10/ Many thanks to the whole team (Zachary Loschinskey,
Munib Hasnain, @briandepasquale.bsky.social, @mikeeconomo.bsky.social) who made this work possible!

9/ Takeaway: Together these results point to a previously unrecognized, fast timescale layer of flexibility in motor control and provide a framework for tracking how cortical circuits dynamically govern behavior as demands evolve.

8/ Takeaway: When future actions and/or their sensory consequences are unpredictable, the cortex stays engaged to monitor and guide movement. We describe how this may represent an instantiation of a sensorimotor form of attention.

7/ During learning, engagement is initially prolonged under high uncertainty and sharpens into transient engagement as animals learn how a task is structured and uncertainty falls.

6/ Using a task in which animals are rewarded optogenetically, we demonstrate that engagement does not reflect a motivated or reward related state – but is sensitive to sensorimotor uncertainty specifically.

5/ Crucially, whether engagement persists depends on the contextual demands of the movement: cortical engagement is sustained when animals are uncertain about upcoming events, such as the delivery of a reward, and it dissipates when outcomes become predictable.

4/ When the motor cortex is engaged, population spike rates are elevated and necessary for initiating and precisely controlling movement kinematics. Remarkably, the cortex can rapidly disengage on sub-second timescales even as movement continues in a smooth, uninterrupted manner.

3/ Using multi-site population recordings, optogenetics, statistical modeling of neural state transitions, and tightly controlled behavioral tasks, we found that cortical control can switch between “engaged” and “disengaged” states during a single continuous movement.

2/ Further, it’s unclear whether cortical involvement is fixed during movement or can toggle on and off within an action.

1/ Many studies have shown that the motor cortex can be essential for some movements but dispensable for others, and its role can change with learning. As a result, when exactly the motor cortex contributes to movement has been difficult to predict, control, and formalize.