference (A boy sits on his sofa when he hears the doorbell ring. He thinks that it is the mailman but he smiles when he sees that it is his friend. Question: Why does the boy seem happy?). Task 4 was composed of 5 stories that the patient had to read before answering questions about possible or impossible inferences [20, 21] e.g. ‘The more Willy loses at the casino, the more affectionate he is with his wife; apparently, he was not very lucky in his game this evening’. Control question without inference: Did Willy win a lot at the casino this evening? (response: no). Question with impossible inference to make: Is Willy’s wife rich? (response: we cannot know). Question with an inference to make:
1 In a fMRI study with healthy subjects, Vogeley et al. [2001], found different patterns of brain activation, as the subjects took the first or third person perspective and Gambini, Barbieri, Scarone [2004], supported this idea in a study involving schizophrenic subjects.
290
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
Is Willy rather affectionate with his wife this evening? (response: yes). Tasks 5 and 6 are beyond ToM strictly speaking but are used to determine clinically a general executive impairment, which can be associated (although not necessarily) with deficits in social cognition [9]. Task 5 required that patients correctly detect an absurdity in 4 stories and accept a control story (e.g. ‘The management of a railroad company noticed that in train accidents, injuries were for more serious in the tail wagons. Consequently, it was decided to remove the last wagon of each train. What do you think of that?’). This test was retained as it constituted an interesting match with the social cognition stories that also required a ‘meta’ representation, albeit without the social component (adapted from Binet and Simon [22]). Task 6 was not directly related to social cognition, but was presented as a general measure of executive function. It is a French adaptation of the Counter test [23] and consists in determining the rules underlying the temporal appearance of colour counters (e.g. black-white-white, black-white-white…). This test was chosen for its ease of use and its good prediction of mental flexibility in frontal impairment [24]. Every verbal or picture task contained control questions measuring literal comprehension but excluding ToM or social aspect. The control questions were not scored but were included to ensure correct linguistic comprehension. The item ‘Jim’ that constituted a full question in the test was an exception. The duration of the whole test ranged from about 40 min in a healthy subject to between 60 and 70 min in the brain-damaged patients. In this latter group, participants could re-read the stories or obtain repetitions of the questions. Patients and control subjects were examined by clinically qualified neuropsychologists. Procedure The Geneva Social Cognition Scale (GeSoCS) was tested in addition to the patients’ standard neuropsychological examination, which will not be detailed in this study. In controls, no additional evaluation was performed, as participants were considered to be cognitively normal. However, a semi-structured interview was carried out to confirm the absence of any indicators of cognitive impairment, either anamnestically or behaviourally. The subjects were individually tested in a single session of about 40–60 min. The sessions took place at the University Hospital of Geneva, in the Neuropsychology Unit. During the test, they were asked to answer the questions freely and no solutions were proposed. If necessary, the question was repeated. For each item, full points were given when the answer was complete and correct. Half points were given if the answer was incomplete, or if the answer was correct but not the explanation. Statistical Analysis The results of the whole group are described first, followed by that of the patients and controls (global scores, then subtests). Significant differences and interactions between the controls and the patients were calculated using an analysis of variance (ANOVA; statistical analyses were performed using Statistica 12 software). For the smaller and more heterogeneous groups (TBI, dementia), additional testing was carried out using nonparametrical tests.
Martory/Pegna/Sheybani/Métral/ Bernasconi Pertusio/Annoni
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
3 weeks after their stroke and after stabilization for other aetiologies). All patients’ diagnoses were based on a clinical evaluation, a laboratory evaluation including an MRI, and a formal neuropsychological assessment. Exclusion criteria included a history of prior psychiatric disorders, delirium, global aphasia and severe dementia (≥CDR3). Subsequent statistical inquiries as to the effects of levels of education on our tests showed that this factor did not influence the results. The different levels were therefore merged.
Social cognition
Mind in the eyes
Cartoons
Inferences
Absurd stories
Temporal rules
Total
0.80
0.40
0
–0.40 Controls
Fig. 1. Comparison of mean z-scores of
–0.80
Patients
controls and patients for all the tasks.
Results
Table 3. Results: mean scores and SD for controls and patients for
all tasks
Comparison of z-Scores: CVA Patients vs. Controls The 31 patients with CVA demonstrated approximately the same pattern as the entire patient group with altered performances compared to the controls as shown in figure 2. GeSoCS as a ToM Clinical Tool
Tasks
Controls (n = 52), mean (SD)
Patients (n = 84), mean (SD)
Social cognition Mind in the eyes ToM cartoons Inferences Absurd stories Temporal rules Total
19.36 (1.20) 15.09 (2.28) 19.55 (0.79) 19.17 (1.49) 9.65 (0.94) 9.76 (0.85) 92.61 (4.51)
16.02 (4.08) 12.11 (3.81) 17.22 (3.5) 15.83 (4.4) 7.78 (2.32) 7.54 (3.42) 76.51 (15.32)
The ANOVA confirmed that the difference was significant (F(15, 395) = 4.4,651, p < 0.05). However, when considering the side of the lesion, a post hoc Fisher test showed different patterns of impairment in the case of right CVA, left CVA or bilateral CVA. Right CVAs did not differ from the controls in the first task of social cognition and were quite correct in ‘faux pas’ stories, which required criticism (p = 0.06), but they failed in cartoons, in inferences and in tasks in which they had to guess a feeling, such as the ‘mind in the eyes’ test (p < 0.05), or an intention, such as in the temporal rule task (p < 0.05). Left CVAs were impaired in all verbal tasks, social cognition tasks (p < 0.05), inferences (p < 0.05), absurd stories, and also in cartoons (p < 0.05). Conversely, they have Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
291
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
Descriptive statistics for the controls and the patients are provided in table 3. The controls obtained a total mean score of 92.6/100 (SD 4.5); for each test, the result was close to the maximum score with a standard deviation equal to 1, with the exception of the ‘mind in the eyes’ test, (mean 15/20, SD 2.3). There was no significant difference due to gender (women 92.4, men 93). The patients obtained a total mean score lower than the controls (76.51) with a large dispersal (SD 15.32) and performances for each test were lower than the controls. For every task, the performance of the patients and the controls were expressed as z-scores of the mean total group’s performances (patients + controls; fig. 1). An ANOVA comparing the 2 groups’ z-scores revealed a significant difference in performance between the controls and the patients (F(1, 134) = 55.035, p < 0.00001), and a post hoc test showed that the patients were impaired compared to the healthy subjects in all tasks. We now considered the mean z-score for the CVA population compared to the mean z-score of the control group.
