Cerebrospinal fluid (CSF) plays a vital role in maintaining brain health, clearing waste, and transporting nutrients. With the development of the fsfMRI by the Lewis Lab, researchers can visualize CSF flow through the 4th ventricle along both spatial and temporal dimensions. However, manually identifying and analyzing voxels displaying CSF inflow is time-consuming and labor-intensive. This project proposes an automated system for segmenting voxels displaying CSF inflow in the 4th ventricle using machine learning techniques. First, the system will automatically segment all the voxels corresponding to the 4th ventricle. Then, it will mark the regions containing CSF flow by analyzing patterns in the spatial progression of flow across three cross-sections within the ventricle. Finally, it will extract the temporal oscillation intensity in these CSF flow regions, allowing researchers to automatically visualize CSF flow patterns from fsfMRI scans.

This SuperUROP project allows me to apply my background in machine learning with my interests in neuroimaging and brain health. As a double major in AI and neuroscience, this will be my first opportunity to combine these two disciplines together. I hope to deepen my understanding of both fields and discover how my skill set can best contribute to the amazing research at the Lewis Lab.