“Resting Brain” In Autism: Functional Deactivation and Connectivity of the Default Mode Network

Background: The default mode network (DMN) is a collection of brain areas consistently deactivated during task performance (Raichle et al., 2001). Several components of the DMN make up the Task Negative Network (TNN) which has been implicated in self-referential processing; and studies have shown reduced task-related deactivation of this network in autism, suggesting social impairments (Kennedy et al., 2006; Kennedy & Courchesne, 2008). These results are further supported by recent evidence of association between default mode brain function and processing information within social contexts (Schilbach et al., 2008).

Objectives: The primary aim of this fMRI study was to investigate deactivation and functional connectivity of the DMN in autism during resting baseline, and to examine the effect of prior task (social cognitive or linguistic) on this response.

Methods: Thirteen high-functioning adults with autism and fourteen typical control participants took part in three fMRI studies (two language tasks: word and sentence comprehension, and a Theory-of-Mind (ToM) task). Each of the studies had separate blocks of fixation baseline, in which participants were asked to relax and think about nothing. The data from these separate fMRI studies (task blocks and fixation blocks) were collated to examine deactivation and functional connectivity using previously established methods (Fair et al., 2007). Data were acquired from a Siemens 3.0T Allegra head-only scanner and analyzed using SPM8.

Results: 1) Within-subject analyses revealed that controls exhibited stronger deactivation in key regions of the DMN (ventromedial prefrontal, anterior cingulate, posterior cingulate cortices) during task-related performance. Participants with autism, however, failed to show significant deactivation in any DMN regions; 2) Between-group analyses, using a 2x2 ANOVA design, revealed that individuals with autism showed lesser deactivation in TNN areas (e.g., left medial prefrontal cortex); 3) When the tasks were separated to examine fixation blocks following social cognitive or linguistic tasks, using a Group (Autism, Control) x Tasks (Fixation, Language, ToM) ANOVA, fixation blocks following the language task revealed no group difference. However, fixation blocks occurring during the ToM task showed less deactivation in autism than in controls in several DMN regions; and 4) A functional connectivity analysis revealed decreased connectivity within the TNN for individuals with autism, specifically between dorsomedial prefrontal cortex and angular gyrus.

Conclusions: The overall lack of deactivation in DMN in autism has implications for self-reflection processes, suggesting that DMN may be a compelling framework for understanding the social brain in autism. Our study examined resting blocks from fMRI tasks, providing a unique dataset for investigating the effect preceding task may have on rest and DMN deactivation. While language tasks did not show deactivation differences between the two groups, the ToM task elicited lack of deactivation in autism. This difference may attest to the difficulty individuals with autism have in modulating the recruitment of DMN regions depending on task; the deactivation becomes less when the task involves self-other reflections. Our findings suggest more in-depth examination of the DMN in autism to better understand not only the resting brain, but also its relationship to social cognition in autism.