Resolving the cognitive clinico-radiological paradox - Microstructural degeneration of fronto-striatal-thalamic loops in early active multiple sclerosis.

BACKGROUND: Associations between cognitive impairment (CI) and both global and regional brain volumes can be weak in early multiple sclerosis (MS), a dilemma known as cognitive clinico-radiological paradox. We hypothesized that white-matter (WM) integrity within fronto-striatal-thalamic networks may be a sensitive marker for impaired performance in speed-dependent tasks, typical for early MS.

METHODS: Twenty-seven patients with early active relapsing-remitting MS (RRMS) received comprehensive neuropsychological assessment and underwent structural and diffusion-weighted brain magnetic resonance imaging (MRI). Global and regional brain volumes were obtained using FreeSurfer software. Fractional anisotropy (FA) was computed from diffusion tensor images to assess microstructural alterations within three anatomically predefined fronto-striatal-thalamic loops known to be relevant for speed-dependent attention and executive functions.

RESULTS: Overall cognitive performance (Spearman's ρ = .51) and performance in the domains processing speed (ρ = .44) and executive functions (ρ = .41) were correlated with patients' mean FA within the right dorsolateral-prefrontal loop. In addition, overall cognitive performance correlated with mean FA within the right lateral orbitofrontal loop (ρ = .39) - but only before controlling for WM lesion count. In contrast, regional volumes of grey-matter structures within these fronto-striatal-thalamic loops (including the thalamus) were not significantly related to CI. The total brain volume was associated with performance in the domain verbal memory (ρ = .43) only.

CONCLUSIONS: Microstructural degeneration within specific fronto-striatal-thalamic WM networks, previously characterized as crucial for task-monitoring, better accounts for speed-dependent CI in patients with early active RRMS than global or regional brain volumes. Our findings may advance our understanding of the neural substrates underlying CI characteristic for early RRMS.