E. J. Chichilnisky

The paper learns a neural response metric from retinal ganglion cell data to improve retinal prosthesis performance by better matching artificial and natural visual signals.

The paper develops a Gaussian Process model to cancel electrical stimulation artifacts and detect neural spikes in multi-electrode recordings, achieving low error rates and real-time compatibility.

A dense multi-electrode array spike sorter that triages collisions, clusters clean waveforms with a coreset Dirichlet process mixture, then deconvolves, giving more stable clusters and faster template learning than KiloSort.

Generalized linear models fit macaque parasol ganglion cells under white noise but fail on naturalistic stimuli, even when fitted on them, and added scalar nonlinearities do not rescue them.