Introduction & research topics

Animals navigate complex environments with remarkable adaptability, generalizing to new situations without requiring extensive learning. The Neural Computation in Vision Group investigates how the brain transforms sensory information into robust neural representations that enable this flexible behavior.

We combine large-scale neurophysiological recordings with advanced machine learning to create 'digital twin' models of visual brain regions. Using the mouse visual system as our primary model, we iterate between detailed computational analysis and targeted experiments to understand neural processing at both cellular and population levels, from the retina through the visual cortex. Our high-throughput approaches for assessing visual function and behavior in mice accelerate both basic research and the validation of potential treatments for visual disorders.

Through our collaboration with Stanford University, we systematically compare mouse and primate visual systems to identify universal principles of neural computation. This comparative approach reveals both conserved and species-specific features of vision, helping to bridge the gap between animal models and human applications.

Our research program integrates experimental neuroscience with machine learning methods, creating a powerful framework for understanding how the brain builds robust sensory representations and how these representations change in disease states.