Spatial Transformations in Frontal Cortex During Memory-Guided Head-Unrestrained Gaze Shifts

We constantly orient our line of sight (i.e., gaze) to external objects in our environment. One of the central questions in sensorimotor neuroscience concerns how visual input (registered on retina) is transformed into appropriate signals that drive gaze shift, comprised of coordinated movement of the eyes and the head. In this dissertation I investigated the function of a node in the frontal cortex, known as the frontal eye field (FEF) by investigating the spatial transformations that occur within this structure. FEF is implicated as a key node in gaze control and part of the working memory network. I recorded the activity of single FEF neurons in head-unrestrained monkeys as they performed a simple memory-guided gaze task which required delayed gaze shifts (by a few hundred milliseconds) towards remembered visual stimuli. By utilizing an elaborate analysis method which fits spatial models to neuronal response fields, I identified the spatial code embedded in neuronal activity related to vision (visual response), memory (delay response), and gaze shift (movement response). First (Chapter 2), spatial transformations that occur within the FEF were identified by comparing spatial codes in visual and movement responses. I showed eye-centered dominance in both neuronal responses (and excluded head- and space-centered coding); however, whereas the visual response encoded target position, the movement response encoded the position of the imminent gaze shift (and not its independent eye and head components), and this was observed even within single neurons. In Chapter 3, I characterized the time-course for this target-to-gaze transition by identifying the spatial code during the intervening delay period. The results from this study highlighted two major transitions within the FEF: a gradual transition during the visual-delay-movement extent of delay-responsive neurons, followed by a discrete transition between delay-responsive neurons and pre-saccadic neurons that exclusively fire around the time of gaze movement. These results show that the FEF is involved in memory-based transformations in gaze control; but instead of encoding specific movement parameters (eye and head) it encodes the desired gaze endpoint. The representations of the movement goal are subject to noise and this noise accumulates at different stages related to different mechanisms.