Visual motion signals in extrastriate visual area MT provide the primary sensory input that guides smooth pursuit eye movements. Because of the broad tuning of MT neurons for target direction and speed, many neurons are active in MT when a target moves with any given direction and speed: as a consequence, any given visual motion is represented in the brain by the discharge of a large population of neurons, called the """"""""population response"""""""". The long-term goal of this application is to understand how the population response is read by the motor system to provide commands for smooth pursuit eye movements. Prior work has allowed us to form the hypothesis that target speed is decoded from the population response in MT by performing a vector-averaging computation on an opponent motion signal, where the computation is biased toward estimating low speeds if the population response is noisy or has a low amplitude. We now will ask how target direction is coded and decoded for pursuit. Direction has been chosen for analysis because it offers advantages for understanding how the decoding computation is done with neurons. We will develop an analysis that is based on the mean and variation of individual neural and behavioral responses in awake, trained rhesus monkeys. We will conduct behavioral experiments to determine how well pursuit can discriminate between targets moving in slightly different directions, for stimuli with and without directional noise. We will record the mean and variation of neural responses in MT during the directional discrimination, and investigate co-variation of neural and behavioral responses as well as correlations between the responses of pairs of MT neurons. Then, we will evaluate possible neural mechanisms for decoding target direction by computer simulations of a neural network model with a realistic population code and neuraly plausible decoding mechanisms. Our proposed approach investigates the situation faced by the pursuit system in real life, when it must estimate target direction on the basis of individual responses of many neurons. It will provide us an understanding of the neural operations performed in neural circuits between the cerebral cortex and cerebellum, shedding light on the normal functions of pathways that are compromised in many strokes and motor disorders and potentially leading to new therapies for assisting in recovery from strokes. ? ?
Showing the most recent 10 out of 84 publications
