search

open-ephys-plugins / neuropixels-pxi

Open Ephys GUI plugin for interfacing with PXIe-based Neuropixels hardware
GNU General Public License v3.0
13 stars 19 forks source link

channels ocassionally shuffle with no error message #22

Open kmcclain001 opened 2 years ago

kmcclain001 commented 2 years ago

While recording in openephys with neuropixel 2.0 I've noticed twice (still checking other recordings) that at some point during recording the spatial characteristics of the signal get scrambled and spikes appear on channels where they previously were not. I'm assuming the channels are some how getting shuffled. When I stop and restart the recording (without restarting software) the problem persists. Is this a known issue? Is there anything to do about it? IMG_1239

jsiegle commented 2 years ago

Can you provide some more details about what's shown in this picture? What's the difference between the green and pink channels?

Would it be possible to share one of the data files that has this issue so we can take a closer look at what's going on?

kmcclain001 commented 2 years ago

Hello, Thanks for the response.

The green and pink are just different shanks (in neuroscope). The issue you can see on the bottom channel of the pink set where after some event, high frequency spikes suddenly appear. Other channels have switched too, though not as evident in this image.

I tried using spikeGLX to see if that might solve the problem but I've now experienced the same issue there. I suppose this means it is an issue with the probes themselves.

In both cases blocks of channels move around after some instantaneous discontinuity. For example, the signal that was being recorded on channel 96 is suddenly being recorded on channel 47.

I am happy to share a data though I'm not sure what would be the best way.

grahamfindlay commented 2 years ago

FWIW, we have seen this channel shuffling several times on Neuropixel 1.0 probes for years, with both OpenEphys and SpikeGLX. Only a subset of channels are affected, never the whole probe IIRC. Usually if we stop and start the recording the problem fixes itself briefly, but quickly reappears. If left to record in the shuffled state, it eventually either fixes itself or the recording crashes (I forget the exact error). The issue was much more frequent when we were using slip rings to commutate the tether, but I think I have had it happen with fresh-out-of-the-box probes, tethers, and headstages. @cavelligonca might be able to tell you more.

kmcclain001 commented 2 years ago

Thanks @grahamfindlay that's useful to know. Have you tried to recover this data in any way?

grahamfindlay commented 2 years ago

We have not tried to permute the channel ordering to correct for the shuffling, no. But we have thought about it -- seems like it should be possible. We often had this shuffling happen in places with a distinctive LFP (e.g hippocampus) so it would probably be relatively easy to discern the permutation. I imagine there might be useful diagnostic information available in that pattern. It has just never been a priority.