Brain's Language Networks Underconnected in Non-Verbal Autism
Source PublicationN/A
Primary AuthorsMontaña-Valverde, Linke, Slušna et al.

Language is universal in humans, develops robustly in infancy, and is rarely absent entirely in aphasia following strokes. Non- or minimally verbal children, in whom language has not developed by school-age in either production or comprehension, thus provide an unparalleled window into the human brain when it is deprived of language function. Yet, insights from functional MRI are absent in this population.
Researchers conducted a first study of its kind, examining intrinsic connectivity in 9 nvASD children, scanned under sedation with propofol, alongside 8 typically developing children (scanned awake). Using resting-state (rs) fMRI, they targeted both functional (FC) and anatomically constrained, generative effective connectivity (GEC) within speech and language networks, the insula and associated networks, and the hippocampus. The findings indicated that nvASD children, compared to their neurotypical peers, showed a widespread pattern of hypoconnectivity—or underconnectivity—across auditory speech perception, frontotemporal, and semantic processing regions, extending to the insula and hippocampus. These patterns were selectively replicated by GEC results and correlated with Autism Diagnostic Observation Schedule (ADOS) behavioral scores within the nvASD group.
Importantly, the researchers evaluated sedation confounds by applying identical analyses to an independent rsMRI dataset of healthy adults scanned both under propofol and while awake. This hypoconnectivity pattern extended neither to the adults scanned under propofol, nor to the visual cortex used as a control region in nvASD, suggesting that the observed pattern of results could not be explained by sedation. As lead author Montaña-Valverde notes in the paper, "Together, this first evidence from intrinsic connectivity reveals a broad pattern of underconnectivity across key cognitive networks, which provides a neural correlate for the significant breakdown of language-related cognitive functions in this population."