BIDS Manager

From raw recordings to a curated BIDS dataset

BIDS Manager reads your raw MRI, PET, MR spectroscopy, EEG, MEG, physiological and eye-tracking recordings, names each one by the BIDS standard, helps you describe what the raw files don't capture, and stays with you after the conversion, in its GUI or from the command line.Reads your raw recordings, names each one by the BIDS standard, helps you describe them, and stays with you after the conversion.

Install BIDS Manager Take the course

Free and open source, for Windows, macOS and Linux.

BIDS Manager is curation software for brain imaging and electrophysiology data. It reads a folder of raw recordings, shows you what it found and the name the standard gives each one, lets you correct and describe them, converts them with the established tools, and then helps you validate, view, check and edit the dataset that results. You do not need to write code.

The result

Every file named, every sidecar described

The course's sample after it was converted and described: each recording under its BIDS name, with the JSON sidecar beside it. In a sidecar, every value says where it came from: read from the recording, written by BIDS Manager, your answer, or a placeholder still to answer.

multimodal_study/
├── dataset_description.json  participants.tsv  README
└── sub-001/ses-01/
    ├── anat/ sub-001_ses-01_acq-tfl3p2_T1w.nii.gz  .json
    ├── func/ sub-001_ses-01_task-rest_bold.nii.gz  .json
    │         sub-001_ses-01_task-rest_sbref.nii.gz  .json
    │         sub-001_ses-01_task-rest_recording-ecg_physio.tsv.gz  .json
    ├── fmap/ sub-001_ses-01_acq-fm2_phasediff.nii.gz  .json
    ├── dwi/  sub-001_ses-01_acq-epb0p2_dir-AP_dwi.nii.gz  .bval .bvec  .json
    ├── pet/  sub-001_ses-01_pet.nii.gz  .json
    ├── mrs/  sub-001_ses-01_svs.nii.gz  .json
    ├── eeg/  sub-001_ses-01_task-rest_eeg.edf  _channels.tsv _events.tsv  .json
    ├── meg/  sub-001_ses-01_task-rest_meg.fif  _channels.tsv  .json
    └── beh/  sub-001_ses-01_task-rest_recording-eye1_physio.tsv.gz  .json

sub-001_ses-01_task-rest_bold.json

{
"TaskName": "rest"written by BIDS Manager
"RepetitionTime": 0.85read from the recording
"EchoTime": 0.03read from the recording
"PhaseEncodingDirection": "j-"read from the recording
"B0FieldSource": "fmap012"written by BIDS Manager
"TaskDescription": "Resting state with the eyes open, looking at a fixation cross."your answer
"Instructions": "Keep your eyes on the cross and let your mind wander."your answer
"CogAtlasID": "TODO"to answer
}

How it works

Four stages, each a part of the GUI

You move through the same four stages for every study. Select a stage to see where it happens in BIDS Manager.

In the marginEvery figure on this site is a picture of BIDS Manager itself, working on the sample datasets you can download.

  1. 1

    Scan the raw data

    Point BIDS Manager at a folder of DICOM series, EEG, MEG and eye-tracking files. It reads every header, groups the files by participant and session, and proposes a BIDS name for each recording.

    Scanning raw data
  2. 2

    Curate in the table

    Each recording is a row. Correct a subject, a session or a task, leave a series out, and compare subjects in the Overview before anything is written.

    The scan table
  3. 3

    Describe and convert

    Give the metadata the raw files don't capture, see the dataset in the BIDS preview, then convert: images, recordings and tables under their BIDS names, with the links between fieldmaps and the runs they correct.

    Metadata you provide
  4. 4

    Check and edit in the Editor

    Validate the dataset against the BIDS standard, go to each finding, edit sidecars and tables, view images and signals, and run a fast quality check.

