Neural Bases of Prosodic Processing Differences in ASD: Resting State Functional Connectivity of Right Planum Temporale

Background: Differences in speech prosody have long been identified as characteristic of autism spectrum disorder (ASD). Transcription-based and acoustic studies of intonation in ASD reveal differences in resonance quality and misplaced utterance stress, as well as elevated pitch, increased pitch range, and pitch excursions. Also reported are deficits in comprehension of affective prosody and discrimination of prosodic tunes. These prosodic differences are observed in individuals who display otherwise normal language development and functioning.

Objectives: In typically-developing (TD) individuals, prosody is right-lateralized, involving particularly the right planum temporale (PT). The neural bases of prosodic processing differences in ASD are poorly understood. To clarify these mechanisms, we use resting-state functional magnetic resonance imaging (rsfMRI) data collected from ASD and TD youth to test the hypothesis that these groups differ in functional connectivity (FC) between right PT and other brain regions.

Methods: Resting-state data for ASD and TD boys aged 7-15 years (y) from the Autism Brain Imaging Data Exchange were analyzed. Inclusion criteria were right-handedness and FSIQ ≥ 80; the ASD group was comprised of participants diagnosed with ASD or Asperger's. Youth were binned into age groups: 7-11y (ASD n = 48, TD n = 55); and 12-15y (ASD n = 69, TD n = 75). FC analysis used the CONN toolbox (v17e) in SPM 12. The default MNI-space preprocessing pipeline was used, with two modifications: removal of initial 4 volumes; smoothing kernel of 6mm³.Participants for whom >20% of volumes were flagged as having excessive motion (defined as ≥ 0.9mm motion, measured by framewise displacement) were excluded from analyses. Study site, eye status (eyes open vs. closed), and full-scale IQ were covariates of no interest. Seed-to-voxel analyses, correlating activity in right PT (10mm³ sphere centered at MNI coordinates [52 -31 15]) with activity in the remainder of the brain, were conducted. Multiple comparison correction used an initial cluster defining threshold of p < 0.001, k = 10, and false discovery rate p < 0.05.

Results: 7-11y: ASD > TD: ASD versus TD youth showed greater FC with clusters in left precentral (PrG) and middle frontal gyri (MFG), and right cerebellar posterior lobe (Crus I, VII). 12:15y: TD > ASD: TD versus ASD youth exhibited increased FC with left parahippocampal gyrus (PHg).

Conclusions: ASD and TD youth differed for FC with the right PT, a key brain region in processing the relatively slow fundamental frequency movements characteristic of intonation. Groups differed in PT connectivity with other brain regions involved in speech production (PrG) or perception (MFG, Crus I), including pitch modulation(PrG) and pitch memory (PHg). These FC differences at rest may plausibly relate to differences seen during prosody processing, given that brain networks identified during rsfMRI are predictive of neural activity during tasks. Future work should directly probe the relationship between PT activation and activity in brain regions reported here during prosody-related task-based fMRI. Additionally, it remains to be determined whether the FC differences shown are attested in ASD girls; ideally, sex-balanced cohorts could be investigated to test for diagnosis- vs. sex-specific differences.