The Relationship Between Resting-State Functional Connectivity Between Language Areas and Oral Comprehension in Children with Autism

Background:  Autism spectrum disorder (ASD) is characterized by socio-communication deficits, including receptive language deficits.  Resting-state functional connectivity has been useful to the study of neural substrates of speech and language but has seldom been used to study receptive language deficits in children with ASD.

Objectives: To determine how resting state functional connectivity affects receptive language by comparing resting-state functional MRI between commonly-recognized anterior and posterior language areas with performance on Oral Comprehension subtest of the Woodcock Johnson Test of Cognitive Skills III (WJ3) in typically developing (TD) children and those with ASD.

Methods: Subjects were scanned during performance of the modified flanker task using an echo-planar imaging (EPI) sequence with the following parameters: TR = 3000ms, TE = 30ms, flip angle = 90^o, matrix size = 64x64, FOV = 192x192mm², 50 slices with a thickness of 2.8mm and a 0.2mm gap for an effective resolution of 3.0x3.0x3.0mm³.  We obtained resting state data from our subjects by analyzing BOLD fluctuations associated with the fixation condition blocks of a modifier flanker task.  There were a total of four fixation blocks (42 secs each) for a total non-task time of 3 minutes, 12 seconds.  We measured functional connectivity between common cortical substrates of language areas and their right-hemispheric homologues using partial-correlation in 24 TD children and 24 children with ASD (ages 6-18).  Subjects who exhibited large motion artifacts were excluded from all groups (ArtRepair: threshold of 10% of scans per subjects having > 1.5% average global whole volume signal change).  Bilateral regions of interest (ROIs) included Brodmann's area (BA) 41+42 (Heschl’s gyrus, including primary auditory cortex), anterior BA 22 (word-form area), posterior BA 22 (Wernicke’s area), BA 44 (Broca’s area, pars opercularis), and BA 45 (Broca’s area, pars triangularis).

Results: Fifty-three percent (24/45) of the correlations comparing functional connectivity between ROIs and standardized WJ3 Oral Comprehension subtest scores in children with ASD were significant (p < 0.05), but none were significant in TD children.  Seventy-one percent (17/24) of these connections were frontotemporal, and 29% were temporotemporal.  Further, 50% of these connections were inter-hemispheric, 29% were intra-hemispheric on the left side, and 21% (5/24) were right intra-hemispheric on the right side.  ROI-pairs that included left BA 41+42, right anterior BA 22, and left or right BA 44 correlated best with oral comprehension.  Correlations were never greater in TD children than in children with ASD, and we statistically verified that differences between correlational coefficients in the TD and ASD groups were significant.  Similar comparisons with other WJ3 subtests were not as robust.  None of these findings extended to functional connectivity within the default mode network. 

Conclusions: Resting state functional connectivity between language areas is a better predictor of oral comprehension in children with ASD than in TD children.  The effect did not extend to the default mode network, indicating that this relationship is specific to the connectivity between cortical areas important for language.  These results illustrate that functional connectivity between language areas is a predictor of receptive language ability in children with ASD.