Social behavioural adaptation in Autism

Abstract

Autism is still diagnosed on the basis of subjective assessments of elusive notions such as interpersonal contact and social reciprocity. We propose to decompose reciprocal social interactions in their basic computational constituents. Specifically, we test the assumption that autistic individuals disregard information regarding the stakes of social interactions when adapting to others. We compared 24 adult autistic participants to 24 neurotypical (NT) participants engaging in a repeated dyadic competitive game against artificial agents with calibrated reciprocal adaptation capabilities. Critically, participants were framed to believe either that they were competing against somebody else or that they were playing a gambling game. Only the NT participants did alter their adaptation strategy when they held information regarding others' competitive incentives, in which case they outperformed the AS group. Computational analyses of trial-by-trial choice sequences show that the behavioural repertoire of autistic people exhibits subnormal flexibility and mentalizing sophistication, especially when information regarding opponents' incentives was available. These two computational phenotypes yield 79% diagnosis classification accuracy and explain 62% of the severity of social symptoms in autistic participants. Such computational decomposition of the autistic social phenotype may prove relevant for drawing novel diagnostic boundaries and guiding individualized clinical interventions in autism.

Heterogeneity in strategy use during arbitration between experiential and observational learning

“…Importantly, these differences remained significant when controlling for additional covariates such as age, gender, education level and cognitive ability. Known difficulties in behavioral adaptation and flexibility during social inference in autism 43,57,58 are consistent with the observed group differences, such that those participants relying on more optimal strategies (arbitration) also exhibit the lowest autism factor score, and those scoring high on autistic traits are more likely to use an irrelevant, non-learner strategy during the task. Trait anxiety, on the other hand, has also been associated with difficulties in adapting learning to changes in uncertainty [35][36][37] , which would also lead to the prediction of reduced reliance on dynamic arbitration in highly anxious individuals.…”

“…Social anxiety has also been found to be associated with excessive deliberation 41 and with suboptimal learning 42 . Finally, autism has been linked to deficits in behavioral adaptation during social inference, specifically suboptimal flexibility and lower mentalizing sophistication 43 , overestimation of the volatility of sensory environment 44 , reduced implicit causal inference about sensory signals 45 , and enhanced observational learning in the aversive domain 46 . Therefore, we hypothesized that anxiety, as well as social dysfunctions associated with autism and social anxiety are likely to be sensitive to the computational heterogeneity in strategy use during EL, OL, and the arbitration between them.…”

Absence of covert face valuation in Autism

“…Thus, likeability may be seen as an attribute of social stimuli that autistic participants would not spontaneously estimate, together with other attributes such as mental states [ 22 , 36 , 37 ]. It could be an explanation for (but not a consequence of) these participants paying little attention to social stimuli in general [ 38 ].…”

Computational models of subjective feelings in psychiatry

“…Computational modeling has led to progress understanding disorders with a social dimension, such as autism, social anxiety, and borderline personality disorder, which have been linked to differences in learning and decision making in interpersonal settings (Fineberg et al, 2018;Forgeot d'Arc et al, 2020;Henco et al, 2020;Hopkins et al, 2021;King-Casas et al, 2008;Siegel et al, 2020). An ongoing challenge is to understand the mechanisms by which affective experience during social decision making is related to psychiatric disorders, something that computational models of subjective feelings have begun to shed light on.…”