Worked in group to develop a simulation and decoding software to be utilized in the next generation of bioluminescence brain machine interface, with the ultimate goal of addressing large ensembles of individual neurons with high specificity and temporal resolution. The approach was to use a combination of smoothing filters to maximize the SNR, k-means with Individual Component (ICA) to cluster firing to neurons, and a matching filter to recreate the fluorescence firing pattern in real time.
