[Question] Extracting Waveform Features

[lcui809]

Hi there, I just wanted to express my sincere gratitude for updating the tutorials; they have been incredibly helpful.

In the meantime, I have another question:

Would it be possible to extract waveform features or metrics for all spikes using only the sorted data? I ask because the current single-cell waveform features were extracted from the electrode with the largest amplitude. We're aiming to group the spikes by frequency and extract the waveform features of different subgroups within the same unit. If this is possible, could you please indicate which NPY file I should refer to for these metrics?

Thanks again for your help! I truly appreciate it.

[hornauerp]

Hi, I am not sure if I understand your question correctly, but it sounds to me like you want to cluster the different units in one recording by their waveform and their firing rate? Should this be the case, then you can look at the SingleCellPhenotype tutorial. If you specify the unit_features as ["ReferenceWaveform","ActivityFeatures"] and then run the louvain clustering, you should get what you described. You can then also use the assignClusterIdx function to permanently assign this cluster id to individual units and use that to calculate the single-cell features for each cluster in each recording separately, using useClustered = true.

I hope that helps!