Data Naming and Storing

Data Naming and Storing: MEG

Data host

The MEG data is securely stored on NYU BOX, access is given through invitations. The Data folder is structured in the BIDS standardized format. Please raise an issue on github repository if you think the structure does not conform to BIDS.

MEG data is stored on NYU Box (invitation only). If you are unable to access the datasets it means you do not have the permission. Kindly send an email to nyuad.meg@nyu.edu

Data naming BIDS protocol

A typical MEG dataset from NYUAD contains the following files and folders in BIDS-specification https://bids-specification.readthedocs.io/ :

Your data should be stored in the BIDS format. Find a template to be used under bids-template, the template contains all the information that you should have in each file.

In the following, [SUB_ID] should be replaced with the ID of the subject for naming purposes. The different data files generated from a MEG experiment are the following.

Note

If you have found a bug in the BIDS naming format please raise an issue on github or create a pull request with your proposed modifications.

Directory structuring

Your file structure should look like this:

  • root level directory named after your project ‘PROJECT_NAME’

  • within the root directory, you should have:
    • README.md: general information on your project

    • dataset_description.json: a JSON object detailing authors, project name, grants and so on

    • sub-[SUB_ID]: a folder for each subject, within which there should be:
      • Optionally a ses-[SES_ID] where SES_ID is a numeric session ID, inside which you will find:
        • meg (must have one): will contain your MEG data and other meta-data

        • sourcedata (not necessary)

      • If the ses-[SES_ID] folder is not there, it means you do not have multiple sessions for the subject, in that case you will directly find meg and sourcedata within the sub-[SUB_ID] folder

Additionally, it can have:
  • events.json should have meta-data about the device used for presenting Stimulus

  • events.tsv should have information about the types of trials, onset and duration

  • task-[PROJECT_NAME]_meg.json should have information about the MEG system used, sampling frequency and so on

If the three above files are not found at the root of your dataset, it means that each subject or session might have a specific configuration, in that case make sure to have these files within the subject or session, while adding the sub/ses ID’s to the name of the file. The three above files can be found at the root of your dataset, while having the same files within each session/subject under the following condition: - when some parameters are specified at the level of the dataset, while other are specific for the session/subject

Note

General BIDS notes: - If a particular file applies to all runs, then it should not have the run descriptor in its name - If a particular file applies to all sessions, then it should not have the ses descriptor in its name

Filename Entities and Ordering

The BIDS standard enforces a strict ordering of components (entities) in filenames. The following table shows the order in which entities must appear if they are present. Entities marked as Mandatory must always be included.

Entity

Status

Description

sub-<label>

Mandatory

Subject ID. Unique identifier for the subject.

ses-<label>

Optional

Session ID. Required if the subject has multiple sessions.

task-<label>

Mandatory

Task name (required for functional data).

acq-<label>

Optional

Acquisition parameter (e.g., acq-laserproject).

run-<index>

Optional

Run index. Used if multiple runs of the same task are acquired.

mod-<label>

Optional

Modality.

split-<index>

Optional

Split index. Specific to .con split files.

proc-<label>

Optional

Processing label (e.g., proc-CALMnoisereduction).

space-<label>

Optional

Coordinate space (e.g., space-ALS for marker files).

Suffix

Mandatory

The modality or file type suffix (e.g., _meg, _markers, _headshape).

Extension

Mandatory

The file extension (e.g., .con, .mrk, .fif, .json).

Warning

The order of these entities is fixed. For example, sub-001_task-rest_ses-01_meg.con is INVALID because ses must come before task. The correct name is sub-001_ses-01_task-rest_meg.con.

MEG-Laserscan files

  1. A .fsn filename that should be named sub-\[SUB_ID\]_acq-laserproject_headshape.fsnThis file is obtained by saving the whole fastscan laser project (File Save)
    • The .fsn file is stored under sourcedata

  2. Two .txt files holding the head surface scan and stylus points should be named:
    • sub-001_ses-01_acq-head_headshape.txt is the head scan of the participant

    • sub-001_ses-01_acq-points_headshape.txt is the stylus points location file of the participant
      • all should be stored under meg

MEG-KIT files

Depending on the experiment, many .con files can be produced by the KIT machine.