1.20
Controls
Right CVA
Left CVA
Bilateral CVA
0.80
0.40
0
–0.40 Social cognition –0.80
Mind in the eyes Cartoons Inferences
–1.20
Temporal rules
Fig. 2. Comparison of mean z-scores of –1.60
shown significant difficulties in the ‘mind in the eyes’ task (p = 0.07). Finally, bilateral CVAs were significantly impaired (p < 0.05) in all 6 tasks. Determining Pathological Function: The Cut-Off Score Based on these data, we attempted to determine a cutoff score for clinical use. The values for each test are given in table 4 using a cut-off score of minus 2 SDs. Using these values, 52 patients fell below the cut-off (84) and 32 remained above; among the 31 CVA patients, 19 were below the cut off (9/15 left, 5/8 right and 5/7 bilateral CVA); 3 healthy subjects fell below the cut-off and 49 were above. The cut-off of 84 allowed to significantly differentiate between controls and patients (χ2(1) = 42.02; p < 0.0001). Sensitivity was at 61.9% and specificity at 94.23%. An analysis computing the receiver operating curve (ROC) with these cut-off scores showed an estimated area of 0.78 under the curve. Thus, the probability of correctly classifying a neurological patient as pathological on the basis of our test is 78% (fig. 3).
Discussion
In order to provide a simple tool for evaluating social cognition, we designed the GeSoCS, which allows rapid and easy assessment of social cognition in hospitalized neu292
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
Table 4. Cut-off for each subtest and the total test
Tasks
Controls, mean (SD)
Cut-off (rounded to nearest unit)
Social cognition Mind in the eyes ToM cartoons Inferences Absurd stories Temporal rules Total
19.36 (1.20) 15.09 (2.28) 19.55 (0.79) 19.17 (1.49) 9.65 (0.94) 9.76 (0.85) 92.61 (4.51)
17 11 18 16 8 8 84
rological patients in a daily clinical practice. This screening test was constructed with items taken from existing tasks previously described in the literature for evaluating the ToM. These tasks were subsequently adapted and grouped in order to form a scale graded from 0 to 100. To the best of our knowledge, it is the first global tool adapted for everyday clinical work as most of the studies on clinical patients until now have used extensive batteries. From a clinical perspective, GeSoCS appears to conform to the criteria for a consistent scale. First, the GeSoCS uses multiple aspects of validated tests for assessment of ToM, a ‘complex, multifaceted cognitive ability’ [25]. On average, GeSoCS lasts 40–60 min. Generally, this duration is easily tolerated by neurological patients, although one cannot exclude potential attentional deficits during evaluation. Martory/Pegna/Sheybani/Métral/ Bernasconi Pertusio/Annoni
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
right CVA, left CVA, bilateral CVA and controls for all the tasks.
Absurd stories
100
Sensitivity
80
60
40
20
0 0
20
40
60
80
100
100 – Specificity
Fig. 3. ROC with an estimated area of 0.78 under the curve.
This scale was administered by experienced neuropsychologists to 136 volunteers: 52 healthy subjects and 84 patients who were either hospitalized in the post-acute stage or stabilized. The results showed that patients performed significantly more poorly than healthy controls on average (p < 0.05). In our sample, patients differed from controls not only in the global score but also in all subtests. This ability to differentiate between both groups, despite the patients’ heterogeneous results (SD 15.368 in the patients’ group vs. 4.5 in the controls), offers a satisfying perspective in the use of this tool. The CVA patients, who constituted the main group, performed in a significantly worse manner in all events. Interestingly, the results differed according to the lateralization of the lesion: left CVAs were significantly different from the controls in all the tests (p < 0.05) except the ‘mind in the eyes’ test. Some of these left CVA patients suffered from aphasia and, although it has been shown that ToM abilities are not fundamentally impaired in this population [26], it is possible that language difficulties and deficient working memory may play a part in these results, and also in ToM picture tasks where all the answers require a verbal argumentation. Nevertheless, this group was not significantly different from the controls in the interpretation of the ‘mind in the eyes’ test (p > 0.5). The right CVAs showed a different pattern with respect to the controls and to the left CVAs: they failed in the ‘mind in the eyes’ test, in temporal rule tasks (which nevertheless do not require an elabGeSoCS as a ToM Clinical Tool
orated verbal response) and ToM pictures (cartoons) but performed correctly in the cognition scale. The bilateral strokes, as expected, had more difficulty in all tasks. As regards CVA, the recent data in the literature show that social cognition is affected during the acute phase, particularly if the lesion involves anatomical regions such as the parieto-temporal junction and fronto-parietal areas. However, this disorder can evolve favourably in the same way as other cognitive functions [27]. Interest and Limitations of Current Study The GeSoCS scale has the advantage of being composed of tests that are described and validated in the literature as being sensitive to social cognition and ToM. Instructions and scoring methods are simple to use for neuropsychologists and behavioural neurologists. Its duration and the absence of heavy equipment allow it to be used quite easily with any patient suspected of presenting disorders of social cognition. In our study, the number of healthy controls was slightly limited, but their scores were high and homogeneous, and most tests have already been reported extensively in the literature and their use has been confirmed. Their average age was younger (11 years) than the average age of the patients. This difference might have influenced our findings. However, a recent study that considers the effect of aging on mind-reading abilities in healthy people has suggested the absence of any difference between younger and older subjects [28]. The test cannot be understood as a screening, but must be considered as a more descriptive test. It took often between 30 min and 1 h, and had to be done in a different session than the usual neuropsychological evaluation. Such durations have also been used in other populations with clinically oriented test (see social cognition impairments in relation to general cognitive deficits, injury severity, and prefrontal lesions in traumatic brain injury patients [29] or in frontotemporal dementia [30]). So it must be considered a clinical test with a capacity of describing the type of deficit in clinical cognition in neurological patients. Our experience suggests that patients easily support this non-challenging evaluation, except strongly demented patients and severe aphasia. Our data showed a good specificity of 94% for a cut-off of 84, making the GeSoCS a highly efficient method for detecting deficits of social cognition. The sensitivity of 61% is of course a limitation in our study and the GeSoCS cannot be considered an interesting screening tool for neurological disorders. It has, however, a strong clinical interest because it can detect, among the neurological paEur Neurol 2015;74:288–295 DOI: 10.1159/000442412
