Big neuro + movement

This is research project collaborated between Rice University and Baylor College of Medicine. This study consists of exploring the correlations between mice movements and their visual responses using data collected on the movements and the neural activities in the primary visual cortex of mice. A video is taken of the mouse as it is fixed on a treadmill and exposed to various visual stimuli from Hollywood blockbusters to generated patterns. Using DeepLabCut, an efficient method for 3D markerless pose estimation based on transfer learning with deep neural networks, the mouse body parts and pupil information were tracked. Patterns are then explored between the movement data and the mice neuron data.

The data used has not been published and multiple teammates have contributed to this projects. Therefore, instead of source code, a poster link is provided. The project gained wide recognition and won the inaugural D2K Learning Lab Showcase at Rice.