spike-sorting

Automate spike sorting for high-channel-count extracellular recordings by loading data into SpikeInterface, preprocessing signals, and running Kilosort2, Kilosort4, or Mountainsort5 to generate single-unit spike trains reproducibly.

Quality metrics used to identify well-isolated units after spike sorting include SNR, ISI violations, presence ratio, amplitude cutoff, and firing rate, which filter low-quality units before exporting curated results to Phy or CSV.

You can run Kilosort2, Kilosort4, and Mountainsort5 within the same SpikeInterface pipeline to compare sorter performance reproducibly across multi-probe recordings before applying quality filtering and exporting results.

This spike sorting workflow supports exporting curated results to Phy for manual inspection and validation, allowing users to review well-isolated units filtered by quality metrics like SNR and ISI violations after automated sorting.

Before processing extracellular electrophysiology data with SpikeInterface, you need to install SpikeInterface and supporting numeric and plotting libraries to execute the load, preprocess, run sorter, compute metrics, and export pipeline.

Preprocess multi-probe electrophysiology recordings within SpikeInterface by loading raw multi-channel data, applying signal preprocessing steps, and then passing the processed signals to Kilosort or Mountainsort5 for automated spike sorting.