The ability of animals to use “intuitive physics” to understand 3D scenes is a hallmark of general biological intelligence. To study this cognitive process, we train rats and mice to play games (like fetch) and challenge them to use intuitive physics to play these games efficiently. To measure and model how rodents learn and use this knowledge, we combine high-resolution computer vision systems, robotics, biomechanics, and machine learning. We take a “behavior-first” approach to neuroscience, but we’ll soon aim to combine high-density electrophysiological recordings in freely moving rats and mice to study how internal models in the rodent brain generate representations needed for physics-based reasoning.
Lab Website