search

open-ephys / plugin-GUI

Software for processing, recording, and visualizing multichannel electrophysiology data
https://open-ephys.org/gui
GNU General Public License v3.0
193 stars 685 forks source link

LFP viewer 'Sort by depth' doesn't work for NP2 4-shank #533

Open LiesDec opened 1 year ago

LiesDec commented 1 year ago

Hi, When using the NP2 4-shanks the channels are not correctly sorted after applying the 'sort by depth' function in the LFP viewer. I would like to double check that this is indeed ONLY a display error and not something more fundamental and that my data is not recorded as I think. From my offline analyses I don't think so but I think it is good to be 100% sure :)

For simplicity I'm using on both probes the same channel map and a channel map that only selects electrodes on one shank. In the image below you can see the Imro file plotted with the associated channel numbers. image

However, when I apply the sort by depht function in the LFP viewer this is the sorting I get. LFP_sorting

As you can see this is not what you would expect from the imro file. Here is also the imro file I used: fourth_density_shank_0.txt

Together with @marinechap we've tried to look for the error in the high level display function but that seems to use correctly the y-pos. Can someone help us with finding the bug in the deeper layers of the code and double check it doesn't influence anything else?

jsiegle commented 1 year ago

Thanks for reporting this! The bug is on this line of NeuropixThread.cpp. The step in which the depth metadata was extracted assumed that the channels are ordered within a bank, which is not the case for Neuropixels 2.0 probes.

I fixed it by reading the selected electrode directly from the probe settings. The add-geometries branch includes this fix, and also updates the signal chain after loading an IMRO file, so the new depths will propagate to the LFP Viewer.

The good news is that the actual channel configuration (and channel metadata in the settings files) was correct; the problem was just with the depth info used by the LFP Viewer.

If you can double check it to be sure the behavior is correct, it would be much appreciated! After that we will merge this into main and put out a new release ASAP.

MarineChap commented 1 year ago

Thanks @jsiegle ! I installed your fix on our computer this morning. We will give you a feedback soon.

LiesDec commented 1 year ago

Indeed thanks @jsiegle for the (quick!!) fix, this works! One question, how are you currently handling the sort by depth visualisation when you have channels on multiple shanks? (so I can see if that checks out with what I see :) ). It would be cool in the future to be able to see the 4 shanks next to each other, or been able to pick only channels on one shank :)

jsiegle commented 1 year ago

When sorting by depth, the shanks are displayed consecutively (all channels on shank 1 first, then all channels on shank 2, etc.)

Currently, you can use a vertical split in the LFP Viewer to view up to 3 shanks side-by-side. We are brainstorming better ways to visualize multi-shank data—let us know if you have any suggestions!