We want to understand the brain, as much of it as possible, all at once. We do this in three ways.
First, we develop technologies for the analysis and visualization of data, and the sharing of data and computation. We leverage recent advances in distributed computing and web technologies, and all of our development is 100% open source. As examples of our projects, check out thunder (distributed analytics), lightning (interactive visualization), and binder (reproducible computing).
Second, we uncover principles of neural coding by designing experiments and analyzing data. We work closely with collaborators across many systems — mouse, zebrafish, fly. Our focus is holistic neural computation: how do sensory events, behavioral states, and past experiences guide behaviors, and how can we watch the underlying computations evolve both within and across brains areas.
Third, we are developing new ways to "close the loop": to analyze neural data online during experiments, and use the results to change an experiment on the fly, or perform targeted manipulations of neural activity. This brings together several exciting challenges — data stream processing, online machine learning, real-time visualization, and closed-loop experimental design — and promises new windows into neural computation.
