Featuring Mark J. Schnitzer, Ph.D., Biology and Applied Physics Professor, Howard Hughes Medical Institute, Stanford University.

Optical techniques have become central to research at the forefront of brain science and are rapidly increasing in their breadth and importance. I will present recent advances in optical brain imaging, which have allowed the visualization of large-scale neural coding in behaving animals and optical readouts of neuronal voltage oscillations. Neuroscientists use these innovations to study information processing in healthy and diseased brain states. As an in-depth example, I will describe imaging experiments that address questions first raised by John von Neumann about how the brain can compute so accurately even though individual neurons appear to be extremely noisy. By using custom optical mesoscopes to image neuronal dynamics across the mouse visual cortex, we found that neocortex supports reliable sensory performance through brief elevations in sensory coding redundancy near the start of perception, neural population codes that are robust to cellular variability, and widespread inter-area fluctuation modes that transmit sensory data and task responses in non-interfering channels. These measurements reveal mesoscale dynamics of cortical visual processing and provide constraints for the design of future brain-machine interfaces.

More broadly, an upcoming generation of optical instruments is poised to lay the basis for rich interactions between experimental neuroscience, machine learning, and human health.