Savants, family, and Social Responsiveness

Post Reply New Topic

Abstract
Savant syndrome is a phenomenon whereby individuals with cognitive impairments have one or more outstanding abilities, inconsistent with their general intellectual functioning. Approximately 50% of savant individuals have autism spectrum disorder (ASD), and 10–30% of people with ASD have savant skills. To shed additional light on this considerable overlap, we compared autistic traits as measured by the Social-Responsiveness-Scale (SRS) between 712 children with at least one reported savant skill, as determined by designated questions from the ADI-R questionnaire (savant group), and 2,032 non-savant children from the Simons-Simplex-Collection (SSC) database. We also examined SRS scores of the parents of these children and compared parent-child differences in SRS scores between the savant and non-savant groups. Savant children had significantly lower SRS scores (less deficiencies) compared to non-savant children (P < 0.05), while no such differences were observed among their parents. Further intra-familial analyses revealed weak pairwise-correlations (r = −0.015–0.141) between SRS scores of parents and their children, and significantly larger parent-child differences in standardized SRS scores within savant families (P < 0.05). These findings suggest that the less severe autistic traits among savant children with ASD compared to other people with ASD is not likely to be a familial trait.

Introduction
Savant syndrome is a fascinating phenomenon whereby individuals with cognitive impairments have one or more outstanding abilities, inconsistent with their general intellectual functioning1. The special skills typically occur in five areas - music, art, calendar calculating, mathematics or mechanical/visual-spatial skills, and are usually associated with outstanding memory that is restricted to the area of expertise2,3,4,5. In most cases, the condition is evident from early childhood5,6,7, however few cases of acquired savant skills have been documented following a CNS injury or disease8,9.

Interestingly, there is a high concordance between savant syndrome and autism spectrum disorder (ASD) where approximately 50% of savant individuals have also ASD5,10,11, and 10–30% of persons with ASD have exceptional abilities4,12,13. Another similarity between these two conditions is that both have a similar male-to-female ratio of 4-6:15,14. In the general population, a relation between special abilities and ASD-like-traits has been observed, where children with special abilities show elevated ASD-like traits compared to those without such abilities15,16,17,18. In contrast, within ASD populations, Individuals with special abilities tend to have less ASD-like traits, or – elevated social skills13,19.

The considerable concordance between ASD and savant syndrome could stem from a shared etiology of these two conditions, and/or from a mechanism by which one condition predisposes the emergence of the other. Theories supporting the first option include the “Enhanced Perceptual Functioning” model that suggests that the over-functioning of brain regions involved in perceptual functions among people with ASD may be also responsible for the high prevalence of special abilities in these people20,21. Veridical mapping has been further proposed as the specific mechanism that accounts for the role of perception in the manifestation of exceptional skills among people with ASD22,23,24. A few rare genetic syndromes support these theories where people with these syndromes have both autistic features and specific enhanced skills25,26,27. Alternatively, the “Empathizing–Systemizing” theory, suggests that the attention to detail observed in ASD individuals occurs in the service of achieving an ultimate understanding of a system, and may predispose the emergence of savant skills28. Similarly, in their “Fractionable Triad Proposal”, Happé and Vital showed that the greater number of ASD-like traits found in savant-autistic children was particularly due to increased restricted and repetitive behaviors and interests (RRBIs), while their social and communication impairments were somewhat reduced17. Therefore, they suggested that the autistic characteristic of a detail focused cognitive style (weak coherence) predisposes to talent, and that the presence of savant skills may aid social competence. Other studies found no indication that elevated RRBIs predict the existence of savant skills, although they did find that autistic savants have better cognition and social skills than people with ASD without any exceptional talent4,13,29.

Today, the high concordance between the savant syndrome and ASD is still an enigma. The main goal of this study was to examine whether children with ASD and exceptional skills (i.e. autistic savants) have different autistic characteristics, than other children with ASD, and whether such differences are familial or not. To address this question, we compared various autistic traits as measured by the Social Responsiveness Scale (SRS) questionnaire30 between children with ASD and their parents. These analyses allowed us to assess whether the elevated social skills observed among individuals with ASD and savant skills may have a familial origin.

