Behaviour-correlated profiles of cerebellar-cerebral functional connectivity observed in independent neurodevelopmental disorder cohorts
Felipe Morgado,Marlee M Vandewouw,Christopher Hammill,Elizabeth Kelley,Jennifer Crosbie,Russell Schachar,Muhammad Ayub,Robert Nicolson,Stelios Georgiades,Paul Arnold,Alana Iaboni,Azadeh Kushki,Margot J Taylor,Evdokia Anagnostou,Jason P Lerch
DOI: https://doi.org/10.1038/s41398-024-02857-4
2024-04-03
Abstract:The cerebellum, through its connectivity with the cerebral cortex, plays an integral role in regulating cognitive and affective processes, and its dysregulation can result in neurodevelopmental disorder (NDD)-related behavioural deficits. Identifying cerebellar-cerebral functional connectivity (FC) profiles in children with NDDs can provide insight into common connectivity profiles and their correlation to NDD-related behaviours. 479 participants from the Province of Ontario Neurodevelopmental Disorders (POND) network (typically developing = 93, Autism Spectrum Disorder = 172, Attention Deficit/Hyperactivity Disorder = 161, Obsessive-Compulsive Disorder = 53, mean age = 12.2) underwent resting-state functional magnetic resonance imaging and behaviour testing (Social Communication Questionnaire, Toronto Obsessive-Compulsive Scale, and Child Behaviour Checklist - Attentional Problems Subscale). FC components maximally correlated to behaviour were identified using canonical correlation analysis. Results were then validated by repeating the investigation in 556 participants from an independent NDD cohort provided from a separate consortium (Healthy Brain Network (HBN)). Replication of canonical components was quantified by correlating the feature vectors between the two cohorts. The two cerebellar-cerebral FC components that replicated to the greatest extent were correlated to, respectively, obsessive-compulsive behaviour (behaviour feature vectors, rPOND-HBN = -0.97; FC feature vectors, rPOND-HBN = -0.68) and social communication deficit contrasted against attention deficit behaviour (behaviour feature vectors, rPOND-HBN = -0.99; FC feature vectors, rPOND-HBN = -0.78). The statistically stable (|z| > 1.96) features of the FC feature vectors, measured via bootstrap re-sampling, predominantly comprised of correlations between cerebellar attentional and control network regions and cerebral attentional, default mode, and control network regions. In both cohorts, spectral clustering on FC loading values resulted in subject clusters mixed across diagnostic categories, but no cluster was significantly enriched for any given diagnosis as measured via chi-squared test (p > 0.05). Overall, two behaviour-correlated components of cerebellar-cerebral functional connectivity were observed in two independent cohorts. This suggests the existence of generalizable cerebellar network differences that span across NDD diagnostic boundaries.