Resting-state and picture-naming EEG reveal aberrant connectomes and outcome-associated dysfunctomes in Cantonese...

Post-stroke aphasia (PSA) is increasingly understood as a network disorder, yet its electrophysiological connectome has been characterized almost exclusively in stress-timed Indo-European languages, and rarely with phase-locking connectivity sampled across multiple frequency bands and recording states in the same cohort. We acquired 61-channel resting-state and overt picture-naming EEG in 31 chronic Cantonese-speaking individuals with aphasia and 14 demographically matched neurologically healthy controls, computed surface-Laplacian-transformed phase-locking-value connectivity in six canonical frequency bands (delta, theta, alpha, low-beta, high-beta, low-gamma), and used Network-based Statistic to identify within-group aberrant naming engagement subnetworks, between-group aberrant subnetwork, and outcome-associated dysfunctomes (subnetworks associated with seven clinical outcomes). Best-supported band × state dysfunctomes were identified through cross-band Akaike model comparison. Results showed that HC and PWA relied on qualitatively different oscillatory subnetworks to shift from rest into naming, with PWA displacing engagement into right-hemisphere and domain-general frontal regions and failing to disengage central alpha. Between-group contrasts highlighted slow-wave-dominant hyperconnective subnetworks in the canonical left temporal regions, reflecting passive cortical dysfunction. Outcome-associated dysfunctomes identified resting-state alpha as the best-fitting candidate for overall aphasia severity, spontaneous speech fluency and informativeness, overall naming, and confrontation naming, and resting-state low-beta as the best-fitting candidate for auditory comprehension. Together, the disrupted naming-engagement subnetworks and peri-lesional slow-wave aberrant connectomes offer mechanistic markers of aphasia-related network reorganization, while the resting-state alpha and low-beta dysfunctomes provide candidate targets to guide outcome-specific precision neuromodulation for PSA.