Home Department: Art Practice and Mathematics
Mentor: E.J. Chichilnisky (Neurosurgery and Ophthalmology)

“Algorithm Development for Distinguishing Neural Firing Consistency in Clustered Response Curves for Developing an Artificial Retina"

Seiji will advance the development of artificial retinas for vision restoration by improving analysis of neural activity after electrical stimulation. Neural activity resulting from repeated stimulation trials are clustered into spiking and non-spiking waveforms. Current algorithms rely on having two unique response clusters to calculate the spiking response probability by eliminating electrical stimulation artifacts. However, when spiking response probabilities are nearly 0 or 1—resulting in only one response waveform cluster—current algorithms malfunction. Seiji will develop an algorithm identifying whether single waveform clusters correspond to spiking probability 0 or 1, enabling rapid calibration of single-neuron responses to electrical stimulation.