Neural Networks for Anxiety? Decreased Integration in ASD of Sensorimotor and Emotional Pathways That Support Classical Fear Conditioning

Background:   Multimodal neuroimaging methods have the potential to elucidate specific networks underlying behavioral symptoms related to autism spectrum disorders. We recently performed a classical fear conditioning fMRI experiment that revealed abnormal right amygdala and left anterior insula activation in ASD during early fear acquisition. We collected diffusion tensor imaging (DTI) data from the same individuals and hypothesized that DTI based on our fMRI results could expand understanding of the neural basis for co-morbid anxiety that is commonly seen in ASD. 

Objectives:   We hypothesized that atypical function during acquisition and extinction of learned fear may be due to differences in white matter connectivity in neural networks related to integrating fear recognition and response. Low connectivity measures may be associated with a decreased ability to adapt to changing contexts and lead to subsequent symptoms of uncertainty and anxiety. 

Methods:  Participants included seventeen adults ages 18-29 diagnosed with ASD and age- and IQ- matched healthy controls. ROIs in right amygdala and left insula, de­fined from our fear conditioning fMRI data, were used as seeds for a tractography analysis using FSL’s probtrackx software; this software estimates a “connectivity distribution” for each participant based on the initial seed. These participant-specific connectivity distributions were used to create a mask for each ROI such that voxels in the mask had connections with the ROI in at least half of the participants (collapsed across groups). We log-transformed the connectivity distribution maps to minimize skew, then performed group t-tests on the connectivity distribution measures for voxels within the masks using a voxel-wise threshold of p<.02 (two-tailed) and spatial extent threshold of >20 contiguous voxels.

Results:   Analyses of the right amygdala revealed significantly lower connectivity within the uncinate fasciculus (see Figure 1), inferior fronto-occipital fasciculus, and inferior temporal gyrus white matter tracts for ASD. Our left anterior insula mask showed greater connectivity of the inferior frontal gyrus in ASD, but overall decreased connectivity in the thalamocortical tract, superior longitudinal fasciculus, corticospinal tract, and superior frontal gyrus white matter compared to controls (see Figure 2). 

Conclusions: These DTI data expand on our fMRI BOLD data showing decreased amygdala and insula activation during fear acquisition but increased activation during contexts (e.g., extinction trials) that should be safe. DTI highlights underconnected networks from amygdala to frontal lobe monitoring and decision making areas, and insula cortex to integration of sensorimotor information. These data support our hypothesis that chronic everyday anxiety in individuals with ASD may arise from uncertainty regarding environmental cues related to fear and safety. Improved understanding of the neural mechanisms that underlie unique manifestations of anxiety in ASD may provide targets for etiological research as well as for better treatment specificity.