Clinical disability in multiple sclerosis (MS) is insufficiently explained by structural damage as measured with standard magnetic resonance imaging (MRI) measures. More advanced measures of brain network atrophy and functional network changes might better explain symptoms and clinical deterioration.
To investigate the relevance of functional network alterations in addition to network atrophy for explaining physical disability in MS.
In this cross-sectional study 143 MS patients and 36 healthy control participants underwent resting-state magnetoencephalography (MEG) and structural MRI. Functional connectivity between regions was estimated using the phase lag index, from which the minimum spanning tree (MST) was constructed, representing the backbone of the functional network. The topology of the MST was described using the so-called tree hierarchy (MST-Th). Gray matter (GM) volume was calculated within literature-based resting-state network maps (i.e. visual, sensorimotor, dorsal attention, ventral attention, limbic, fronto-parietal, default mode, deep gray matter, and cerebellar networks). Physical disability was quantified with the Expanded Disability Status Scale (EDSS), Nine Hole Peg Test (9HPT) and Timed 25-Foot Walk Test (TWT). Network atrophy and topology were compared between groups and related to disability.
Atrophy was apparent in all resting-state networks. All volumes correlated positively (p<.001) with EDSS and 9HPT: Spearman’s ρ between .289 and .567, highest correlations for sensorimotor, default mode, fronto-parietal and dorsal attention networks. EDSS correlated negatively with MST-Th in the lower alpha band (α1) (p < 0.008), while 9HPT correlated negatively with MST-Th in the upper and lower alpha, gamma, delta and theta bands (p <0.05), indicating a less efficient network relating to worse disability. TWT was related to atrophy in all networks, but not network topology. Together, MST-Th-α1, age, cerebellar and fronto-parietal atrophy explained 36% of EDSS variance, while 19% of 9HPT variance was explained by deep GM atrophy and MST-Th-α1. Lesion volume had no added significant effect on variance.
These results suggest that more advanced measures of network atrophy and functional network topology can explain a significant degree of disability variance in MS. In addition, mobility scores were not related to network changes, which could imply different underlying pathological substrates compared to those that underlie upper limb dexterity.