The abstract of his talk from NCM 2026:
Evolving thoughts on motor cortex
Stephen Scott, Queens’s University
How does the human motor system produce goal-directed movements given multiple feedback pathways: spinal, subcortical, and cortical? This question occupied much of my career. Early on, as a graduate student studying cat locomotion, I viewed motor control through the traditional Jackson/Sherrington hierarchical framework, where supraspinal regions influence the spinal cord, which generates muscle activity. This model is straightforward to simulate using engineering principles. However, during my postdoctoral work on voluntary control in non-human primates, this strict hierarchy seemed less convincing. Primates have corticospinal projections targeting both spinal interneurons and motoneurons, indicating more complex interactions. Later, approaching the problem with optimal feedback control theory, our research emphasized the importance of transcortical feedback for goal-directed actions, with spinal feedback playing a limited role. Still, I found it challenging to explain such shifts in feedback processing across species and behaviors. Modeling multiple parallel pathways often felt speculative, with strong but conflicting views about the dominance of each level. Our recent work offers fresh insight by making one feedback pathway trainable, allowing it to learn the contributions of the other non-trainable pathway—effectively one part of the motor system modeling another. As predicted, increased spinal feedback during mechanical disturbances corresponded with reduced motor cortex activity. This suggests that goal-directed movements arise from the combined contributions of multiple feedback pathways, with transcortical feedback providing motor commands that are the difference between what is generated by subcortical pathways and what is required for goal-directed behaviour.