    Validating in the Editor
The Converter after scanning the multimodal sample: the raw folder on the left, the structure the scan found, the table of recordings with their proposed BIDS names, and the properties of the T1w row.The Converter after scanning the multimodal sample: the raw folder on the left, the structure the scan found, the table of recordings with their proposed BIDS names, and the properties of the T1w row.
The Converter after curation: the Recordings view lists each kind of recording, and the Overview counts one subject, one session and fifteen recordings, with nothing worth a look.The Converter after curation: the Recordings view lists each kind of recording, and the Overview counts one subject, one session and fifteen recordings, with nothing worth a look.
The Converter before the conversion: the BIDS preview shows the PET sidecar the conversion will write, each value coloured by where it comes from, and Properties shows the PET row.The Converter before the conversion: the BIDS preview shows the PET sidecar the conversion will write, each value coloured by where it comes from, and Properties shows the PET row.
The Editor on the converted sample: the BIDS tree with its validation marks, the BOLD run in the Planes + 3-D layout with its time course and physiological recordings, and the Validation panel.The Editor on the converted sample: the BIDS tree with its validation marks, the BOLD run in the Planes + 3-D layout with its time course and physiological recordings, and the Validation panel.

What runs underneath

Established tools, and the work around them

Converting a file format is a solved problem, and BIDS Manager does not solve it again: each kind of recording goes to the tool built for it. What BIDS Manager adds is everything around them. Pick a kind of recording to follow its path.

Follow

  1. 01

    Scan

    Reads every file's header, groups the files by participant and session, and proposes a BIDS name for each recording. Nothing is written into the raw data.

    • pydicomreads every DICOM headerMRI, MR spectroscopy, PET, Physiology
    • dcm2niixproposes a name from the header strings (a sidecar only, no image)MRI, MR spectroscopy, PET
    • the parameter classifiernames each MR series from how it was acquiredMRI
    • nibabelreads an ECAT file's headerPET
    • MNE-Pythonreads each recording's header and channelsEEG, MEG
    • eyelinkioreads an EyeLink EDF (an ASC file is read by BIDS Manager)Eye tracking

  2. 02

    Curate

    The table: one row per recording, every name editable, rules and bulk edits, the Overview across subjects, and the metadata no file records: templates per kind of file, dose sheets, participants.

    • the scan tableone row per recording; every edit recorded in the projectevery kind
    • Dataset metadatatemplates per kind of file, the dataset description, dose sheets, participantsevery kind
    • the BIDS previewevery file the conversion will writeevery kind

  3. 03

    Convert

    Each recording goes to the tool built for it, in a staging folder per subject; a subject is written to the dataset only when all of it converted.

    • dcm2niixDICOM to NIfTI (and NIfTI-MRS)MRI, MR spectroscopy, PET
    • nibabelECAT images to NIfTIPET
    • PET2BIDSECAT header fields, the PET DICOM fields dcm2niix leaves out, PMOD blood curvesPET
    • MNE-BIDSEEG and MEG with their channels, events and coordinatesEEG, MEG
    • bidsphysiothe physiological log DICOM to _physio.tsv.gzPhysiology
    • BIDS Managereye tracking to _physio and _physioevents per eyeEye tracking

  4. 04

    Describe

    Fills what each converter leaves out: fieldmap links, identifiers removed, PET and MRS sidecars repaired, participants and scans tables, the dataset description, TODO placeholders for what is still unanswered.

    • fix-upsfieldmap links, identifiers removed, PET and MRS headers, eye-tracking and EEG sidecarsevery kind
    • the metadata stepyour answers, participants.tsv, _scans.tsv, TODO placeholdersevery kind

  5. 05

    Check and edit

    Validation against the standard, the Editor's viewers and tools, a fast quality check of anatomical and diffusion images, defacing.

    • bidsvalvalidates the dataset against the BIDS schema (kept inside BIDS Manager)every kind
    • the viewersimages, spectra, signals, gradient tables, documentsevery kind
    • brainchopbrain masks and tissue classes for the quality check and skull strippingMRI
    • niimathdefacing, and the quality check's registrationMRI, PET

done by BIDS Manager itself an established tool it installs and drives

How series are named

Named from how it was acquired, not from what it was called

A series name is typed by a person at the scanner, and people make mistakes there: a protocol changed from a T2-weighted image into a FLAIR keeps its old name, a copied protocol keeps another study's task. BIDS Manager reads what the scanner recorded about how it acquired each MRI series (echo and repetition times, an inversion, the readout, the diffusion gradients, the phase encoding, the image type) and names the series from that, with dcm2niix's guess as a second opinion. On 256 series checked by hand it named 99.2 % right, against 89.1 % without it.

