Resting-state functional connectivity of subcortical locomotor centers explains variance in walking capacity

Abstract

Walking capacity influences the quality of life and disability in normal aging and neurological disease, but the neural correlates remain unclear and subcortical locomotor regions identified in animals have been more challenging to assess in humans. Here we test whether resting-state functional MRI connectivity (rsFC) of midbrain and cerebellar locomotor regions (MLR and CLR) is associated with walking capacity among healthy adults. Using phenotypic and MRI data from the Nathan Kline Institute Rockland Sample (n =119, age 18-85), the association between walking capacity (6-min walk test distance) and rsFC was calculated from subcortical locomotor regions to 81 other gait-related regions of interest across the brain. Additional analyses assessed the independence and specificity of the results. Walking capacity was associated with higher rsFC between the MLR and superior frontal gyrus adjacent to the anterior cingulate cortex, higher rsFC between the MLR and paravermal cerebellum, and lower rsFC between the CLR and primary motor cortex foot area. These rsFC correlates were more strongly associated with walking capacity than phenotypic variables such as age, and together explained 25% of the variance in walking capacity. Results were specific to locomotor regions compared with the other brain regions. The rsFC of locomotor centers correlates with walking capacity among healthy adults. These locomotion-related biomarkers may prove useful in future work aimed at helping patients with reduced walking capacity.

Keywords: brain; gait; locomotion; magnetic resonance imaging; network.

Figure 1

Conceptual models of locomotor control. Arrows show supraspinal connections thought to be the most relevant for automatic and conscious locomotor control, with the most emphasized connections for each type of locomotor control shown in black. Known inhibitory connectivity is shown by dashed lines. MLR, midbrain locomotor region; CLR, cerebellar locomotor region; M1F, primary motor cortex foot area; SMA, supplemental motor area; PMd, dorsal premotor cortex; PPN, pedunculopontine nuclei; CN, cuneiform nuclei; FN, fastigial nuclei; DN, dentate nuclei; Lat Cblm, lateral cerebellum; PMRF, pontomedulary reticular formation; CPGs, central pattern generators

Figure 2

Primary MLR and CLR rsFC associations with walking capacity (n = 119). Seed‐to‐ROI rsFC significantly associated with 6MWT distance (two‐sided nonparametric p _FDR < 0.05), shown in MNI152 space at ROI centroid coordinates. Areas with positive and negative mean seed‐to‐voxel rsFC (|T| > 2) are translucently shown in pink and purple, respectively, with black outlines. The bottom row shows Fisher z rsFC values for each gait‐related rsFC pair plotted against 6MWT distance, with color coding for participant age and sex. MLR, midbrain locomotor region; CLR, cerebellar locomotor region; rsFC, resting‐state functional connectivity; 6MWT, 6‐min walk test; SFG, superior frontal gyrus; ACC, anterior cingulate cortex; Cblm, cerebellum; M1F, primary motor cortex foot area; FDR, false discovery rate