A Human Gait Circuit Derived from Brain Lesions and Deep Brain Stimulation

ObjectivesThe relevant neuroanatomy for gait dysfunction remains unclear. We sought 1) to identify a brain circuit for gait impairment post stroke, 2) to identify a brain circuit for gait changes after subthalamic DBS for Parkinsons disease, 3) to test for convergence between these two circuits. MethodsThis cross-sectional retrospective study used data from four independent datasets including 109 individuals with stroke and 125 patients with PD who received subthalamic DBS. Gait impairment post stroke was measured using the Combined Walking Index. Gait changes after subthalamic DBS was measured using the gait subscore on the UPDRS Part III. Connectivity between lesion locations or DBS sites and two a priori locomotor regions was computed using a large normative connectome. We repeated this analysis in a data-driven fashion to identify additional connections. ResultsConnectivity between lesion locations and a priori regions in the pedunculopontine nucleus and cerebellar locomotor region were associated with gait impairment post stroke (p< 0.05), along with connectivity to a distributed circuit of other brain regions. Connectivity between DBS sites and our a priori gait regions and connectivity between DBS sites and our lesion-based gait circuit were all associated with gait changes post DBS (p < 10-3). Data-driven gait circuits derived from lesions and DBS showed similar topography and converged on a common brain circuit (spatial r = 0.68, p = 0.0063, 10000 permutations). InterpretationLesion locations impairing gait and DBS sites modulating gait converge on a common brain circuit, which may provide a neuromodulation target for gait dysfunction.