Abstract
BACKGROUND: Post-stroke cognitive impairment (PSCI) involves complex interactions between gray matter (GM) and white matter (WM) pathology, yet their covarying patterns remain poorly characterized. METHODS: We conducted a multimodal meta-analysis following PRISMA guidelines (PROSPERO registration number: CRD420251078162). Systematic literature searches identified 7 gray matter volume (GMV) (292 PSCI patients/231 controls) and 8 diffusion tensor imaging fractional anisotropy (DTI-FA) studies (339 PSCI patients/258 controls). Seed-based d Mapping with Permutation of Subject Images (SDM-PSI) was used for voxel-wise meta-analyses. Spatial integration identified GM-WM covarying pathology. RESULTS: PSCI patients showed significant global GMV reduction versus controls (p = 0.011). Family-wise error (FWE)-corrected analyses revealed GM atrophy in the right medial superior frontal gyrus (BA 10), right superior temporal gyrus/temporal pole (BA 38), and left orbital middle frontal gyrus (BA 47) (all p < 0.05, FWE-corrected). Concurrent GMV reduction and FA decrease occurred in the left anterior cingulate/paracingulate gyri (BA 11/25; p = 0.043, uncorrected). Meta-regression demonstrated negative associations between age and GMV in the right medial superior frontal gyrus (BA 10) (p = 0.006) and FA in the left median network, cingulum (p = 0.021). Heterogeneity was low (I (2) < 50%). Egger's test suggested potential publication bias for the right temporal pole and DTI findings. CONCLUSION: This study applied SDM-PSI multimodal meta-analysis to delineate characteristic neurostructural damage patterns in PSCI. Key findings encompass gray matter atrophy within frontotemporal cognitive hubs (FWE-corrected). Additionally, exploratory analyses suggested potential covarying GM-WM pathology in the left anterior cingulate/paracingulate gyri (BA 11/25; p = 0.043, uncorrected). These findings elucidate PSCI pathogenesis and suggest potential anatomical targets for future research.