CRCNS: Computations and Neural Mechanisms in Sustained Perception in Humans

Studies of human perception mostly focus on neural responses to short stimuli and specifically on their initial recognition. How the brain supports sustained perception of longer stimuli is poorly understood. The fundamental novelty of our project is in addressing this neglected segment of human perceptual experience. We recently discovered that distributed neural representations of stimulus category persist for the duration of the stimuli in ventral temporal cortex, whereas in fronto-parietal cortex the stimulus category could only be decoded shortly after stimulus onset. These findings motivate our Specific Aims: 1) How does attention and awareness affect persistent representations of visual stimuli? 2) What is the structure and information content of field potentials and single neuron spikes in individual brain regions and how do they interact during sustained perceptual experience? 3) What are the computational principles and mechanisms underlying sustained perception? The project is embedded in the Pis long term goals of understanding the electrophysiological and computational basis of perception, attention and awareness. We will use advanced computational methods and unique intracranial data of local field potentials (LFP) and single units (SU) from human patients, combining novel experimental paradigms, data analysis and decoding approaches, and modeling and analytic frameworks. The project will provide training to graduate and undergraduate students in interdisciplinary, collaborative research on fundamental issues in neural computations and cognition, with cross-site visits to develop new approaches and discuss results.