Delayed Reversal Learning in Autism

Background:  Up to 70% of individuals diagnosed with autism spectrum disorders (ASD) also experience severe, debilitating symptoms of anxiety (Lopata et al., 2010; Reaven, 2010). Anxiety-inducing situations for ASD individuals include changes to the environment or daily routines. Difficulties responding to such changes may lead to consequences for ASD children, their families and communities. Understanding the mechanisms that lead to such difficulties may improve the specificity of treatment for these symptoms. We adapted a reversal learning paradigm used in healthy adults (Schiller et al., 2008) to study older children and adolescents diagnosed with ASD, in order to improve understanding of connections between ventromedial prefrontal cortex (VMPFC) and the amygdala regarding the acquisition and maintenance of fear. To our knowledge this is the first such reversal study in ASD.

Objectives:  In the context of frequent severe anxiety in autism, including core symptoms of insistence on sameness, we sought to identify underlying differences in fear learning and a later switch in the learned contingencies, in ASD and a matched control group.

Methods:  Participants viewed a series of a yellow or blue square presented on the monitor in pseudorandom order. During the Acquisition stage, Color A coterminates on 1/3 of trials with a burst of air directed at the participant’s neck. After 30 total trials of 4 seconds each (12 trials of Color B, 12 of Color A with no reinforcement, and 6 reinforced trials), the task is reversed so that Color B, instead of Color A, may be accompanied by the puff of air. Disposable electrodes are used to collect skin conductance response (SCR) for the duration of the experiment.

Results:  Participants (typical CON n=30; ASD n=30) matched for age (range 11-16 years) and IQ (range 80-135). Both groups showed robust fear learning during Early and Late Acquisition phases, measured as the difference between SCR response to Color A and Color B. During Early Reversal, the ASD group did not differentiate between Color A and B, unlike the CON group who showed the expected greater response to the new threat stimulus. The ASD group finally made the switch during the Late Reversal phase.

Conclusions:  Delayed reversal learning in the ASD group suggests possible disruption in the circuit linking the amygdala and ventromedial prefrontal cortex (VMPFC). The emotional responses found in the amygdala are unable to make the connection to the VMPFC and back again, as necessary for reversal learning. The inability to switch efficiently may underlie symptoms of anxiety in autism including insistence on sameness. Future research will focus on fMRI studies to further observe brain regions used during this reversal task. Present SCR results suggest that difficulty dealing with changes in ASD may occur at a fundamental, early processing level. We also discuss the viability of using the air puff device as a safe but effective aversive unconditioned stimulus.