293
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
Color version available online
Total
tients those who have effectively a deficit in social cognition. As mentioned, the high specificity allows predicting that impaired results at the GeSoCS in neurological patients are indeed related to the disease and less probably to non-neurological effects and are a sign of an acquired deficit in social cognition. In order to confirm the ecological validity of the GeSoCs, we made a post hoc correlation between the GeSoCS scores and scores obtained at the dysexecutive questionnaire (DEX) [31], a standardized 20 items rating scale used for quantifying behavioural disturbances commonly associated with executive impairment. In an unselected subgroup of 28 patients, a close relative completed the DEX questionnaire. In this subgroup, there was a correlation between total GeSoCS score and DEX score (CC = –0.450, p = 0.02). Another limitation of this study is the fact that various aetiologies were included, some of which contained smaller samples of patients, with the exception of the CVA group. Aetiologies therefore could not be investigated systematically and only the effect of the side of the lesion was taken into account because of the absence of prior known behavioural deficits. The aim was actually to test different neurological populations to show that different pathologies could be detected as causing ToM and social cognition deficits. However, in the second part of the results, we analysed separately dementia, stroke patients and the other aetiologies, in order to show that all pathologies may have social cognition impairment, but the pattern varies with the aetiologies. GeSoCS sub scores seemed also sensitive to the side of the lesions as described in our stroke subgroup: right CVAs failed in cartoons, inferences, ‘mind in the eyes’, and in the temporal rule
task while left CVAs were impaired in verbal/discourse tasks (social cognition, inferences, absurd stories, and cartoons). Furthermore, we did not attempt any qualitative analysis of the answers or the different questions for social cognition (ToM, point of view, faux pas) and did not establish correlations between the GeSoCS and neuropsychological (e.g. executive functions and language) or ecological measures. Finally, the test was intended to be as short as possible and thus, the test samples used may not have had the same diagnostic value as the test as a whole, e.g. the ‘reading the mind in the eyes’ test consists of 36 items, but only 10 were selected for our task. More elaborate studies on the different patient groups should be considered and, in particular, their correlation with other ecological measures, including questionnaires and behavioural scales, could be examined in more detail. Nevertheless, the data obtained from the current-screening test for disorders of social cognition and ToM provide a useful bedside assessment tool that should allow the clinician to determine the presence of ToM deficits in braindamaged patients. Acknowledgements We thank S. Baron Cohen, A. Duchêne May-Carle, M. Monfort and I. Monfort, Ortho Editions for allowing us to use parts of their tests to devise our scale and the neuropsychologists who administrated the GeSoCS.
Disclosure Statement The authors declare no conflict of interest.
References
294
7 8 9
10
11
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
rietal junction in ‘theory of mind’. Neuroimage 2003;19:1835–1842. Saxe R, Powell LJ: It’s the thought that counts: specific brain regions for one component of theory of mind. Psychol Sci 2006;17:692–699. Samson D: Reading other people’s mind: insights from neuropsychology. J Neuropsychol 2009;3:3–16. Martin-Rodriguez JF, Leon-Carrion J: Theory of mind deficits in patients with acquired brain injury: a quantitative review. Neuropsychologia 2010;48:1181–1191. Aboulafia-Brakha T, Christe B, Martory MD, Annoni JM: Theory of mind tasks and executive functions: a systematic review of group studies in neurology. J Neuropsychol 2011; 5: 39–55. Dunn J, Brophy M: Communication, relationships, and individual differences in chil-
dren’s understanding of mind; in Astington JW Baird JA (eds): Why Language Matters for Theory of Mind. Oxford, Oxford University Press, 2005, pp 50–69. 12 Baron-Cohen S, Jolliffe T, Mortimore C, Robertson M: Another advanced test of theory of mind: evidence from very high functioning adults with autism or Asperger syndrome. J Child Psychol Psychiatry 1997; 38: 813–822. 13 Baron-Cohen S, Wheelwright S, Hill J, Raste Y, Plumb I: The ‘reading the mind in the eyes’ test revised version: a study with normal adults, and adults with Asperger syndrome or high-functioning autism. J Child Psychol Psychiatry 2001;42:241–251. 14 Stone VE, Baron-Cohen S, Knight RT: Frontal lobe contributions to theory of mind. J Cogn Neurosci 1998;10:640–656.
Martory/Pegna/Sheybani/Métral/ Bernasconi Pertusio/Annoni
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
1 Ward J: The Student’s Guide to Social Neuroscience. Hove and New York, Psychology Press, 2012. 2 Premack D, Woodruff G: Does the chimpanzee have a theory of mind? Behav Brain Sci 1978;1:515–526. 3 Wimmer H, Perner J: Beliefs about beliefs: representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition 1983; 13: 103–128. 4 Baron-Cohen S, Leslie AM, Frith U: Does the autistic child have a ‘theory of mind’? Cognition 1985;21:37–46. 5 Frith U, Frith CD: Development and neurophysiology of mentalizing. Philos Trans R Soc Lond B Biol Sci 2003;358:459–473. 6 Saxe R, Kanwisher N: People thinking about thinking people. The role of the temporo-pa-
GeSoCS as a ToM Clinical Tool
21 Duchêne May Carle A: La gestion de l’implicite. Théorie et évaluation. Ortho Edition, 62330, Isbergues, 2007. 22 Binet A, Simon T: Le développement de l’intelligence chez les enfants. L’année Psychologique, 1907, vol 14, pp 1–94. 23 Luria AR, Prieram KH, Homskaya ED: An experimental analysis of the behavioral disturbance produced by a left frontal arachnoidal endothelioma (meningioma). Neuropsychologia 1964;2:257–280. 24 Messerli P, Seron X, Tissot R: Quelques aspects des troubles de la programmation dans le syndrome frontal. Arch Suisses Neurol Neurochir Psychiatrie 1979;125:23–35. 25 Muller F, Simion A, Reviriego E, Galera C, Mazaux JM, Barat M, Joseph PA: Exploring theory of mind after severe traumatic brain injury. Cortex 2010;46:1088–1099. 26 Apperly IA, Samson D, Carroll N, Hussain S, Humphreys G: Intact first- and second-order false belief reasoning in a patient with severely impaired grammar. Soc Neurosci 2006; 1: 334–348. 27 Bertrand-Gauvin C, Faucher L, Bocti C, Gagnon MJ, Joanette Y: Cognition sociale et accident vasculaire cérébral: perspective neuropsychologique. Rev Neuropsychol 2014; 6: 99–109.