Read moreHow names are decided: three real series judged by their name, by dcm2niix and by their parameters, and the order in which every step of a scan runs.

What it reads

Many kinds of recording, one scan

A single scan handles a folder that mixes all of these (PET blood curves are not scanned: you attach them to their PET run in **Properties**). Each is converted by the tool built for it, which BIDS Manager installs and drives.

The enginesdcm2niix, MNE-BIDS, nibabel and PET2BIDS do the converting, eyelinkio reads EyeLink files, and BIDS Manager reads Siemens physiological logs itself. BIDS Manager decides what each is asked to do, and repairs what each leaves behind.

  • MRI

    Anatomy, functional runs with their references, diffusion and fieldmaps.

    From
    DICOM from Siemens, Philips, GE, Canon and others
    To
    NIfTI in anat, func, dwi, fmap, perf
    With
    dcm2niix

  • MR spectroscopy

    Single-voxel spectra and their water references.

    From
    Siemens DICOM
    To
    NIfTI-MRS in mrs
    With
    dcm2niix

  • PET

    Dynamic and static scans, with the doses and blood data no scanner records.

    From
    DICOM, ECAT 7, PMOD blood curves
    To
    NIfTI and blood tables in pet
    With
    dcm2niix, nibabel, PET2BIDS

  • EEG

    Recordings with their channels, events, montage and reference.

    From
    EDF, BDF, BrainVision, EEGLAB, GDF, EGI MFF
    To
    The original format (EDF, BDF, BrainVision or EEGLAB) in eeg, or EDF when re-encoding is on
    With
    MNE-BIDS

  • MEG

    Task, rest and empty-room recordings, with sessions from their dates.

    From
    Elekta/MEGIN FIF, CTF, KIT
    To
    The original format in meg
    With
    MNE-BIDS

  • Physiology

    The heartbeat, pulse, breathing and triggers recorded with an MRI run.

    From
    The physiological log DICOM (PhysioLog) that Siemens CMRR sequences store with the run
    To
    _physio.tsv.gz beside the run, one table per sampling rate or one for the run
    With
    BIDS Manager's own reader

  • Eye tracking

    Gaze and pupil per eye, with the calibration, paired with the run they belong to.

    From
    SR Research EyeLink EDF and ASC
    To
    _physio.tsv.gz per eye
    With
    eyelinkio for EDF; BIDS Manager reads ASC itself

Why it exists

Curation, not only conversion

A converter turns files into files. A dataset others can use also needs the facts the raw files don't capture, names that agree across sources, no identifiers left behind, and a way to check all of it.

Read moreWhy BIDS Manager exists, and what curation means in practice.

i.

A table between the raw data and the dataset

Every recording is a row you can read and correct before anything is written: subjects that disagree across sources, a task the file name does not say, a run that was aborted.

ii.

Metadata nobody records

Authors, task descriptions, the EEG reference, the PET dose: asked once, for the dataset or each kind of file, and marked TODO where still missing.

iii.

Checked, and kept

Validation against the standard, a quality check of every anatomical and diffusion image, identifiers removed, and every change recorded in the project so the work can be resumed and traced.

Where to start

Pick the path that fits you

New to BIDSStart with BIDS in ten minutes: what a dataset looks like and how its files are named. Then follow the course. New to BIDS ManagerInstall it, download the multimodal sample, and take it from raw files to a valid dataset in about fifteen minutes. Converting your own dataGo to the guide for your kind of recording: MRI, PET, EEG, MEG, eye tracking, or several at once.

Cite

If BIDS Manager helped your work

Please cite the software, and the engines it relies on, which deserve the credit for the converting.

How to cite BIDS Manager What changed in each release