search

int-brain-lab / iblrig

Main repository for IBL rig code
MIT License
21 stars 20 forks source link

Seattle rig consistently failing QC #656

Closed k1o0 closed 1 month ago

k1o0 commented 2 months ago

Sessions have QC outcomes such as this. The frame2ttl has been calibration and the issue lies with the Bonsai workflow. This needs to be investigated (separate issue) and the extraction should be modified to ignore consistent sync square changes before stim on (see #654).

There is also an issue with the Harp sound card (see https://github.com/int-brain-lab/iblrig/issues/655). It seems particularly bad on this rig so we should investigate if the TTLs are indeed failing in the same manner, or whether it is a wiring issue. Checking Bpod and DAQ TTLs in the QC viewer will tell us more information.

oliche commented 2 months ago

The QC Viewer can't be run on NP2.4 This is due to the dynamic pipeline creation that attempts to recreate a pipeline after the splitting of the original probes in 4 shanks.

Here is a quick fix to address this at the QC stage, which is not desirable @mayofaulkner @k1o0
https://github.com/int-brain-lab/ibllib/compare/taskqc?expand=1

oliche commented 1 month ago

This issue underwent 2 phases of mitosis and is now split in 4 issues

  1. 655 Harp

  2. int-brain-lab int-brain-lab/ibllib#768 task qc with np2.4
  3. ephysCW v8 extracts with the wrong sequence #658
  4. frame2TTL extraction and unexpected flicker before stimOn (probably due to ITI definition?) #654