Atypical Brain Connectivity during a Theory-of-Mind Task in Adults with and without Autism Spectrum Disorder

Background: Individuals with autism spectrum disorder (ASD) exhibit atypical functional brain connectivity across a variety of tasks, generally showing a pattern of long-range underconnectivity and short-range hyperconnectivity in task-relevant networks. For instance, studies find reduced connectivity between the nodes of the theory-of-mind (ToM) network, which mainly consists of the medial prefrontal cortex, temporoparietal junction, posterior superior temporal sulcus, and precuneus, and this decreased connectivity is thought to contribute to the social cognitive deficits that characterize ASD. However, the specific differences in the temporal-spatial dynamics of this ToM network have yet to be determined, and analyzing how brain connectivity differs across time and frequency bands may demonstrate that these differences are more nuanced than previously found.

Objectives: To utilize the excellent temporal and spatial resolution of magnetoencephalography (MEG) to examine timing and frequency differences in brain connectivity between adults with and without ASD during a ToM task.

Methods: We studied 35 control adults (23 M, 12 F; mean age = 26.4±4.9 years) and 43 adults with ASD (30 M, 13 F; mean age = 26.7±5.5). Participants performed a false-belief task, based on the classic Sally-Anne task, to assess ToM, while in the MEG scanner. The task required participants to infer whether a character had a true or false belief about an object’s location. Time series of activity during this task at various brain areas were reconstructed from the MEG data using a vector beamformer. Connectivity between sources was estimated using the weighted phase lag index (wPLI) metric, which was then averaged within several canonical frequency bands (theta, alpha, beta, and low and high gamma) and within a series of 100-ms time windows from 100-600 ms following stimulus onset. Brain connectivity in the false vs. true belief trials were compared. Significant results are reported at a FWE-corrected p<0.05.

Results: Whole-brain network analyses revealed a mixed pattern of under- and overconnectivity in adults with ASD during our ToM task. Between 100-200 ms, adults with ASD showed decreased right temporoparietal junction and ventromedial prefrontal cortex connectivity, and decreased anterior and posterior cingula connectivity in the theta frequency band, as well as increased precuneus and ventromedial prefrontal cortex connectivity, and increased left inferior frontal gyrus and dorsomedial prefrontal cortex connectivity in the beta frequency band. Between 300-400 ms, the ASD group also showed reduced left superior parietal lobule and ventromedial prefrontal cortex connectivity in the low gamma frequency band.

Conclusions: Our results indicate that adults with ASD show distinct patterns of brain connectivity compared to controls across a number of frequency bands, between areas within and outside the ToM network. The most prominent effects seen in the ASD group were the reduced connections in the theta frequency band, which is known to reflect long-range connectivity during cognitive processing. Additional analyses will correlate these connectivity findings with behavioural performance on tasks involving understanding the emotions and intentions of others.