LFP subsampling #2
Conversation
code/run_capsule.py
Outdated
| channel_ids_to_keep = channel_ids[0:len(channel_ids):SPATIAL_CHANNEL_SUBSAMPLING_FACTOR] | ||
| channel_ids_to_remove = [channel_id for channel_id in channel_ids if channel_id not in channel_ids_to_keep] | ||
| recording_lfp_spatial_subsampled = recording_lfp.remove_channels(channel_ids_to_remove) |
There was a problem hiding this comment.
You can also use recording_lfp_spatial_subsampled = recording_lfp.channel_slice(channel_ids_to_keep)
There was a problem hiding this comment.
Thanks @arjunsridhar12345
I added a few comments and suggestions
|
@arjunsridhar12345 do you mind if I push directly to your PR? |
Yeah sure |
code/run_capsule.py
Outdated
| recording_lfp = spre.common_reference( | ||
| recording_lfp, | ||
| reference="single", | ||
| groups=groups, | ||
| ref_channel_ids=reference_channel_ids, | ||
| ) |
There was a problem hiding this comment.
@arjunsridhar12345 not sure this is correct. If I understand correctly, you want to subtract the median of the channnels in agar (from surface_channel_index to the last channel) to the remaining channels, correct?
The current solution will subtract to each channel group (so each channel in your case), the corresponding reference_channel_id, so all channels out of the brain will be zeroed out, with no effect on the other channels. Is this the wanted behavior?
There was a problem hiding this comment.
yes, want to subtract the median of channels from surface index to last to the remaining channels. Then, should the reference be "global" with only ref_channel_ids set? Or what parameters should be passed to get the desired effect?
There was a problem hiding this comment.
Mmm I don't think we have that option yet, but we could cook in a workaround if needed.
See this issue: SpikeInterface/spikeinterface#2985
There was a problem hiding this comment.
ah ok cool. that seems close, basically just need to subtract the remaining traces from the median of those channels. is there any sort of set traces type functionality or another way you know of?
There was a problem hiding this comment.
Not really...I'll think about a smart way to do it! We will probbaly need some custom recording functions
There was a problem hiding this comment.
@arjunsridhar12345 fixed in this PR: SpikeInterface/spikeinterface#3139 (and SpikeInterface/spikeinterface#3140)
Until we release, I added a custom installation in the postinstall
There was a problem hiding this comment.
ok great, thanks!
LFP Subsampling ported from allensdk. Peforms spatial and temporal subsampling: