Alterations of the Theory of Mind Network in Adults with ASD

Background: Deficits in solving Theory of Mind (ToM) tasks have been associated with social impairments in children with autism spectrum disorder (ASD) (Frith, 2001). However, high-functioning adults with ASD perform well on ToM tasks while noticeable social deficits remain (Frith, 2004). Thus, is the ToM brain network functioning neurotypically or do network alterations persist? The few existing imaging studies indicate network alterations but do not deliver consistent conclusions.

Objectives: The aim was to characterize the ToM network in high-functioning adults with ASD using an fMRI dataset obtained during a classical False-Belief/False-Photograph task (Saxe and Kanwisher, 2003) and multiple statistical techniques. 

Methods: The participants (17 high-functioning ASD patients, 17 age/handedness-matched controls) read 24 short stories followed by a statement to be rated as true or false.  Twelve statements required inferring someone’s true or false belief (Belief), twelve were about the representation of the reality on a picture (Photo). Differences in network recruitment (Belief > Photo) and the relative contribution of the different ToM network nodes during mentalizing were assessed. Functional connectivity modulated by mentalizing was compared using a PPI analysis with a seed in the right temporoparietal junction (R-TPJ), which is specifically involved in reasoning about false beliefs (Saxe and Wexler, 2005). The within-subject variability of ToM network node activations was analyzed as described in (Dinstein et al., 2010). A voxel-based morphometry (VBM) analysis was conducted to reveal regional grey matter differences using the Dartel algorithm. 

Results: Both groups showed high task performances (about 90% correct), and activated all the ToM network nodes during mentalizing. Nevertheless, a different weighting of the nodes in the ToM network was observed. The increases of activity in R-TPJ and the right superior temporal sulcus (R-STS) were larger in controls, while activations in L-TPJ and precuneus were larger in the ASD group. A percent-signal-change analysis confirmed  that the balance in the level of activity between the ToM areas was different between groups (group x area interaction). Outside the ToM network, the left inferior frontal gyrus (L-IFG) was hyper-active during mentalizing in the ASD group. Mentalizing strengthened the connectivity between R-TPJ and R-STS as well as L-TPJ in both groups, but the R-TPJ – L-TPJ and R-TPJ – L-IFG connectivity was enhanced in the ASD group. The within-subject variability in cerebral activity during mentalizing was not enlarged in the ASD group. Finally, the VBM analysis revealed decreased hippocampal and increased L-IFG volumes in ASD, in agreement with a recent meta-analysis (Via et al., 2011). 

Conclusions: High-functioning adults with ASD perform well in this ToM task and all the nodes of the ToM network are recruited. However, there are differences in network weighting and connectivity that cannot be explained by within-subject variability and are possibly due to differences in network development.