Saccade Target Selection in Naturalistic Visual Search

点击翻译按钮获取中文摘要

Project Summary Understanding how we select targets for saccadic eye movements is a basic research question of far-reaching importance. These rapid eye movements allow the fovea to fixate objects of interest, and we typically make more than 170,000 saccades each day. Saccades are essential for efficient visual perception and action: from reading to cooking to driving, most common behaviors heavily engage the saccadic system, and a dysfunctional mechanism for selecting saccade targets would impair performance in all of these everyday activities. The superior colliculus (SC), a key midbrain structure responsible for controlling saccades, is comprised of two subdivisions: the superficial layers (SCs), which respond predominantly to visual stimuli, and the intermediate layers (SCi), which can show both visual and saccade-related responses. Despite its importance, we still know little about how the SC selects saccade targets in realistic conditions. This study will address this critical gap in our knowledge. A major obstacle to progress in this area has been the complexity of analyzing neural responses under naturalistic conditions, and tackling this problem requires a model that can provide testable predictions of neural responses in naturalistic conditions. We will use a state-of-the-art neural model, MASC (Model of Attention in the Superior Colliculus), which incorporates constraints based on SC anatomy and physiology, and does a superior job of predicting saccade endpoints and scanpaths in a variety of search and free-viewing tasks. In conjunction with neural recordings, we will elucidate how superficial- and intermediate-layer SC neurons differ in their integration of activity related to salience, relevance, inhibitory tagging, and movement selection during multi-saccade visual search. In addition, we will test the contributions of the frontal eye field, a cortical area providing input to the SC, to these search-related signals in the SC.