OpenSorter is a stand-alone package for sorting neural spike data and is the latest addition to TDT’s OpenEx software suite. OpenSorter offers a number of powerful sorting methods including Bayesian expectation-maximization, k-means, and closest-centers algorithms in addition to manual cluster cutting and waveform selection.
Capabilities include manual, semi-automated, and fully automated processing. Spikes can be sorted as individual channels, batched for fast processing of groups of datasets, or combined across blocks or channels and sorted as a pooled superset.
Data is displayed and can be sorted in principal components feature space, waveform parameter feature space, an event timeline, or piled waveform shape windows. Sorting results can be edited and spikes re-assigned using feature or waveform space editors. Post-hoc analyses including Pseudo-F, isolation distances, L-ratios, and silhouette indices indicate sort quality and help to guide manual resorting and editing.