28 Castelli I, Baglio F, Blasi V, Alberoni M, Falini A, Liverta-Sempio O, Nemni R, Marchetti A: Effects of aging on mindreading ability through the eyes: an fMRI study. Neuropsychologia 2010;48:2586–2594. 29 Spikman JM, Timmerman ME, Milders MV, Veenstra WS, van der Naalt J: Social cognition impairments in relation to general cognitive deficits, injury severity, and prefrontal lesions in traumatic brain injury patients. J Neurotrauma 2012;29:101–111. 30 Bertoux M, Volle E, Funkiewiez A, de Souza LC, Leclerq D, Dubois B: Social Cognition and Emotional Assessment (SEA) is a marker of medial and orbital frontal functions: a voxel-based morphometry study in behavioral variant of frontotemporal degeneration. J Int Neuropsychol Soc 2012; 18: 972– 985. 31 Wilson BA, Alderman N, Burgess PW, Emslie H, Evans J: Behavioural Assessment of the Dysexecutive Syndrome (BADS) Manual. Bury St Edmonds, Thames Valley Test Company, 1996.
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
295
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
15 Henry A, Tourbah A, Chaunu MP, Rumbach L, Montreuil M, Bakchine S: Social cognition impairments in relapsing-remitting multiple sclerosis. J Int Neuropsychol Soc 2011; 17: 1122–1131. 16 Henry JD, Cowan DG, Lee T, Sachdev PS: Recent trends in testing social cognition. Curr Opin Psychiatry 2015;28:133–140. 17 Desgranges B, Laisney M, Bon L, Duval C, Mondou A, Bejanin A, et al: TOM-15: une épreuve de fausses croyances pour évaluer la théorie de l’esprit cognitive. Rev Neuropsychol 2012;4:216–220. 18 Happé F, Brownell H, Winner E: Acquired ‘theory of mind’ impairments following stroke. Cognition 1999;70:211–240. 19 Monfort M, Monfort I: L’esprit des autres. Madrid, Entha Ediciones Pour Ortho Edition, 62330, Isbergues, 2001. 20 Duchêne May-Carle A: La gestion des inférences chez les cérébro-lésés droits. Thèse de doctorat en neuropsychologie. France, Université de Lyon, 1997.
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
Assessment of Social Cognition and Theory of Mind: Initial Validation of the Geneva Social Cognition Scale Marie-Dominique Martory a Alan John Pegna a, b Laurent Sheybani c Mélanie Métral a Françoise Bernasconi Pertusio a Jean-Marie Annoni d
a
Neuropsychology Unit and Laboratory of Experimental Neuropsychology, Department of Neurology, Geneva University Hospital, b Faculty of Psychology and Educational Science and c Department of Neuroscience, University Medical Center (CMU), University of Geneva, Geneva, and d Neurology Unit, Department of Medicine, Faculty of Sciences, University and Hospital of Fribourg, Fribourg, Switzerland
Abstract Background: Social cognition is widely studied in neurology. At present, such evaluations are designed for research or for specific diseases and simple general clinical tools are lacking. We propose a clinical evaluation tool for social cognition, the Geneva Social Cognition Scale (GeSoCS). Methods: The GeSoCS is a 100-point scale composed of 6 subtests (theory of mind stories, recognition of social emotions, false beliefs, inferences, absurdity judgement and planning abilities) chosen from different validated tests of social and cognitive evaluation. Eighty-four patients with neurological disorders and 52 controls participated in the study. Evaluation duration lasted 20–60 min. Results: Mean scores were 92.6 ± 4.5 for controls and 76.5 ± 15.3 for patients and differentiate patients and controls in all subtests. With a cut-off score of 84, the scale had a sensitivity of 62% and a specificity of 94%. In our stroke subgroup, right CVAs failed in cartoons, inferences, ‘mind in the eyes’, and in the temporal rule task while
© 2015 S. Karger AG, Basel 0014–3022/15/0746–0288$39.50/0 E-Mail [email protected] www.karger.com/ene
left CVAs were impaired in verbal/discourse tasks (social cognition, inferences, absurd stories, and cartoons. Conclusions: The GeSoCS is a medium duration assessment tool that appears to detect and characterize significant social impairment in neurological patients. © 2015 S. Karger AG, Basel
Introduction
The last decade saw the development of social neuroscience, which integrated social psychology, cognitive neuropsychology and neurosciences and raised the question of whether social cognition exists as a specific domain (for review see Ward [1]). Social cognition refers to the abilities and processes allowing people to interact and to understand the behaviour of others. Among these different processes, the capacity to attribute mental states such as feelings and beliefs and then to predict behaviour is called theory of mind (ToM) and plays a determining role in social adaptation. ToM was first studied in primates [2], and subsequently in developmental psychology [3]. Its first clinical application involved autistic persons [4]. Social cognition has also been addressed extenDr. Marie-Dominique Martory Neuropsychology Unit, Department of Neurology Geneva University Hospital 4, Gabrielle-Perret-Gentil, 1205 Geneva (Switzerland) E-Mail mdmartory @ bluewin.ch
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
Key Words Clinical scale · Neuropsychological evaluation · Brain damage · Social emotions · False beliefs · Inferences · Absurdity judgement · Executive planning
GeSoCS as a ToM Clinical Tool
Table 1. Demographic characteristics of control and patient groups
Age, mean (SD) Gender Male Female Handedness Right Left Level of education, years 12
Control group (n = 52)
Patient group (n = 84)
42.2 (16.4)
53.3 (15.5)
20 32
49 35
51 1
71 13
1 51
11 73
Table 2. Patient group aetiologies Patient group (n = 84) Age, mean (SD) Gender Male Female Handedness Right Left
Stroke Trauma Tumor Dementia (n = 31) (n = 17) (n = 7) (n = 17)
Encephalopathy (n = 11)
58 (12.8) 41 (13)
44.8 (16.6)
51 (14) 66.4 (8.8)
15 16
13 4
4 3
11 6
7 4
28 3
14 3
6 1
14 3
8 3
neurological patients rather than a short screening tool. Deficits identified with this tool would then allow subsequent detailed examination using the existing tests.
