Table 3. Currently available spike sorters in Spikeinterface.
TM = Template Matching; SL = Spike Localization; DB = Density-based clustering.
| Name | Method | Notes | Reference |
|---|---|---|---|
| Klusta | DB | Python-based, semi-automatic, designed for low channel count, dense probes. | Rossant et al., 2016 |
| Mountainsort4 | DB | Python-based, fully automatic, unique clustering method (isosplit), designed for low channel count, dense probes and tetrodes. | Chung et al., 2017 |
| Kilosort | TM | MATLAB-based, GPU support, semi-automated final curation. | Pachitariu et al., 2016 |
| Kilosort2 | TM | MATLAB-based, GPU support, semi-automated final curation, designed to correct for drift. | Pachitariu et al., 2018 |
| SpyKING Circus | TM | Python-based, fast and scalable with CPUs, designed to correct for drift. | Yger et al., 2018 |
| HerdingSpikes2 | DB + SL | Python-based, fast and scalable with CPUs, scales up to thousands of channels. | Hilgen et al., 2017 |
| Tridesclous | TM | Python-based, graphical user interface, GPU support, multi-platform | Garcia and Pouzat, 2015 |
| IronClust | DB + SL | MATLAB-based, GPU support, designed to correct for drift. | Jun et al., 2020 |
| Wave clus | TM | Matlab-based, fully automatic, designed for single electrodes and tetrodes, multi-platform. | Chaure et al., 2018 |
| HDsort | TM | Matlab-based, fast and scalable, designed for large-scale, dense arrays. | Diggelmann et al., 2018 |