Abstract
Background: Slowed information processing speed (IPS) is an early and prevalent cognitive feature of relapsing–remitting multiple sclerosis (RRMS), although its neurobiological mechanisms remain poorly understood. This study investigated the multimodal basis of IPS impairment in mildly disabled RRMS patients by integrating structural MRI, resting-state functional connectivity (FC), and serum biomarkers of neuroaxonal (sNfL) and astrocytic (sGFAP) injury.Methods: Forty-three individuals with RRMS and 29 healthy controls (HCs) underwent neuropsychological assessment, blood sampling, and MRI scanning. Gray matter (GM) voxel-based morphometry and resting-state independent component analysis (ICA) were performed. RRMS patients were stratified into high-IPS and low-IPS subgroups based on their performance in neuropsychological tests.Results: Compared with HCs, RRMS patients exhibited significant IPS impairment, bilateral thalamic GM atrophy, and elevated sNfL levels. Subgroup stratification revealed a clear dissociation: low-IPS patients showed pronounced bilateral thalamic atrophy, higher white matter lesion volume, and elevated sNfL and sGFAP levels relative to high-IPS patients. In contrast, high-IPS patients showed no evidence of GM atrophy or abnormal sNfL levels compared with HCs, but exhibited significantly increased FC within sensorimotor cortical regions. Stepwise multiple regression identified thalamic GM volume as the strongest independent predictor of IPS, with sGFAP levels and sensorimotor network FC providing significant incremental explanatory value.Conclusions: These findings reveal a clinico-biological dissociation in RRMS. While IPS impairment is driven by neuroaxonal injury and structural collapse, cognitive preservation is maintained through adaptive sensorimotor network reorganization. This functional compensation appears to mitigate cognitive decline until a critical structural damage threshold is reached.