Method Participants All subjects gave written informed consent for their participation in accordance with the Declaration of Helsinki. The study was approved by the local Ethics Committee, (Commission d’Ethique de la Recherche Clinique, Geneva University Hospital, Switzerland; CER 10-073). The population was composed of 52 healthy French-speaking, volunteers, with no known neurological or psychiatric disorders. Demographic data are summarized in table 1. Eighty-four brain-damaged French-speaking patients were examined consecutively during their hospitalization in the neurological clinic. The patient population presented various aetiologies (stroke, epilepsy, trauma, tumour, dementia; table 2). Participants with CVA were assessed in the post-acute stage (about 10 days to
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
289
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
sively using brain imaging techniques. Along these lines, fMRI studies, for example, have shown that the temporal lobe, the temporal-parietal-junction, the medial prefrontal cortex and the posterior cingulate cortex are specifically involved in ToM [5–7]. In the field of clinical neuroscience, investigations of brain-injured patients with acute or neurodegenerative disease have also revealed different types of impairments in social cognition and the understanding of people’s behaviour [8]. Indeed, ToM constitutes a complex domain that involves numerous cognitive processes, including executive functions [9, 10] and, in certain conditions, normal language development [11]. Thus, the evaluation of social cognition and ToM in adults relies on a wide range of tasks of varying complexity. The most common involves false beliefs and can be used with young children. In particular, first-order beliefs necessitate understanding another person’s mental state, while second-order beliefs, which are more complex, predict what a person thinks about a third person’s feelings [3, 4]. More complex tasks used to assess the mental state of others require that participants identify a person’s emotional state on the basis of the expression in their eyes (the ‘mind in the eyes’ test) [12, 13], detect a social ‘faux pas’ [14], or understand mental inferences or indirect speech as in metaphors and jokes. Children are able to resolve these tasks at different ages (age 4 for the first order false belief and 9 or 11 for the faux pas). In brain-damaged patients, a number of authors have proposed that patients with brain injuries would process simple false beliefs stories correctly but would fail in more complex tasks [14]. Despite the great number of studies in social neuroscience investigating patients with acute acquired or neurodegenerative brain damage, research in the field has not yet provided tests that are broad enough for use in the clinic and have focused either on specific pathologies, such as multiple sclerosis [15] or neurodegenerative groups (for review see Henry et al. [16]), or alternately, have addressed specifics such as false beliefs [17], or faux pas recognition [14]. Additional global testing, as described earlier, is often too long for clinical practice. To our knowledge, there is yet no general test that allows neuropsychologists to assess ToM and social cognition function more globally in everyday clinical practice. The aim of this study is thus to offer a short assessment tool that is easy to use in everyday neuropsychological practice, which has been developed from existing tests described in the literature such as the mind in the eyes test or false beliefs. This scale was intended as a clinical tool for potential impairment in social cognition found in
Assessment of ToM and Social Cognition The test was developed internally on the basis of the literature reported above and spanned aspects most frequently explored in social cognition. Each of them was evaluated through specific tasks as described in the literature and combined in order to form a scale with a total score of 100 points (online suppl. appendix, see www. karger.com/doi/10.1159/000442412). Tasks 1, 2, 3 and 4 were specific to social cognition and comprehension of inferences. For each of them, the patients were asked to read a story and to answer specific questions involving ToM, but also to respond to simple control questions in order to test linguistic comprehension. The lengths of the stories were similar. In addition, precautions were taken to ensure that the questions and the syntactic structures of the sentences were equivalent and simple enough to be understood by most patients. Task 1 was composed of 5 written ToM and social cognition stories: one ‘faux pas’ situation, one ‘normal’ situation (‘Helen’ and ‘Jim’ stories [14]), one ToM story (‘burglar’ story [18]), and one in-house ToM story involving a self-perspective1. Task 2 involved 10 items chosen from the ‘reading the mind in the eyes’ test [12, 13]. The original test is composed of 36 pairs of eyes each evoking an emotion. The subject has to choose between 4 written propositions the one that best describes the emotion. The items selected (1-4-5-8-14-16-18-30-33-35) were retained as being the most easily identifiable on the basis of a pilot experiment with 25 participants. Task 3 was made up of 5 ToM picture stories [19] representing first- and second-order false beliefs (e.g. A man is sitting on a bench with his dog beside him and he is preparing to eat a sandwich; while he is throwing the paper packaging, a bird steals the sandwich; on seeing the sandwich missing, the man scolds the dog. Question: why does the man scold his dog?) and one story not requiring inference (A boy sits on his sofa when he hears the doorbell ring. He thinks that it is the mailman but he smiles when he sees that it is his friend. Question: Why does the boy seem happy?). Task 4 was composed of 5 stories that the patient had to read before answering questions about possible or impossible inferences [20, 21] e.g. ‘The more Willy loses at the casino, the more affectionate he is with his wife; apparently, he was not very lucky in his game this evening’. Control question without inference: Did Willy win a lot at the casino this evening? (response: no). Question with impossible inference to make: Is Willy’s wife rich? (response: we cannot know). Question with an inference to make:
1 In a fMRI study with healthy subjects, Vogeley et al. [2001], found different patterns of brain activation, as the subjects took the first or third person perspective and Gambini, Barbieri, Scarone [2004], supported this idea in a study involving schizophrenic subjects.
290
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
Is Willy rather affectionate with his wife this evening? (response: yes). Tasks 5 and 6 are beyond ToM strictly speaking but are used to determine clinically a general executive impairment, which can be associated (although not necessarily) with deficits in social cognition [9]. Task 5 required that patients correctly detect an absurdity in 4 stories and accept a control story (e.g. ‘The management of a railroad company noticed that in train accidents, injuries were for more serious in the tail wagons. Consequently, it was decided to remove the last wagon of each train. What do you think of that?’). This test was retained as it constituted an interesting match with the social cognition stories that also required a ‘meta’ representation, albeit without the social component (adapted from Binet and Simon [22]). Task 6 was not directly related to social cognition, but was presented as a general measure of executive function. It is a French adaptation of the Counter test [23] and consists in determining the rules underlying the temporal appearance of colour counters (e.g. black-white-white, black-white-white…). This test was chosen for its ease of use and its good prediction of mental flexibility in frontal impairment [24]. Every verbal or picture task contained control questions measuring literal comprehension but excluding ToM or social aspect. The control questions were not scored but were included to ensure correct linguistic comprehension. The item ‘Jim’ that constituted a full question in the test was an exception. The duration of the whole test ranged from about 40 min in a healthy subject to between 60 and 70 min in the brain-damaged patients. In this latter group, participants could re-read the stories or obtain repetitions of the questions. Patients and control subjects were examined by clinically qualified neuropsychologists. Procedure The Geneva Social Cognition Scale (GeSoCS) was tested in addition to the patients’ standard neuropsychological examination, which will not be detailed in this study. In controls, no additional evaluation was performed, as participants were considered to be cognitively normal. However, a semi-structured interview was carried out to confirm the absence of any indicators of cognitive impairment, either anamnestically or behaviourally. The subjects were individually tested in a single session of about 40–60 min. The sessions took place at the University Hospital of Geneva, in the Neuropsychology Unit. During the test, they were asked to answer the questions freely and no solutions were proposed. If necessary, the question was repeated. For each item, full points were given when the answer was complete and correct. Half points were given if the answer was incomplete, or if the answer was correct but not the explanation. Statistical Analysis The results of the whole group are described first, followed by that of the patients and controls (global scores, then subtests). Significant differences and interactions between the controls and the patients were calculated using an analysis of variance (ANOVA; statistical analyses were performed using Statistica 12 software). For the smaller and more heterogeneous groups (TBI, dementia), additional testing was carried out using nonparametrical tests.
