Cerebral Basis of the Decision-Making Difficulties Experienced By Persons with Autism Spectrum Disorder in an Unstable Context

Background:  Constructing an accurate representation of the world, despite its uncertainty, is a challenge for people with autism. These difficulties could be even more obvious in a social environment because of its very changing nature. In a previous study (Robic et al., submitted), we found that an unstable context constitutes a major source of difficulty for people with Autism Spectrum Disorders (ASD) that impacts the decision-making process. A cross effect between the influence of the social nature of the environment and its instability was also found.

Objectives: We conducted an fMRI study, aiming at determining the neural correlates of the difficulties experienced by ASD people in an unstable context. We investigated the influence of an environment either stable (associated with constant probabilities) or unstable (associated with changing probabilities), either social or non-social, on the brain activity of persons with High Functioning Autism (HFA) and Asperger Syndrome (AS) compared to controls. 

Methods: We proposed a lottery task, adapted from Behrens et al. (2008), to sixteen subjects with High-Functioning Autism (HFA) or Asperger Syndrome (AS) and 16 controls matched with respect to age, gender and IQ. Participants were presented with two boxes with hidden reward probabilities and were asked to choose between them. The goal was to score as many points as possible. A cue was presented at the beginning of each trial before the participants made their choice. In the non-social environment the cue was an arrow, in the social one the cue was a movie, which depicted the choice of an actor. Participants received feedback about their choice after each trial and their score updated accordingly. In each kind of environment (social and non-social), two conditions were proposed: a stable (constant probabilities) and an unstable one (probabilities switched every 20 trials). The study was conducted using a 1,5 Tesla MRI scanner. Functional data were analyzed using the SPM8 software package.  

Results: HFA/AS participants showed lower performance than controls. They also showed a difficulty to redirect their attention when contextual cues were not relevant. We found in HFA/AS participants a weaker engagement of a fronto-parietal attentional network in an unstable context (regardless whether the cue was social or non-social). Importantly, this difference was not observed in the stable context. We found in controls a greater activation of the ACC in the unstable context compared to the stable one, but not in HFA/AS participants. Despite the fact that HFA/AS subjects succeeded in the task in the same proportion than controls in the social environment, social information processing activated regions involved in action understanding and emotion processing to a greater extent in controls than in HFA/AS subjects.  

Conclusions: We propose that the impaired ability to build prediction in an unstable context experienced by persons with autism could be linked to a less important engagement of both the attentional cerebral network and the ACC, a key structure involved in the estimation of the context instability. This deficit might not be specifically linked to the social nature of the information.