Low-dimensional dynamics of brain-muscle networks during gait

The neural control of walking requires the interaction between multiple sensory and motor systems. We assess brain-muscle networks during overground gait to assess the dynamic interactions between neural systems involved in gait control. Mobile electroencephalography (EEG) and electromyography (EMG) was recorded from 24 healthy young adults, 24 healthy older adults, and 21 individuals with Parkinson's Disease during overground walking. We estimated dynamic functional connectivity between bilateral sensorimotor cortices and eight leg muscles within a gait cycle. Using multivariate analysis, we extracted brain-muscle networks and investigated the inter-subject variability in the network activation patterns. We identified three orthogonal brain-muscle networks within a gait cycle: 1) a bilateral network that is left-right symmetric and mostly active during the double support phase, 2) a left-lateralised network that is activated during the left swing phase, 3) a right-lateralised network active during the right swing phase. These networks span a low-dimensional subspace in which neuromuscular gait dynamics unfold: the trajectories reveal that the networks are simultaneously co-expressed during the double support phase but with different networks dominating the initial and terminal double support phase. Assessing the inter-subject variability in trajectories, we found an age-related contraction of the trajectories corresponding to a general reduction in neuromuscular connectivity at older age and a selective enhancement of the bilateral network in participants with impaired tactile sensitivity of the foot. The three brain-muscle networks provide a parsimonious description of neuromuscular connectivity during overground walking. The inter-subject variability in low-dimensional trajectories is associated with age and somatosensory function. The selective increase of the bilateral network may reflect a compensation strategy to maintain mediolateral gait stability in response to impaired somatosensory information from the feet. ### Competing Interest Statement The authors have declared no competing interest.