Martory/Pegna/Sheybani/Métral/ Bernasconi Pertusio/Annoni
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
3 weeks after their stroke and after stabilization for other aetiologies). All patients’ diagnoses were based on a clinical evaluation, a laboratory evaluation including an MRI, and a formal neuropsychological assessment. Exclusion criteria included a history of prior psychiatric disorders, delirium, global aphasia and severe dementia (≥CDR3). Subsequent statistical inquiries as to the effects of levels of education on our tests showed that this factor did not influence the results. The different levels were therefore merged.
Social cognition
Mind in the eyes
Cartoons
Inferences
Absurd stories
Temporal rules
Total
0.80
0.40
0
–0.40 Controls
Fig. 1. Comparison of mean z-scores of
–0.80
Patients
controls and patients for all the tasks.
Results
Table 3. Results: mean scores and SD for controls and patients for
all tasks
Comparison of z-Scores: CVA Patients vs. Controls The 31 patients with CVA demonstrated approximately the same pattern as the entire patient group with altered performances compared to the controls as shown in figure 2. GeSoCS as a ToM Clinical Tool
Tasks
Controls (n = 52), mean (SD)
Patients (n = 84), mean (SD)
Social cognition Mind in the eyes ToM cartoons Inferences Absurd stories Temporal rules Total
19.36 (1.20) 15.09 (2.28) 19.55 (0.79) 19.17 (1.49) 9.65 (0.94) 9.76 (0.85) 92.61 (4.51)
16.02 (4.08) 12.11 (3.81) 17.22 (3.5) 15.83 (4.4) 7.78 (2.32) 7.54 (3.42) 76.51 (15.32)
The ANOVA confirmed that the difference was significant (F(15, 395) = 4.4,651, p < 0.05). However, when considering the side of the lesion, a post hoc Fisher test showed different patterns of impairment in the case of right CVA, left CVA or bilateral CVA. Right CVAs did not differ from the controls in the first task of social cognition and were quite correct in ‘faux pas’ stories, which required criticism (p = 0.06), but they failed in cartoons, in inferences and in tasks in which they had to guess a feeling, such as the ‘mind in the eyes’ test (p < 0.05), or an intention, such as in the temporal rule task (p < 0.05). Left CVAs were impaired in all verbal tasks, social cognition tasks (p < 0.05), inferences (p < 0.05), absurd stories, and also in cartoons (p < 0.05). Conversely, they have Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
291
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
Descriptive statistics for the controls and the patients are provided in table 3. The controls obtained a total mean score of 92.6/100 (SD 4.5); for each test, the result was close to the maximum score with a standard deviation equal to 1, with the exception of the ‘mind in the eyes’ test, (mean 15/20, SD 2.3). There was no significant difference due to gender (women 92.4, men 93). The patients obtained a total mean score lower than the controls (76.51) with a large dispersal (SD 15.32) and performances for each test were lower than the controls. For every task, the performance of the patients and the controls were expressed as z-scores of the mean total group’s performances (patients + controls; fig. 1). An ANOVA comparing the 2 groups’ z-scores revealed a significant difference in performance between the controls and the patients (F(1, 134) = 55.035, p < 0.00001), and a post hoc test showed that the patients were impaired compared to the healthy subjects in all tasks. We now considered the mean z-score for the CVA population compared to the mean z-score of the control group.
1.20
Controls
Right CVA
Left CVA
Bilateral CVA
0.80
0.40
0
–0.40 Social cognition –0.80
Mind in the eyes Cartoons Inferences
–1.20
Temporal rules
Fig. 2. Comparison of mean z-scores of –1.60
shown significant difficulties in the ‘mind in the eyes’ task (p = 0.07). Finally, bilateral CVAs were significantly impaired (p < 0.05) in all 6 tasks. Determining Pathological Function: The Cut-Off Score Based on these data, we attempted to determine a cutoff score for clinical use. The values for each test are given in table 4 using a cut-off score of minus 2 SDs. Using these values, 52 patients fell below the cut-off (84) and 32 remained above; among the 31 CVA patients, 19 were below the cut off (9/15 left, 5/8 right and 5/7 bilateral CVA); 3 healthy subjects fell below the cut-off and 49 were above. The cut-off of 84 allowed to significantly differentiate between controls and patients (χ2(1) = 42.02; p < 0.0001). Sensitivity was at 61.9% and specificity at 94.23%. An analysis computing the receiver operating curve (ROC) with these cut-off scores showed an estimated area of 0.78 under the curve. Thus, the probability of correctly classifying a neurological patient as pathological on the basis of our test is 78% (fig. 3).
Discussion
In order to provide a simple tool for evaluating social cognition, we designed the GeSoCS, which allows rapid and easy assessment of social cognition in hospitalized neu292
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
Table 4. Cut-off for each subtest and the total test
Tasks
Controls, mean (SD)
Cut-off (rounded to nearest unit)
Social cognition Mind in the eyes ToM cartoons Inferences Absurd stories Temporal rules Total
19.36 (1.20) 15.09 (2.28) 19.55 (0.79) 19.17 (1.49) 9.65 (0.94) 9.76 (0.85) 92.61 (4.51)
17 11 18 16 8 8 84
rological patients in a daily clinical practice. This screening test was constructed with items taken from existing tasks previously described in the literature for evaluating the ToM. These tasks were subsequently adapted and grouped in order to form a scale graded from 0 to 100. To the best of our knowledge, it is the first global tool adapted for everyday clinical work as most of the studies on clinical patients until now have used extensive batteries. From a clinical perspective, GeSoCS appears to conform to the criteria for a consistent scale. First, the GeSoCS uses multiple aspects of validated tests for assessment of ToM, a ‘complex, multifaceted cognitive ability’ [25]. On average, GeSoCS lasts 40–60 min. Generally, this duration is easily tolerated by neurological patients, although one cannot exclude potential attentional deficits during evaluation. Martory/Pegna/Sheybani/Métral/ Bernasconi Pertusio/Annoni
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
right CVA, left CVA, bilateral CVA and controls for all the tasks.
