Spike Width

[ffredes]

Does anybody has any experience what spike width distribution you get from sp. temps file using neuropixel probe and kilosort3? If I estimate spike width (spike width at half amplitude) from the average waveforms stored in sp.temps file I do not get a typical separation of narrow and wide spiking units. The distribution is rather uniform instead of bimodal. I know that the sorted units are all coming from the cortex, so the data should be comparable to the published report.

[ffredes]

Does Kilosort3 store a limited number of spike waveforms for each unit? I am trying to recalculate the spike waveform features, but do not want to extract spike waveforms from the raw continuous data. Since phy plots spike waveforms for many spikes from a given unit, I thought some matlab or python variable stores these waveforms.