Structural and functional magnetic resonance imaging correlates of fatigue and dual-task performance in progressive multiple sclerosis

Abstract

Background Frontal cortico-subcortical dysfunction may contribute to fatigue and dual-task impairment of walking and cognition in progressive multiple sclerosis (PMS). Purpose To explore the associations among fatigue, dual-task performance and structural and functional abnormalities of frontal cortico-subcortical network in PMS. Methods Brain 3 T structural and functional MRI sequences, Modified Fatigue Impact Scale (MFIS), dual-task motor and cognitive performances were obtained from 57 PMS patients and 10 healthy controls (HC). The associations of thalamic, caudate nucleus and dorsolateral prefrontal cortex (DLPFC) atrophy, microstructural abnormalities of their connections and their resting state effective connectivity (RS-EC) with fatigue and dual-task performance were investigated using random forest. Results Thirty-seven PMS patients were fatigued (F) (MFIS >= 38). Compared to HC, non-fatigued (nF) and F-PMS patients had significantly worse dual-task performance (p <= 0.002). Predictors of fatigue (out-of-bag [OOB]-accuracy = 0.754) and its severity (OOB-R-2 = 0.247) were higher Expanded Disability Status scale (EDSS) score, lower RS-EC from left-caudate nucleus to left-DLPFC, lower fractional anisotropy between left-caudate nucleus and left-thalamus, higher mean diffusivity between right-caudate nucleus and right-thalamus, and longer disease duration. Microstructural abnormalities in connections among thalami, caudate nuclei and DLPFC, mainly left-lateralized in nF-PMS and more bilateral in F-PMS, higher RS-EC from left-DLPFC to right-DLPFC in nF-PMS and lower RS-EC from left-caudate nucleus to left-DLPFC in F-PMS, higher EDSS score, higher WM lesion volume, and lower cortical volume predicted worse dual-task performances (OOB-R-2 from 0.426 to 0.530). Conclusions In PMS, structural and functional frontal cortico-subcortical abnormalities contribute to fatigue and worse dual-task performance, with different patterns according to the presence of fatigue.