Fundamental detection and estimation limits of spike sorting

Spike sorting refers to the detection and classification of electric potentials (spikes) from multi-neuron recordings, a difficult but essential pre-processing step before neural data can be analyzed for information content. While several spike sorting algorithms have been proposed, our goal was to determine the ultimate limits of spike classification and to characterize this error, regardless of spike sorting algorithm. We account for the major factors influencing the sorting procedure: SNR, relative amplitude ratio and inter-spike correlation in both time and waveform morphology. Using an ideal detection/estimation system we calculated detection probabilities and time delay estimation errors as they vary with these parameters, establishing upper bounds on spike classification in terms of these metrics.
 

Authors: Mona Sheikh

Publications: Fundamental Detection and Estimation Limits in Spike Sorting (ICASSP '07), Favorable Recording Criteria for Spike Sorting (CNS '06), Fundamental Limits in Spike Sorting, M.S. Thesis

Related research at Rice