SUMMARY The ability to modify ongoing gait with precise, voluntary adjustments is what allows animals to navigate complex terrains and to execute skilled actions during ongoing pursuits. However, how the nervous system generates the signals to precisely control the lower limbs while simultaneously maintaining ongoing locomotion is still poorly understood. While there is some consensus in the role of spinal cord (SC) central pattern generators (CPGs) in the maintenance of locomotion, little is known about the role of primary motor cortex (M1), especially in the case of volitional adjustments to ongoing locomotion movements. Our own work suggests that M 1 plays at least a dual role in both the monitoring and the adjustment of the hindlimbs during locomotor adjustments, expressed in different subspaces (manifolds) of its neuronal spiking activity. We hypothesize that the emergence of these distinct neural subspaces plays a critical role in the dynamic sensorimotor cortical control of locomotion. Specifically, the emergence of distinct low-dimensional subspaces allows M1 to track sensed ongoing locomotion states, which are primarily driven by spinal-cord CPGs, and to generate occasional descending motor commands to drive volitional adjustments without destructive interference. Besides the basic neuroscience relevance, fully understanding the role of M1 in modulating locomotion is essential for the development of new mobility therapies, including brain-machine interfaces (BMls), for restoring walking and volitional leg control in people with paralysis. We have recently developed a new experimental paradigm to test these hypotheses in nonhuman primates. It enables wireless recordings from microelectrode arrays implanted in multiple sensorimotor cortical areas during natural locomotion and navigation around obstacles visually cued on a treadmill and other spaces. In addition to M 1, we plan to simultaneously record primary sensorimotor cortex (S1) and lower limb electromyo