ABSTRACT Background While structural and functional connectivity changes in multiple sclerosis (MS) are well documented, their complex interplay remains poorly understood. This study identifies co‐fluctuating patterns of structural and functional changes in MS using an MRI‐based multimodal fusion approach and assesses the added value to clinical outcomes. Methods Linked independent component analysis (ICA) was applied to spatial maps of white matter (WM) lesions, fractional anisotropy (FA), gray matter (GM) volume, and functional network connectivity to detect regions with differential co‐fluctuations. Graph theory (GT) was then used to reveal clusters of interconnected brain regions. Linear mixed‐effect models targeted regions with significantly different co‐fluctuating patterns between MS and HC. Multivariate stepwise regressions analyzed the associations between co‐fluctuating patterns and disability and cognitive dysfunction. Results The study included 147 patients with MS and 57 HC. Significant co‐fluctuating patterns of decreased FA, increased lesion probability in the thalamic radiation and corpus callosum, GM atrophy in sensorimotor and thalamic areas, and enhanced functional connectivity in the temporal parietal network distinguished MS from HC ( p < 0.001). GT revealed eight brain sub‐networks of spatially connected clusters. Regional‐ and modality‐specific loadings and GT changes explained physical (adjusted R 2 = 0.51, p < 0.001) and cognitive (adjusted R 2 = 0.44, p < 0.001) disability better than traditional MRI measures (adjusted R 2 = 0.12–0.33, p < 0.001). Conclusions Our multimodal MRI approach revealed co‐fluctuating regional patterns of lesions, structural disconnection, and functional hyperconnectivity in MS, offering a more comprehensive explanation of clinical outcomes than traditional MRI metrics.
Zhang et al. (2025) studied this question.