The Role of Sensory Predictions in Updating Perception and Action

To move the body from one location to another, the brain must know its current position and how that position changes with movement (the body state). While sensory feedback from the body (e.g., vision, proprioception/kinesthesia) can provide this information, the brain is slow to process it in real time, which means it is always delayed relative to the actual movement.

To compensate for this time delay, the brain may generate predictions of the sensory consequences of movement in order to detect and correct errors before the actual feedback information is available. How does the brain combine predictions and feedback signals to form estimates of current and future body state? Elucidating the nature of this process is relevant to many aspects of day-to-day human functioning, such as walking from tile to carpet, working in different gravity environments (air, water, astronauts in zero gravity), and using different types of equipment (gas pedals in cars, computer track pads).

This knowledge is also important to neurological conditions, such as stroke, in which the generation of sensory predictions degrades.

This project examines how the brain combines sensory predictions with sensory feedback to estimate the body's state, using subjective reports of the perceived position and movement of the hand (i.e., sensory perception) as a measure of state estimation. The project employs a motor learning paradigm that has participants adapt target-directed reaching movements to a perturbing force field, which causes sensory predictions about reaching movements to update to account for the novel environmental physics.

Sensory prediction updates are known to alter the selection of movement commands. In the case of force field adaptation, this produces reaching movements that counter the perturbing force and successfully reach the target. This project compares sensory perceptions before and after this intervention to additionally discern whether and how sensory predictions influence brain estimates of body state.

Specifically, this project examines (1) whether sensory predictions influence brain estimates of body position and movement differently, (2) whether sensory predictions influence body state estimates in the same ways that they influence movement command generation (i.e., by assessing overlap in the properties of this influence across the motor and sensory domains), and (3) how the influence of sensory predictions affects brain estimates of the starting point of a movement. This project advances our understanding of the circumstances under which sensory predictions contribute to body state estimates and sensory perception and has the potential to update current models of how these signals influence movement.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.