Neuron Identification and Classification

Neuroscientists often study the neuron function of fruit flies (drosophila) as a simplified substitute for more complicated brains (such as the human brain). There are many established tools for studying drosophila neural circuits, but most of the approaches involve studying one cell type at a time. Our approach captures multiple cell types, allowing us to study more of the network at once including any potential interactions.

Our project, a collaboration between a Stanford neuroscientist and University of Virginia Data Science utilizes the CaImAn python library from the Flatiron Institute to isolate neurons in recorded video. We then create a classification model to identify the neuron type, using shape of the neuron, reaction to stimulus, calcium deconvolution and other output from the CaImAn model.

This project is still in progress.