Absurd stories
100
Sensitivity
80
60
40
20
0 0
20
40
60
80
100
100 – Specificity
Fig. 3. ROC with an estimated area of 0.78 under the curve.
This scale was administered by experienced neuropsychologists to 136 volunteers: 52 healthy subjects and 84 patients who were either hospitalized in the post-acute stage or stabilized. The results showed that patients performed significantly more poorly than healthy controls on average (p < 0.05). In our sample, patients differed from controls not only in the global score but also in all subtests. This ability to differentiate between both groups, despite the patients’ heterogeneous results (SD 15.368 in the patients’ group vs. 4.5 in the controls), offers a satisfying perspective in the use of this tool. The CVA patients, who constituted the main group, performed in a significantly worse manner in all events. Interestingly, the results differed according to the lateralization of the lesion: left CVAs were significantly different from the controls in all the tests (p < 0.05) except the ‘mind in the eyes’ test. Some of these left CVA patients suffered from aphasia and, although it has been shown that ToM abilities are not fundamentally impaired in this population [26], it is possible that language difficulties and deficient working memory may play a part in these results, and also in ToM picture tasks where all the answers require a verbal argumentation. Nevertheless, this group was not significantly different from the controls in the interpretation of the ‘mind in the eyes’ test (p > 0.5). The right CVAs showed a different pattern with respect to the controls and to the left CVAs: they failed in the ‘mind in the eyes’ test, in temporal rule tasks (which nevertheless do not require an elabGeSoCS as a ToM Clinical Tool
orated verbal response) and ToM pictures (cartoons) but performed correctly in the cognition scale. The bilateral strokes, as expected, had more difficulty in all tasks. As regards CVA, the recent data in the literature show that social cognition is affected during the acute phase, particularly if the lesion involves anatomical regions such as the parieto-temporal junction and fronto-parietal areas. However, this disorder can evolve favourably in the same way as other cognitive functions [27]. Interest and Limitations of Current Study The GeSoCS scale has the advantage of being composed of tests that are described and validated in the literature as being sensitive to social cognition and ToM. Instructions and scoring methods are simple to use for neuropsychologists and behavioural neurologists. Its duration and the absence of heavy equipment allow it to be used quite easily with any patient suspected of presenting disorders of social cognition. In our study, the number of healthy controls was slightly limited, but their scores were high and homogeneous, and most tests have already been reported extensively in the literature and their use has been confirmed. Their average age was younger (11 years) than the average age of the patients. This difference might have influenced our findings. However, a recent study that considers the effect of aging on mind-reading abilities in healthy people has suggested the absence of any difference between younger and older subjects [28]. The test cannot be understood as a screening, but must be considered as a more descriptive test. It took often between 30 min and 1 h, and had to be done in a different session than the usual neuropsychological evaluation. Such durations have also been used in other populations with clinically oriented test (see social cognition impairments in relation to general cognitive deficits, injury severity, and prefrontal lesions in traumatic brain injury patients [29] or in frontotemporal dementia [30]). So it must be considered a clinical test with a capacity of describing the type of deficit in clinical cognition in neurological patients. Our experience suggests that patients easily support this non-challenging evaluation, except strongly demented patients and severe aphasia. Our data showed a good specificity of 94% for a cut-off of 84, making the GeSoCS a highly efficient method for detecting deficits of social cognition. The sensitivity of 61% is of course a limitation in our study and the GeSoCS cannot be considered an interesting screening tool for neurological disorders. It has, however, a strong clinical interest because it can detect, among the neurological paEur Neurol 2015;74:288–295 DOI: 10.1159/000442412
293
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
Color version available online
Total
tients those who have effectively a deficit in social cognition. As mentioned, the high specificity allows predicting that impaired results at the GeSoCS in neurological patients are indeed related to the disease and less probably to non-neurological effects and are a sign of an acquired deficit in social cognition. In order to confirm the ecological validity of the GeSoCs, we made a post hoc correlation between the GeSoCS scores and scores obtained at the dysexecutive questionnaire (DEX) [31], a standardized 20 items rating scale used for quantifying behavioural disturbances commonly associated with executive impairment. In an unselected subgroup of 28 patients, a close relative completed the DEX questionnaire. In this subgroup, there was a correlation between total GeSoCS score and DEX score (CC = –0.450, p = 0.02). Another limitation of this study is the fact that various aetiologies were included, some of which contained smaller samples of patients, with the exception of the CVA group. Aetiologies therefore could not be investigated systematically and only the effect of the side of the lesion was taken into account because of the absence of prior known behavioural deficits. The aim was actually to test different neurological populations to show that different pathologies could be detected as causing ToM and social cognition deficits. However, in the second part of the results, we analysed separately dementia, stroke patients and the other aetiologies, in order to show that all pathologies may have social cognition impairment, but the pattern varies with the aetiologies. GeSoCS sub scores seemed also sensitive to the side of the lesions as described in our stroke subgroup: right CVAs failed in cartoons, inferences, ‘mind in the eyes’, and in the temporal rule
task while left CVAs were impaired in verbal/discourse tasks (social cognition, inferences, absurd stories, and cartoons). Furthermore, we did not attempt any qualitative analysis of the answers or the different questions for social cognition (ToM, point of view, faux pas) and did not establish correlations between the GeSoCS and neuropsychological (e.g. executive functions and language) or ecological measures. Finally, the test was intended to be as short as possible and thus, the test samples used may not have had the same diagnostic value as the test as a whole, e.g. the ‘reading the mind in the eyes’ test consists of 36 items, but only 10 were selected for our task. More elaborate studies on the different patient groups should be considered and, in particular, their correlation with other ecological measures, including questionnaires and behavioural scales, could be examined in more detail. Nevertheless, the data obtained from the current-screening test for disorders of social cognition and ToM provide a useful bedside assessment tool that should allow the clinician to determine the presence of ToM deficits in braindamaged patients. Acknowledgements We thank S. Baron Cohen, A. Duchêne May-Carle, M. Monfort and I. Monfort, Ortho Editions for allowing us to use parts of their tests to devise our scale and the neuropsychologists who administrated the GeSoCS.
