Recommended pre processing

To maximize the quality of your data, we suggest:

• minimizing the noise and artefacts on your data. To do this, we recommend recording with either both a ground wire and reference wire going from the probe to the mouse, or with a ground wire and the internal reference (if using probes other than the 3A - the internal reference on these doesn't work well). To remove artefacts from your recorded data, either temporally align your channels with each other and common-average reference your data with Bill Karsh's function CatGT, before feeding this data into kilosort or use pyKilosort, where this is implemented.

• fine-tuning kilosort's spike detection parameters. We recommend looking at your raw data and spikes detected by kilosort, to assess whether most spikes are being detected. If not, consider lowering kilosort's detection thresholds.