Event Details:
Tracking the coordinated activity of individual cellular events across volumes of intact tissue is a major challenge in biology that has inspired considerable technological innovation. However, synchronous, unbiased measurement of the high-speed activity of individual neurons across millimeter-scale volumes in behaving animals has not yet been possible. Noting that several proof-of-concept experiments have demonstrated that light field microscopy can be effective for 3D functional imaging in small, optically transparent neurobiological systems [1], we developed a computational imaging approach (SWIFT3D) that integrates high-dimensional, structured statistics with light field microscopy to allow the synchronous acquisition of single-neuron resolution activity throughout large intact tissue volumes at up to 100 Hz during behavior in mammalian tissue. This talk will highlight some of the results of applying SWIFT3D for volumetric in vivo imaging during behavior and for volumetric functional imaging of live human brain tissue from patients undergoing elective brain surgery. Notably, computation on large data with tools like Amazon Web Services and Apache Spark have been integral at every step of the process, from the reconstruction of volumes from light field images to nonlinear system identification on hundreds to thousands of neurons.