Inference or Integration? Mechanisms of Mental State Understanding in High-Functioning Autism

Thursday, May 12, 2016: 11:30 AM-1:30 PM
Hall A (Baltimore Convention Center)
J. Korman1, B. F. Malle1, M. Leboyer2,3, A. Gaman2,3 and T. Zalla4, (1)Cognitive, Linguistic, & Psychological Sciences, Brown University, Providence, RI, (2)Department of Psychiatry, INSERM U 955, IMRB & University Paris Est Créteil, AP-HP, Henri Mondor-Albert Chenevier Hospitals, Créteil, France, (3)Fondation FondaMental, French National Science Foundation, Créteil, France, (4)Institut Jean Nicod, CNRS, Ecole Normale Supérieure, PSL Research University, Paris, France

By the time many high-functioning individuals on the autism spectrum (henceforth HFA individuals) reach adulthood, they have acquired partial competence in dealing with other people’s mental states—a partial “Theory of Mind” (ToM). Nevertheless, HFA adults do not make full use of mental state information for moral judgment (Moran et al., 2011; Zalla and Leboyer, 2011; Zalla et al., 2011), and in faux pas tasks (e.g., Baron-Cohen et al., 1999; Zalla et al., 2009). In particular, they detect unintentional offenses (or “faux pas”) but struggle to understand that such offenses stem from the offenders’ false beliefs.  


In this project we probe the specific mechanisms underlying continued ToM deficits in adulthood. We address two possible accounts of HFA adults’ struggle to understand the offender’s false belief in the faux-pas task.  First, HFA adults may lack the ability to infer novel mental state information from behavior.  Second, even if HFA adults might glean such mental state information, they may lack the ability to integrate the information into a coherent understanding of behavior. We seek to distinguish between these two accounts by manipulating the participants’ task to either infer mental state information or integrate provided mental state information.  


15 HFA adults (diagnoses confirmed with the ADOS and ADI-R, Lord et al., 2000), and 34 typically developing, age, gender, and IQ-matched controls completed the study online.  Each participant saw a total of 8 “faux pas” stories, in which a character committed an unintentional offense because he or she had a false belief.  The stories represented a three-level within-subject factor.  In the “no information” condition, the stories contained no explicit references to the character’s mental states.  In the “belief” condition, the false belief underlying the character’s behavior was made explicit; in the “desire” condition, the character’s desire was made explicit.   For each story, participants (1) answered a question about whether the character possessed a false belief, and (2) provided open-ended explanations for the faux pas.   


Control participants scored barely higher on the belief question (M = 1.65 correct, SD = .49) in the no-information condition than did HFA participants (M = 1.47 correct, SD = .52), t(47) = 1.18, ns.  While both control and HFA participants benefited from receiving explicit belief information F(1, 47) = 17.69, p < .001, only HFA participants benefited from explicit desire information, t(14) = 3.06, p < .01.  Overall HFA and control participants gave comparable numbers of mental states in their behavior explanations, F(1, 26) = .09, ns; but in the “no-information” condition, HFA participants provided even more mental state explanations than control participants, F(1, 29) = 3.62, p =.06.  


We showed some evidence for the integration hypothesis, as HFA adults improved on the belief question when being given either belief information or desire information. Interestingly, when explaining no-information stories, HFA individuals offered more mental states than controls (even though they had marginally poorer performance on the belief question), suggesting that merely mentioning mental state information does not constitute genuine faux-pas understanding.