coins-meg README

COINS-MEG analysis repo.

Folder Structure

See .gitignore for more details of which files are ignored in commits.

Scripts in the repo should write to /ohba/pi/lhunt/datasets/coins-meg_data/derivatives. (Note that any references to .../coins-meg_meg-analysis/derivatives should be removed, as this was before the repository and dataset were restructured.

• analysis contains:

  • 1_maxfilter-bulk.py which maxfilters the data.
  • 2_preprocess-bulk.py which bulk-preprocesses multiple runs/multiple subjects using osl's batch preprocess tool
    • osl batch preprocess function also outputs the config.yaml (preprocess settings) and fnames.txt (paths of all files to be preprocessed).
    • overview: bandpass filtered from 0.25-40Hz, downsampled to 250Hz, labelling of bad segments/bad channels, ICA
  • 3_coregister-and-source-recon-bulk.py which runs co-regristration, and LCMV beamforming to a parcellation in MNI space
  • preprocess.py - can ignore; an earlier manual version of the preprocessing script, which implements each preprocessing step with MNE Python.
  • bulkprepro-prelim-analysis.ipynb - preliminary data visualisation and ERF plotting
    • bulkprepro-prelim-analysis_bulk.py - does the same thing, but bulk-runs it for specificed participants/runs, and suppresses plot pop-ups while saving them inside preprocessed/sub-{subj}/run-{run}/meg/auto-max/plots
  • glm-prelim.ipynb does some very preliminary experimenting with GLM-spectrum.
  • A virtual environment folder venv which should be ignored as it will not work across computers.
  • a number of python files begining cf_trf... written by @cedricfoucault for temporal response function esimtation

• docs contains:

  • Four data-related docs.
    • coinsmeg_counterbalancing.xlsx: information about counterbalancing of key/button-action mapping, source-volability mapping, and blocks orders within runs ('sessions').
      • E.g., for sub-22, a/s keys (corresponding to buttons 1 & 2 in the MEG scanner) controlled clockwise/counterclockwise movement (respectively) of the shield, and k/l keys (buttons 3 & 4 in the MEG scanner) controlled decreasing/increasing shield size respectively.
    • coinsmeg_if-data-exist.csv: does this data exist at all? Headers contain relevant subdirectory paths.
      • If 'y', the data corresponding to this subdirectory exists even if it may not yet be in the directory.
      • If 'no', the data does not exist and we most likely cannot obtain it (i.e., for structurals, participant has left the country; no response despite multiple follow-ups).
      • If 'tbc', we have been in touch with people who say they have this data, but are still waiting on them to share it.
    • coinsmeg_if-data-in-directory.csv: is this data in the relevant subdirectory? ('y' = yes; 'n' = no)
    • coinsmeg_participant_notes.xlsx: any notes relating to this participant (e.g., any issues, truncated training due to participant lateness, internal number of the MEG scan denoted by #XXX).
  • Documentation that is not data-related is not yet in this repo. This includes participant screener responses, a password-protected linking document (linking participant names to IDs), and the study protocol (coinsmeg_protocol.md).

• experiment contains:

  • two folders containing PsychoPy code to run the experiment.
    • ses-1-training contains the task as set-up for initial behavioural training.
    • ses-2-meg contains the task as set-up for in-person testing in the MEG scanner at the OHBA centre.
  • note that both of these folders contain stimgen subdirectories, with scripts written in MATLAB that control the generative processes for the experimental design.
  • (note also that the readme.md files contained within these two subdirectories is currently out-of date. For a clean, up-to-date version of the experimental design repository, visit https://github.com/lilianAweber/cogpsy_laser_task).

Current status

• A first preprocessing and ERF pipeline has been built for sub-12.
• All subjects/runs have been maxfiltered.

Participants with incomplete data

• The following subjects have existing anatomical scans, but we are still waiting on copies: 5, 6, 8, 22
• The following subjects do not have existing anatomical scans: 1, 2, 3

Trigger information

Below are the trigger descriptors/labels, corresponding trigger values, and corresponding trigger meanings.

• expStart = 100; start of the experiment

• expEnd = 105; end of experiment

• lastFrame = 99; last frame of experiment

• practiceMove = 101; practiced moving shield

• practiceSize = 102; practiced changing shield size

• blockStart = 10; start of block (there were 4 blocks per run)

• blockEnd = 20; end of block

• laserHit = 1; a laser had hit the shield (i.e., the laser was successfully caught)

• laserMiss = 2; a laser had missed the shield

• keyRight = 3; key pressed to move shield to the right

• keyLeft = 4; key pressed to move shield to the left

• keyUp = 5; key pressed to make shield size larger

• keyDown = 6; key pressed to make shield size smaller

• keyRelease = 7; ?