Language network functional connectivity varies by aphasia type and severity

Aphasia is increasingly understood as a disorder of disrupted language networks rather than damage to isolated language regions. However, most resting-state functional connectivity studies treat people with aphasia as a single group or classify individuals by overall severity, potentially obscuring qualitative differences in network organization across aphasia types. The present study examined how resting-state functional connectivity varies across aphasia types and how aphasia severity relates to network organization following left-hemisphere stroke. Resting-state fMRI data were analyzed from 89 individuals in the chronic stage of recovery drawn from the open-source Aphasia Recovery Cohort dataset. Network-level and ROI-to-ROI functional connectivity analyses were conducted across 32 predefined dorsal and ventral stream language regions. Lesion-overlap analyses were additionally performed to characterize group-level lesion distributions. Network-level analyses revealed relatively limited effects, whereas ROI-to-ROI analyses demonstrated substantial heterogeneity across aphasia subgroups. Broca's aphasia and severe aphasia demonstrated clearer and more spatially convergent connectivity patterns, whereas anomic, mild, and moderate aphasia groups showed greater heterogeneity and limited group-level effects. Groups showing clearer functional connectivity patterns were also characterized by greater convergence in lesion distributions. Within Broca's aphasia, milder impairment was associated with stronger left-hemisphere and interhemispheric connectivity. Overall, aphasia type-based analyses revealed more differentiated connectivity patterns than severity-based groupings alone. These findings suggest that post-stroke language network organization varies across aphasia types characterized by partially shared clinical and lesion features and may not be fully captured by severity measures alone.