Our brain is responsible for all our perceptions, thoughts, and actions. Despite the incredible array of processes the brain performs – from memory to emotion – its elementary units are the nerve cell and the synaptic junction. How is that a collection of neurons and their synapses gives rise to all of sensation, cognition, and action?
The goal of our laboratory is to reveal the neural basis of perception. More specifically, we want to understand exactly how patterns of neural activity in space and time drive all of our sensory experience. Despite incredible progress towards this goal in the last half century, we still have a minimal ability to interpret the underlying neural code. The primary approach neuroscientists have used to address this question is to correlate neural activity with specific stimuli or behavioral states. Yet deriving causal relationships between specific patterns of neural activity and specific perceptions has until recently remained impossible.
To overcome this major technical gap, our team develops novel optical brain machine interfaces that can read and write neural activity with cellular resolution, millisecond precision, and operate across large volumes of brain tissue simultaneously. We then leverage these new devices to write in specific patterns of activity in the brains of behaving animals to reveal the logic and the syntax of the neural code of perception. This same technology holds great therapeutic potential for treating a wide array of brain disorders.
Current Projects
There are three major aims of our research. First, we aim to understand exactly how neural circuits in the visual cortex enable complex visual processing. Humans and many other species effortlessly recognize, track and segment the visual world with blazing speed despite the fact that visual processing is among the most challenging computational goals achieved by any natural or artificial neural system. To address this challenge, we combine large-scale recording and targeted perturbation of the visual cortex in rodents and higher species to deeply understand the mechanisms of scene segmentation and object recognition at the synaptic, cellular, circuit, and behavioral levels.
Second, we are interested in visual learning and memory. How is it that we become so good at visual processing and object recognition? Humans learn from very limited experience to extract meaning from a nearly limitless set of sensory contexts. We also can remember and mentally reconstruct visual experiences for years and sometimes decades. Although our understanding of learning in the brain centers on synaptic plasticity, we have a minimal appreciation for how learning updates synaptic weights throughout the visual cortical network to store memory and fine tune object recognition. To address this goal, we are developing all-optical electrophysiological approaches that enable us to repeatedly map synaptic weights across thousands of neurons over the lifetime of individual animals. By combining these powerful optical technologies with computational modeling, we aim to reveal the learning algorithms of the visual system. These discoveries may also enable breakthroughs in treating circuit disorders of the nervous system by rebalancing synaptic weights that have gone awry.
Third, we are ultimately interested in understanding why and how processing sensory information gives rise to subjective experience. The neocortex is constantly generating and updating an incredibly detailed model of the external world. This internal model interacts with bottom-up sense data, attention systems, and prior experience to give rise to our remarkably stable, coherent, and self-consistent stream of experience. We can access this internal model experimentally through visual illusions and highly targeted neural perturbations. We combine our powerful optical technologies with carefully designed visual stimuli, behavioral psychophysics, and close-loop perturbations to unravel how the game engine of our mind generates our experience and our memories.