Jan Zimmermann

This group works on the intersection of neuroscience, physics, and engineering.

The primary goal of the Z-Lab is understanding the neuroscience of decision making in non-human primates. To this end, they employ a multitude of engineering / computational approaches that require high performance computing.

• One project employs the acqusition of high-density neural recordings (512 channels++) wireless in an open enclosure that is encased by 62 high-frame-rate cameras. From these acquisitions the researchers need to extract pose and feature information using a deep learning approach. Their current data rate is 1.6 Terabyte per hour of data acquisition.
• Another project records whole brain hemodynamic functional magnetic resonance images from non-human primates at 10.5 Tesla with ultra-high resolution. While standard analysis of these datasets (regression, linear models) is feasible on the lab's compute infrastructure, more complex questions such as factor decomposition, latent factor analysis, and nonlinear classification on these datasets require high performance computing.
• A final project involves modeling approaches of decisio-making by employing differential equation modeling that is related to neural and behavioral recordings. Evaluating the equation systems requires high performance computing to match the temporal resolution of the neural recordings.