Bottom-up and Top-down Mechanisms of Visually-Guided Movements
Interacting with the world is a two-step process of accurate sensing followed by coordinated movement. Optimization of biologically-inspired robotic systems benefits from the quantification and modeling of natural sensorimotor behavior, including the bottom-up circuits that mediate it and top-down cognitive influences that modulate it. A critical sensorimotor behavior in everyday life is the generation of rapid eye movements, called saccades. By making saccades 2-3 times/second, we scan visual scenes and integrate the incoming visual signals to construct an internal representation of what is around us. Much is still unknown about the neural processes that act on visual input and the nature of the resulting internal construct. To study this, we first created a model with architecture inspired by known visuomotor circuits in the brain. By training the model to achieve visuomotor stability while varying its visual and motor inputs, we found that it converged onto a solution that resembled and explained a dynamic neural process that had been documented electrophysiologically. Second, in a psychophysical experiment, we kept constant the visual stimuli and motor actions but manipulated the expectations of what subjects thought would happen. We found that visual perception systematically changes based on expectation, providing evidence for cognitive influences on visuomotor integration and continuity. Third, we expanded the work to whole-body orienting in an immersive virtual environment. While performing a marksmanship task, subjects learned to precisely intercept moving targets. Analysis and modeling of the dynamics of movement revealed mechanisms of learning in this realistic behavioral context. Taken together, the studies provide a link between the ensemble activity of neurons and perceptual experience, demonstrate that perception is a combination of incoming signals and prior beliefs, and move the field toward the study of perception-action cycles during natural human behavior.
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.
Rights for Collection: Duke Dissertations