Neural Fingerprints of Behavioural Rigidity in Autism

Background:  Recent literature on behavioural rigidity in autistic individuals has suggested that this core symptom might be associated with difficulties in generating task intentions. Empirical evidence for this idea comes from experimental designs in which participants were not only required to switch between two simple cognitive tasks but were also instructed to deliberately choose which of the two tasks to perform next. These voluntary task-switching (VTS) studies revealed that compared to their neurotypical (NT) peers, autistic individuals have a generally stronger tendency to continue repeating tasks (i.e., increased repetition bias) and that once they decide to switch tasks, their performance declines more (i.e., increased switch costs). These observations are considered to reflect behavioural rigidity of autistic individuals within experimental settings.

Objectives:  The current study was developed to follow on this line of research, and its main objective was to specify the neural fingerprints of behavioural rigidity in autistic individuals as reported in the VTS studies.

Methods:  Participants were 31 autistic individuals and 32 NT controls, matched on their age and IQ. They were instructed to choose between two simple cognitive tasks at the beginning of each trial. The tasks included responding to the location or responding to the shape of the presented stimuli. The participants pressed a spacebar to indicate that the choice had been made, which triggered the stimulus presentation. They were then required to respond to either the stimulus location or the stimulus shape, dependent on their task choice for that trial. Using this VTS paradigm, we could investigate the intentional component (task choice) separately from its implementation at the level of task execution (responding to the presented stimulus). We recorded their task choices, as well as their subsequent responding to the stimuli in terms of reaction times (RTs) and error rates. Importantly, the corresponding brain activity was measured with electroencephalography (EEG). The EEG markers of interest were the preparatory contingent negative variation at frontal sites (CNV) as a marker of (the formation of) task intentions and the stimulus-locked P3 at centro-parietal sites as a marker of stimulus-related task execution.

Results:  Our behavioural findings replicate previous observations of a significantly stronger repetition bias and significantly larger switch costs in terms of RTs in autistic participants. Crucially, they demonstrated a significant attenuation of the CNV, which is typically associated with a weaker intentional task preparation. However, no significant differences in the latency or peak amplitude of the P3 were observed between the groups. Specifically, the P3 component demonstrated a usually reported attenuation in switch trials compared to repeat trials, with this pattern being similar in both groups.

Conclusions:  Our results imply that the way in which global task intentions are formed differs between autistic individuals and their NT peers. The way that tasks are actually being executed seem to differentiate only at the behavioural level. Altogether, the present study suggests that the tendency of autistic individuals to engage in repetitive behaviours is associated with the formation of task intentions when tasks are chosen voluntarily.