Discussion
The main goal of this study was to examine whether children with ASD and exceptional skills (i.e. autistic savants) have different autistic characteristics, as measured by the SRS, than other children with ASD, and whether such differences are familial or not. We show that parents of children with and without savant skills do not have different social responsiveness characteristics. Yet, in line with previous studies13,19, children with autism and savant skills have significantly lower SRS scores (less deficiencies) in a range of social impairment domains, as well as having higher IQ, and lower scores in the ADOS and ADI-R tests than non-savant autistic children. The observed differences in age and IQ between these two groups could contribute to the reported differences is social skills. Alternatively, the less apparent autistic features among the savant children with ASD could lead to a delayed diagnosis in these children. Of note, these differences tend to be larger among children with specific skill types (Fig. 1, white columns), suggesting that individuals with different savant skills may have distinct patterns of autistic traits.

Our findings of no significant difference in the autistic mannerisms domain of the SRS between the savant and non-savant groups join other findings (e.g.4,29) that partially negate the “Fractionable Triad Proposal” which suggests that the autistic characteristic of a detail focused cognitive style (weak coherence, RRBIs) predisposes to talent17,21,35,36. In contrast, the observation that the savant group is characterized by better social awareness in our study is compatible with the suggestion that the presence of special skills may aid social competence17,37. Importantly, similar results were seen when we compared each of the five savant subgroups (defined by their talents) to the non-savant group. This may also support the premise that the presence of exceptional talent among people with ASD enhances their social skills, regardless of the type of talent.

Interestingly, there were significant gender differences in all domains of the adult version of the SRS questionnaire, however no such differences were seen in the child questionnaires. A possible explanation of such observation is a rater bias, due to the different gender ratio of the raters of these questionnaires; Child questionnaires are more commonly filled by mothers than by fathers38 whereas adult questionnaires are always filled by one parent, on the second parent. Thus, the observed differences in SRS scores between the mothers and fathers in this study most likely reflect differences in the rater’s judgment in the different rater groups. Yet, the possibility that these differences reflect true differences in social behavior is still valid. Unfortunately, the lack of information regarding the gender of the child SRS rater in our dataset disallowed us from testing this hypothesis.

Finally, we observed larger intra-familial differences in standardized SRS scores among the savant families than in the non-savant families. This finding implies that the differences in autistic traits between the savant and non-savant children with ASD are unlikely driven by familial factors. We therefore suggest that the relationship between talent and social behavior observed in this study and others13,17 is more likely a result of one trait nurturing the other. Nevertheless, our data cannot indicate whether the existence of exceptional talent improves social competence, or whether the predisposition for better social skills promotes talent. In addition, a mutual synergism between these two traits whereby the existence of one enhances the other and vice-versa is also a valid possibility. Further empirical studies of talented children with ASD are thus warranted to gain better understanding of the mechanisms underlying talent in these children. Results of such studies will have remarkable social and clinical implications to people with ASD as well as to other human beings.

The strengths of this study include its large sample size, especially the relatively large number of female subjects which is a recurring limitation among studies of autistic populations. The large sample size allowed us to examine gender-based differences, which was not possible in previous, smaller samples. Nevertheless, our findings should be interpreted in the context of several limitations. First, the savant definition was based on parent report, and was not validated by empirical tests. Using the questions from the ADI-R to determine savant skills in children with ASD may result in either underestimation or overestimation of savant skills in the study. Multiple other large-scale studies (e.g.4,12,13,29,33 used the parental reports to determine exceptional talents among children with ASD. Some of these studies, which used both parental reports and empirical tests of some of the reported skills4,39, suggested that parental reports provide a good proxy for their child’s exceptional talent. Nevertheless, this approach is certainly imperfect, and may cause inclusion of empirically non-talented children in the savant group. Second, despite the relatively large sample size of this study, the classification of the savant group into talent-specific subgroups resulted in small sample sizes, which together with the need of multiple testing correction reduced the statistical power to draw robust conclusions about differences between these groups. Third, the Simons simplex sample is not a genuine representation of the autistic population in the US or elsewhere. Thus, caution should be taken in generalization of the conclusions of this study. Fourth, this is a cross-sectional study whereby single measurements of both the SRS and savant skills were taken at the same time. Therefore, we could not determine which of these traits predispose the other one. Finally, we did not have data regarding the occurrence of exceptional talents among the parents of the children in our dataset and hence could not assess whether the talents reported in the children are inherited or acquired.

Conclusions
Our results suggest that the better social responsiveness observed among people with ASD and savant skills compared to other people with ASD is not likely to be a familial trait. Further studies are required to examine whether any of these distinct autistic features predisposes the other, or whether they co-develop simultaneously.