Introduction: Post-stroke aphasia (PSA) affects 20–40% of stroke survivors and is characterized by heterogeneous impairments across phonological, lexical, semantic, and syntactic domains. Traditional scales largely collapse these dimensions, limiting mechanistic insights into their neural bases. Here, we combined a theory-driven language battery with multimodal rs-fMRI to delineate dissociable neural signatures of compensation and inhibition. Methods: 48 patients with acute PSA (<2 weeks post-onset) and 45 matched healthy controls underwent standardized testing and rs-fMRI. The Chinese Aphasia Language Battery (CALB) was subjected to principal component analysis, yielding latent dimensions of language capacity. Imaging indices included amplitude of low-frequency fluctuations (ALFF), fractional ALFF (fALFF), regional homogeneity (ReHo), and network connectivity via independent component analysis. Group differences and brain–behavior associations were assessed with FWE/FDR-corrected statistics. Results: Behavioral analysis revealed two principal dimensions: lexical–semantic processing and phonological–auditory conversion. PSA patients showed widespread hypoactivity in left perisylvian and frontal regions, coupled with hyperactivity in the cerebellum and right posterior cingulate. Lexical–semantic ability positively correlated with ALFF in the left superior temporal gyrus, underscoring its pivotal role in semantic integration. Conversely, phonological–auditory ability negatively correlated with ReHo in the right middle temporal gyrus, directly supporting the interhemispheric inhibition hypothesis. Network-based statistics further identified large-scale disintegration of interhemispheric connectivity centered on the left middle temporal gyrus, disrupting links to bilateral temporal, parietal, and occipital regions. Conclusion: Our findings provide convergent evidence that acute PSA reflects both compensatory and inhibitory processes. Preservation of left superior temporal activity appears critical for semantic function, whereas maladaptive right-hemisphere overactivation may hinder phonological recovery. Widespread interhemispheric disconnection highlights network-level integration failure and suggests candidate biomarkers for prognosis. By integrating multidimensional behavioral phenotyping with rs-fMRI, this study advances mechanistic models of PSA and offers novel targets for stratified neurorehabilitation and neuromodulation.
Zhang et al. (Thu,) studied this question.