Human vision relies on rapid eye movements called saccades that occur a few times every second. These saccades bring peripherally identified objects to the fovea for high resolution inspection. This split-second sampling of the world defines the perception-action cycle of natural vision and profoundly impacts perception. While much is known about the neural mechanisms involved in the onset of this cycle, how targets are selected before saccades, virtually nothing is known about how pre-saccadic representations would impact processing of post-saccadic information about targets that now appear at the fovea. The proposed research addresses fundamental questions about how pre- and post-saccadic visual information are combined across saccades to maximize information for attended targets. Experiments will be performed in the common marmoset, a New World primate whose smooth cortex affords advantages for recording with high-density silicon arrays across cortical representations and in the longer term for imaging and optogenetic manipulation of large scale neural circuits.
In Aim 1, we will test if pre-saccadic information is enhanced in early visual cortex to favor higher acuity information that is most relevant to the fovea and thus could provide a ?foveal preview? to prime post- saccadic processing.
In Aim 2 we test if pre-saccadic enhancement is target specific and involves a selection of target features, such as orientation or motion.
In Aim 3 we record from foveal populations in early visual cortex to determine how pre-saccadic selection alters post- saccadic processing, and specifically, if pre-saccadic feature selection of the target biases processing to favor the continued selection of the target at the fovea based on increasing gain for its features. This research will provide the first evaluation at the level of single neurons and neural codes of post-saccadic foveal enhancement and the role of pre-saccadic target selection. This has broader impacts for understanding attention under natural viewing conditions where the eyes parse visual scenes, and more generally, for how sensory-motor predictions influence visual perception.
Neurological and neurodevelopmental disorders, such as autism and schizophrenia, have been suggested to involve deficits in perception related to how top-down cognitive processes, such as selective attention or expectations, modulate sensory processing. The proposed work will have a broad long-term impact on our understanding of these deficits by allowing us to examine how sensory expectations influence processing at the level of cortical circuits during active perception tasks. We will perform these experiments in a small New World primate, the common marmoset, a novel animal model which offers opportunities to study genetic models of human disease and to use advanced neurophysiological and molecular tools that were previously only available in rodent models.