Disclosure Statement The authors declare no conflict of interest.
References
294
7 8 9
10
11
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
rietal junction in ‘theory of mind’. Neuroimage 2003;19:1835–1842. Saxe R, Powell LJ: It’s the thought that counts: specific brain regions for one component of theory of mind. Psychol Sci 2006;17:692–699. Samson D: Reading other people’s mind: insights from neuropsychology. J Neuropsychol 2009;3:3–16. Martin-Rodriguez JF, Leon-Carrion J: Theory of mind deficits in patients with acquired brain injury: a quantitative review. Neuropsychologia 2010;48:1181–1191. Aboulafia-Brakha T, Christe B, Martory MD, Annoni JM: Theory of mind tasks and executive functions: a systematic review of group studies in neurology. J Neuropsychol 2011; 5: 39–55. Dunn J, Brophy M: Communication, relationships, and individual differences in chil-
dren’s understanding of mind; in Astington JW Baird JA (eds): Why Language Matters for Theory of Mind. Oxford, Oxford University Press, 2005, pp 50–69. 12 Baron-Cohen S, Jolliffe T, Mortimore C, Robertson M: Another advanced test of theory of mind: evidence from very high functioning adults with autism or Asperger syndrome. J Child Psychol Psychiatry 1997; 38: 813–822. 13 Baron-Cohen S, Wheelwright S, Hill J, Raste Y, Plumb I: The ‘reading the mind in the eyes’ test revised version: a study with normal adults, and adults with Asperger syndrome or high-functioning autism. J Child Psychol Psychiatry 2001;42:241–251. 14 Stone VE, Baron-Cohen S, Knight RT: Frontal lobe contributions to theory of mind. J Cogn Neurosci 1998;10:640–656.
Martory/Pegna/Sheybani/Métral/ Bernasconi Pertusio/Annoni
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
1 Ward J: The Student’s Guide to Social Neuroscience. Hove and New York, Psychology Press, 2012. 2 Premack D, Woodruff G: Does the chimpanzee have a theory of mind? Behav Brain Sci 1978;1:515–526. 3 Wimmer H, Perner J: Beliefs about beliefs: representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition 1983; 13: 103–128. 4 Baron-Cohen S, Leslie AM, Frith U: Does the autistic child have a ‘theory of mind’? Cognition 1985;21:37–46. 5 Frith U, Frith CD: Development and neurophysiology of mentalizing. Philos Trans R Soc Lond B Biol Sci 2003;358:459–473. 6 Saxe R, Kanwisher N: People thinking about thinking people. The role of the temporo-pa-
GeSoCS as a ToM Clinical Tool
21 Duchêne May Carle A: La gestion de l’implicite. Théorie et évaluation. Ortho Edition, 62330, Isbergues, 2007. 22 Binet A, Simon T: Le développement de l’intelligence chez les enfants. L’année Psychologique, 1907, vol 14, pp 1–94. 23 Luria AR, Prieram KH, Homskaya ED: An experimental analysis of the behavioral disturbance produced by a left frontal arachnoidal endothelioma (meningioma). Neuropsychologia 1964;2:257–280. 24 Messerli P, Seron X, Tissot R: Quelques aspects des troubles de la programmation dans le syndrome frontal. Arch Suisses Neurol Neurochir Psychiatrie 1979;125:23–35. 25 Muller F, Simion A, Reviriego E, Galera C, Mazaux JM, Barat M, Joseph PA: Exploring theory of mind after severe traumatic brain injury. Cortex 2010;46:1088–1099. 26 Apperly IA, Samson D, Carroll N, Hussain S, Humphreys G: Intact first- and second-order false belief reasoning in a patient with severely impaired grammar. Soc Neurosci 2006; 1: 334–348. 27 Bertrand-Gauvin C, Faucher L, Bocti C, Gagnon MJ, Joanette Y: Cognition sociale et accident vasculaire cérébral: perspective neuropsychologique. Rev Neuropsychol 2014; 6: 99–109.
28 Castelli I, Baglio F, Blasi V, Alberoni M, Falini A, Liverta-Sempio O, Nemni R, Marchetti A: Effects of aging on mindreading ability through the eyes: an fMRI study. Neuropsychologia 2010;48:2586–2594. 29 Spikman JM, Timmerman ME, Milders MV, Veenstra WS, van der Naalt J: Social cognition impairments in relation to general cognitive deficits, injury severity, and prefrontal lesions in traumatic brain injury patients. J Neurotrauma 2012;29:101–111. 30 Bertoux M, Volle E, Funkiewiez A, de Souza LC, Leclerq D, Dubois B: Social Cognition and Emotional Assessment (SEA) is a marker of medial and orbital frontal functions: a voxel-based morphometry study in behavioral variant of frontotemporal degeneration. J Int Neuropsychol Soc 2012; 18: 972– 985. 31 Wilson BA, Alderman N, Burgess PW, Emslie H, Evans J: Behavioural Assessment of the Dysexecutive Syndrome (BADS) Manual. Bury St Edmonds, Thames Valley Test Company, 1996.
Eur Neurol 2015;74:288–295 DOI: 10.1159/000442412
295
Downloaded by: Emory University 198.143.47.65 - 1/26/2016 7:00:06 PM
15 Henry A, Tourbah A, Chaunu MP, Rumbach L, Montreuil M, Bakchine S: Social cognition impairments in relapsing-remitting multiple sclerosis. J Int Neuropsychol Soc 2011; 17: 1122–1131. 16 Henry JD, Cowan DG, Lee T, Sachdev PS: Recent trends in testing social cognition. Curr Opin Psychiatry 2015;28:133–140. 17 Desgranges B, Laisney M, Bon L, Duval C, Mondou A, Bejanin A, et al: TOM-15: une épreuve de fausses croyances pour évaluer la théorie de l’esprit cognitive. Rev Neuropsychol 2012;4:216–220. 18 Happé F, Brownell H, Winner E: Acquired ‘theory of mind’ impairments following stroke. Cognition 1999;70:211–240. 19 Monfort M, Monfort I: L’esprit des autres. Madrid, Entha Ediciones Pour Ortho Edition, 62330, Isbergues, 2001. 20 Duchêne May-Carle A: La gestion des inférences chez les cérébro-lésés droits. Thèse de doctorat en neuropsychologie. France, Université de Lyon, 1997.