  1. .con files are named:
    • sub-[SUB_ID]_ses-[SES_ID]_task-[TASK_NAME]_run-[RUN_ID]_split-[SPLIT_INDEX]_meg.con

    • example: sub-001_ses-01_task-audiovisualmotor_run-03_meg.con

    • if you applied the CALM filtering or other kind of pre-processing to your con file then:
      • sub-[SUB_ID]_task-[TASK_NAME]_proc-CALMnoisereduction_meg.con or replace the proc value with the name of your filter

      • Example: sub-001_task-attention_proc-CALMnoisereduction_meg.con

    • the split label indicates when, in a single session, the scan results in multiple acquisition files, use this label to indicate the order of the scan

    • all .con files should be stored under meg

  2. .mrk files are named:
    • sub-[SUB_ID]_ses-[SES_ID]_task-rest_acq-[MARKER_ORDER_NUMBER]_space-ALS_markers.mrk where:
      • [MARKER_ORDER_NUMBER] is the chronological order in which the marker has been acquired at in that session

      • The task should be set to rest because there is no stimuli events involved (if not we get warning from the BIDS-Validator)

      • The space refers to the coordinate system type, for Yokogawa/KIT it is Anterior on X-axis, Left on Y-axis, Superior on Z-axis, therfore it is ALS

    • all .con files should be stored under meg

Note

  • The ses is optional, but must be present if the same subject had multiple sessions

  • The run is optional, but must be present if in the same session, a subject had multiple MEG acquisitions leading to multiple .con files

MEG-OPM files

The OPM (FieldLine) system writes .fif files with its own native names (for example 20240415_113034_sub-NA_file-emptyRoom-raw.fif). These must be renamed to a BIDS-valid form before storage: drop the -raw suffix and add a task label.

  1. The .fif files are named: * sub-[SUB_ID]_ses-[SES_ID]_task-[TASK_NAME]_meg.fif * example: sub-001_ses-01_task-restingstate_meg.fif * all .fif files and their channels.tsv and meg.json sidecars are stored under meg

Empty-room (noise) recordings

Empty-room recordings (noise recordings) are acquired with no participant in the magnetically shielded room, to characterize system noise and to estimate noise covariance. BIDS has a dedicated convention for them, which NYUAD follows.

Store every empty-room recording under a single reserved subject sub-emptyroom, with a session label equal to the recording date (ses-YYYYMMDD) and the task noise:

sub-emptyroom/
    ses-<YYYYMMDD>/
        sub-emptyroom_ses-<YYYYMMDD>_scans.tsv
        meg/
            sub-emptyroom_ses-<YYYYMMDD>_task-noise_acq-<kit|opm>_run-<NN>_meg.<con|fif>
            sub-emptyroom_ses-<YYYYMMDD>_task-noise_acq-<kit|opm>_run-<NN>_channels.tsv
            sub-emptyroom_ses-<YYYYMMDD>_task-noise_acq-<kit|opm>_run-<NN>_meg.json

Conventions:

  1. sub-emptyroom (used literally) is the only subject for noise data. Do not assign a numeric subject ID.

  2. ses-<YYYYMMDD> is the date the empty room was recorded, for example ses-20240628.

  3. task-noise is always the task for empty-room data.

  4. acq-kit marks the KIT/Yokogawa SQUID system and acq-opm marks the FieldLine OPM system. This is how both systems coexist in one dataset, because BIDS has no dedicated system entity.

  5. run-<NN> numbers multiple recordings taken on the same day with the same system (run-01, run-02, and so on).

  6. proc-CALMnoisereduction is appended when an on-system noise-reduced copy of a run is stored next to the raw run.

  7. Conditions that must not go in the file name (for example “helium heater on”, “OPM present or removed”, or the original file name) are recorded in the comments and original_filename columns of the session scans.tsv, documented by a sub-emptyroom_scans.json data dictionary.

Examples:

sub-emptyroom_ses-20240628_task-noise_acq-kit_run-03_meg.con
sub-emptyroom_ses-20240415_task-noise_acq-opm_run-01_meg.fif
sub-emptyroom_ses-20241023_task-noise_acq-kit_run-01_proc-CALMnoisereduction_meg.con

Note

Empty-room recordings contain no participant data, so no demographic fields are required beyond the reserved sub-emptyroom entry in participants.tsv.

Note

For the KIT system the channels.tsv typing depends on the recording date: from 2018 onward the sensor mapped to MEG 092 is recorded as a reference magnetometer (a known hardware fault), giving 207 MEGMAG plus 17 MEGREFMAG; recordings before 2018 have 208 MEGMAG plus 16 MEGREFMAG.

Note

This layout is the one produced for the NYUAD empty-room dataset on NYU Box. See the BIDS empty-room specification at https://bids-specification.readthedocs.io/en/stable/modality-specific-files/magnetoencephalography.html.

Data uploading

Steps

  1. Access the folder of NYU Box

  2. Identify the project that the data belongs to, or request for a new project folder

  3. Access the folder for that project or create a new folder if non-existent

  4. Follow the BIDS structure format described above, within each project file,

  5. Make sure that all files have been uploaded to the folder

Data Naming and Storing: MRI

MRI data is hosted on XNAT