Error: peak_channel of the clsuters is the reference channel

[OlyOlympic]

Hi, I have encountered an issue while working with the data in metrics.csv. Specifically, I noticed that there are clusters with peak_channel = 191. However, when using ecephys_spike_sorting/scripts/create_input_json.py, we set reference_channel to 191. Based on this, I would expect no clusters to have a peak_channel of 191.

Could you please advise if I might have made an error in my setup or processing? This situation is causing an additional problem because there is no position corresponding to 191 in channel_map or channel_position, which prevents me from determining the position of those neurons.

Additionally, I would like to ask: If reference_channel = [191], how should I configure the IMRO table or any other relevant parameters to ensure compatibility with Neuropixels during recording?

Thank you for your guidance, and I look forward to your response.

Best regards,
Oly

[jsiegle]

If you look at the waveforms for units with peak channel at 191, do they look like real action potentials, or are they noise? If they are noise, then they can be safely ignored – it's unlikely that real waveforms would have a peak on the reference channel, since there's no signal there.