Identification of Cortical and Subcortical Correlates of Cognitive Performance in Multiple Sclerosis Using Voxel-Based Morphometry

Abstract

Objective: Cognitive impairment is an important feature in multiple sclerosis (MS) and has been associated to several Magnetic Resonance Imaging (MRI) markers, but especially brain atrophy. However, the relationship between specific neuropsychological tests examining several cognitive functions and brain volumes has been little explored. Furthermore, because MS frequently damage subcortical regions, it may be an interesting model to examine the role of subcortical areas in cognitive functioning. Our aim was to identify correlations between specific brain regions and performance in neuropsychological tests evaluating different cognitive functions in a large series of patients with MS. Methods: A total of 375 patients were evaluated with a comprehensive neuropsychological battery and with MRI. Voxel-based morphometry was conducted to analyse the correlation between cognitive performance and gray matter damage, using Statistical Parametric Mapping with the toolboxes VBM8 and Lesion Segmentation Tool. Results: The following correlations were found: Corsi block-tapping test with right insula; Trail Making Test with caudate nucleus, thalamus, and several cortical regions including the posterior cingulate and inferior frontal gyrus; Symbol Digit Modalities Test with caudate nucleus, thalamus, posterior cingulate, several frontal regions, insula, and cerebellum; Stroop Color and Word Test with caudate nucleus and putamen; Free and Cued Selective Reminding Test and Rey-Osterrieth Complex Figure with thalamus, precuneus, and parahippocampal gyrus; Boston Naming Test with thalamus, caudate nucleus, and hippocampus; semantic verbal fluency with thalamus and phonological verbal fluency with caudate nucleus; and Tower of London test with frontal lobe, caudate nucleus, and posterior cingulate. Conclusion: Our study provides valuable data on the cortical and subcortical basis of cognitive function in MS. Neuropsychological tests mainly assessing attention and executive function showed a stronger association with caudate volume, while tests primarily evaluating memory were more strongly correlated with the thalamus. Other relevant regions were the posterior cingulate/precuneus, which were associated with attentional tasks, and several frontal regions, which were found to be correlated with planning and higher order executive functioning. Furthermore, our study supports the brain vertical organization of cognitive functioning, with the participation of the cortex, thalamus, basal ganglia, and cerebellum.

Keywords: cognitive impairment; executive functioning; magnetic resonance imaging; memory; multiple sclerosis; neuropsychological assessment; voxel-based morphometry.

Figure 1

Statistical parametric map overlaid on an MRI template (neurological orientation, axial view), showing brain regions positively correlated with SDMT performance.

Figure 2

Statistical parametric map overlaid on an MRI template (neurological orientation, axial view), showing brain regions positively correlated with BNT.

Figure 3

Statistical parametric map overlaid on an MRI template (neurological orientation, axial view), showing brain regions positively correlated with FCSRT (total free recall).

Figure 4

Statistical parametric map overlaid on an MRI template (neurological orientation, axial view), showing brain regions positively correlated with verbal fluencies. Animals, in purple.

Figure 5

Statistical parametric map overlaid on an MRI template (neurological orientation, axial view), showing brain regions positively correlated with verbal fluencies. “M” words, in red.

Figure 6

Statistical parametric map overlaid on an MRI template (neurological orientation, sagittal view), showing brain regions positively correlated with Tower of London (uncorrected p < 0.001).