Matthew Kaufman’s research focuses on the intersection of systems and computational neuroscience, with the goal of understanding how neurons work together to implement computations relevant to motor control and decision-making. His background includes extensive experimental work in multiple model systems (mice, rats and monkeys), which has allowed him to gain experience with the various strengths of those scientific subcultures: innovation in experimental techniques, careful behavior, sophisticated analysis and high-throughput science. By following this path he has become fluent in a variety of experimental methods and gained substantial experience designing, implementing and debugging novel analytical techniques. This combination has proved successful in a series of publications collecting data and applying various sophisticated models in each of the three systems. Most notably, these works have provided insight into the dynamical generation of motor control and the extent to which representations for decision-making and movement are intertwined.
In Kaufman’s laboratory, his team acquires data using two-photon calcium imaging of motor and decision-related cortical areas in the mouse while these animals perform sophisticated behaviors with high-quality movement tracking. From these data and in multiple collaborations, they develop state-of-the-art methods for understanding neural activity in terms of decoders, dynamical systems and other forms of state-space analysis. They then aim to tie these abstract levels of computation to their biological substrates to understand how cell types, layers, areas and connectivity interact to implement effective controllers.