2016 |
Rao, Hrishikesh |
F31Activity Code Description: To provide predoctoral individuals with supervised research training in specified health and health-related areas leading toward the research degree (e.g., Ph.D.). |
Visual Continuity Across Saccades
? DESCRIPTION (provided by applicant): We make multiple rapid eye movements, called saccades, every second. Each saccade disrupts the flow of visual information to the brain, yet our perception of the world is not that of jumpy images, but of a smooth and stable scene. One hypothesis behind this visual continuity is that the brain uses presaccadic visual information, combined with a copy of the oculomotor command, to predict the post-saccadic visual scene. After each eye movement, the afferent visual information may be compared with the prediction to evaluate object stability and update our internal model of the world. It was shown recently that neurons in the frontal eye field (FEF) may contribute to this pre-/postsaccadic comparison, because they report whether visual stimuli change across saccades. My overall goal is to build on this previous work, using my background in biomedical engineering, to investigate transsaccadic visual stability through computationally grounded psychophysics and single neuron recordings. Our world is a complex landscape of stationary and dynamic objects, and therefore, my overall hypothesis is that our beliefs about object motion, our priors, affect visul perception during saccadic scanning of a scene. My first aim is to use novel psychophysics experiments to reveal how priors, coupled with visual evidence, influence perception across saccades. All of the experiments are grounded in a classic task called saccadic suppression of displacement, which permits fine-scale quantification of whether the visual scene appears stable across saccades. I will test the perceptual effects of three types of priors using the task: expectations about object movement, expectations about object identity, and expectations about the statistics of images that are not discrete objects. My second aim is to determine how priors are encoded by neurons in the FEF and influence their signals about transaccadic changes. I will record from FEF neurons in non-human primates as they perform a task of transaccadic visual stability that involves a range of learned priors about object movements. My prediction is FEF activity will not only report transaccadic changes in visual objects, but also will be modulated by priors just as behavior is modulated. The overall outcome of my project will be to advance our understanding of visual continuity across saccades at both the perceptual and neural levels, by taking into account the fact that subjects have a variety of expectations about the diverse elements of a scene.
|
1 |