Functional Connectivity of BA 44 During Language Processing In ASD and TD Subjects

Background: Delayed language acquisition and marked deficits in communication skills are hallmark features of autism (Bailey et al., 1996). Converging evidence suggests that Autism Spectrum Disorder (ASD) is characterized by deficient long-range connectivity and excessive local connectivity (Belmonte et al., 2004). Just examined connectivity during sentence comprehension (2004) and proposed that social and behavioral deficits in ASD are related to cortical underconnectivity, in a variety of brain systems. Previous fMRI and functional connectivity studies have implicated dysfunction of the left inferior frontal gyrus (Broca’s area) and left superior temporal gyrus (Wernicke’s area) in ASD (Just et al., 2004). 

Objectives: We sought to further test the hypothesis that individuals with ASD have diminished language-related modulation of functional connectivity in language-relevant networks. Specifically, we examined functional connectivity with the pars opercularis in the left inferior frontal gyrus (LIFGpo), in children and adolescents with ASD using psychophysiological interaction (PPI) functional connectivity analysis.

Methods: 14 typically-developing (TD) and 14 ASD subjects underwent functional magnetic resonance imaging (fMRI) while listening to a stream of artifical syllables (see Scott-Van Zeeland et al., 2010). Prior to analysis, we subjected fMRI data to ICA denoising and regression of nuisance parameters. Time-series from the LIFGpo (Harvard Oxford atlas, 40% probability) were extracted for each subject and multiplied by a hrf-convolved task regressor to create a PPI term. This was then correlated with every voxel in the brain to generate whole brain positive and negative connectivity maps. We have previously presented results using an anatomically-defined seed ROI in LIFGpo (Moore et al., SFN, 2010), which showed deficient connectivity with bilateral superior temporal gyrus. In order to avoid possible confounds based on differential group activation in the LIFG, we verified that both groups had overlapping peaks of activation within our anatomical ROI region. Here we report PPI results using a functionally-based spherical seed ROI, 12mm in diameter, centered on overlapping clusters of activity (Z > 2.0, corrected) in the LIFGpo for ASD and TD subjects.

Results: Between-group analyses found that ASD subjects showed no significant task-related changes in LIFGpo connectivity, while TD subjects showed a wide pattern of task-related increased connectivity with left superior temporal gyrus, left caudate, left amygdala, and posterior and anterior cingulate cortex.

Conclusions: These results suggest that delayed and abnormal language processing in ASD may reflect deficits in long-range connectivity between areas associated with comprehension and production of speech. In addition, deficient connectivity with areas associated with emotional processing and social cognition may underlie additional deficits in processing social and emotional content.