search

open-ephys-plugins / neuropixels-pxi

Open Ephys GUI plugin for interfacing with PXIe-based Neuropixels hardware
GNU General Public License v3.0
14 stars 20 forks source link

Setting.xml seems not saving the shank with multishank recording #23

Closed MarineChap closed 2 years ago

MarineChap commented 2 years ago

Hello, I think we caught a bug when using a multishank probe in fklab lab. Or we did not use the plugin as intended ? We are on latest open-ephys / plugin version.

An external IMRO file is used to configure the channelmap with a multi-shank probe, then a recording is run. If I interpreted correctly the setting file, all shanks in the setting.xml are set to 0 while the recording is being done on different shanks. The display in the plugin and the recording is done correctly. The problem seems to be only in the saving part of the setting file. That causes some problems when using the spike interface to load the data. see this issue in probeinterface package.

Also, that would be super useful to have some documentation about how the setting.xml is constructed and how to retrieve the data. Thanks ! Have a good day

medengineer commented 2 years ago

Hi Marine,

I ran a recording using your attached IMRO file and the generated settings file is consistent with the IMRO config. I've attached it here.

Can you confirm which version of the Neuropixels plugin you are seeing this with?

Thanks!

MarineChap commented 2 years ago

Hello, thank you for checking on your side. Happy it works and I will use this setting to fix ours.

I keep you posted as soon as I heard a feedback from the doctorant working on it for the exact version. I am wondering what happened because i was super confident to be on the latest as I did a clean installation myself (removed everything/reinstalled everything from latest) two months ago. But clearly the problem is on our side.

Have a nice day.

MarineChap commented 2 years ago

open ephys version : 0.5.5.3 neuropix-pix version: 0.3.4-API7

jsiegle commented 2 years ago

Those versions should be fine. Can you try uploading the imro again and re-saving the settings.xml file?

MarineChap commented 2 years ago

I will ask the doctorant but she did 2 sessions and both are like this.

LiesDec commented 2 years ago

@jsiegle hi, I'm the PhD student @MarineChap talked about. After reinstalling the plugin I now indeed have the channel map saved in the settings.xml file.

However, I think there is still a small bug for the multistank probe. In the code (#L1293) you add 250 to the y coordinate. But that is not necessary. In the x coordinate that does make sense 'cause you jump shank.

jsiegle commented 2 years ago

You're right, that's a mistake...we'll get that fixed before the next release.

LiesDec commented 2 years ago

@jsiegle thank you for your reply. When will you do a new release? It would be really nice if that could be soon as I'm currently doing experiments using the plugin.

jsiegle commented 2 years ago

Hi @LiesDec...we've fixed this issue here, in the version of the plugin that will be released along with GUI v0.6.0, likely later this week.

We are not planning to release a patched version of the Neuropixels plugin targeting GUI version 0.5.5.4. To implement the fix, you'll have to change this line yourself and compile the plugin locally.