Biophysics-based generation of virtual spiking activity for evaluation of spike-sorting algorithms

New silicon-based multielectrodes comprising hundreds or more electrodecontacts offer unprecedented possibilities for simultaneous recordings of spike trains from thousands of neurons, that can only be realized if accurate,validated methods for automatic spike sorting are provided. This in returnrequires benchmarking data sets with known ground-truth spike times. Here, Iwill present a general simulation methodology for computing benchmarking datafor evaluation of spike-sorting algorithms, based on a well-establishedbiophysical forward-modeling scheme