Organizing EEG data using the Brain Imaging Data Structure
1 like380 views
The document discusses the organization of EEG data using the Brain Imaging Data Structure (BIDS) and highlights the importance of file formats, channels, and metadata in making EEG data interoperable and accessible. It emphasizes the transition from traditional data sharing methods to modern collaborative open science tools and platforms for data documentation and reproducibility. The talk also covers the creation and reuse of BIDS-compliant EEG datasets and advocates for the sharing of EEG data to enhance research impact.
Organizing EEG data using the Brain Imaging Data Structure
- 1. Organizing EEG data using the Brain Imaging Data Structure Robert Oostenveld Donders Institute, Radboud University, Nijmegen, NL Karolinska Institutet, Stockholm, SE r.oostenveld@donders.ru.nl
- 2. https://bids-specification.readthedocs.io/ http://bids.neuroimaging.io these slides will be shared on https://www.slideshare.net/RobertOostenveld
- 3. Science – methods and tools are changing
- 4. Open Science – infrastructure and tools Git and GitHub, Gitlab, BitBucket Work together on code for analyses Open Science Framework (osf.io) Work together on documenting DataVerse, Zenodo, etc Sharing of data Code Ocean, Microsoft Azure, Anaconda Clould Cloud-based computational reproducibility platform Past - Black-and-white version of article printed on paper and stored in library Present - PDF for download, sometimes online supplementary material Future - Online notebooks that allow full reproduction of results (and more) Lab notebook Science is getting more exiting – but also requires new skills
- 5. Outline of this talk How is BIDS extended to EEG? What is specific for EEG? File formats Channels and electrodes Coordinates Other details Re-using BIDS EEG datasets Creating BIDS EEG datasets
- 6. Brain Imaging Data Structure https://www.nature.com/articles/sdata201644 https://www.nature.com/articles/sdata2018110
- 7. https://psyarxiv.com/63a4y/ https://psyarxiv.com/r7vc2/ Both accepted for publication in Scientific Data and soon to appear
- 8. BIDS for MRI, MEG, EEG, iEEG,etc. data/README CHANGES dataset_description.json participants.tsv /sub-01/anat/… /sub-01/meg/… /sub-01/eeg/sub-01_task-auditory_eeg.edf /sub-01/eeg/sub-01_task-auditory_eeg.json /sub-01/eeg/sub-01_task-auditory_channels.tsv /sub-01/eeg/sub-01_task-auditory_events.tsv /sub-01/eeg/sub-01_electrodes.tsv /sub-01/eeg/sub-01_coordinates.json raw EEG data general metadata electrode positions
- 9. Fileformats for EEG data in BIDS Recommended BrainVision Core file fomat (*.vhdr, *.vmrk, *.eeg) European Data Format (*.edf) Allowed Biosemi (*.bdf) EEGLAB (*.set) -> MATLAB, HDF5-based Limited number of formats that should … 1) address the needs of a large portion of the global EEG community 2) be a FAIR format, with a focus on Interoperability 3) meet the technical requirements of neuroscientific workflows https://www.brainproducts.com/productdetails.php?id=21&tab=5 https://www.biosemi.com/faq/file_format.htm https://www.edfplus.info https://www.hdfgroup.org/solutions/hdf5/ and EEGLAB
- 10. EEG data sidecar file Capture information that is not present in the data file • What was the task or instruction to the subject? • Which electrode cap was used? • Which amplifier was used? • Is line noise at 50Hz or 60Hz? • Where was the reference electrode placed? • … Makes some information more easily available • Number of channels with EEG, EOG, etc. • …
- 11. Channels and electrodes Electrode A single point of contact between the acquisition system and the recording site Channel A single analog-to-digital converter in the recording system that regularly samples the value of a transducer, which results in the signal being represented as a time series in the digitized data. Note that GND and REF are electrodes, but usually not channels.
- 12. Coordinates If EEG electrode positions are digitized and specified, then the coordinate system should also be specified. Fiducials objects with a well defined location used to facilitate the localization of electrodes and co-registration with other geometric data such as the participant's own T1 weighted magnetic resonance head image, a T1 weighted template head image, or a spherical head model. Anatomical landmarks locations on a research subject such as the nasion, which is the intersection of the frontal bone and two nasal bones of the human skull.
- 13. Task and event details http://www.hedtags.org http://wiki.cogpo.org https://www.cognitiveatlas.org
- 14. Outline of this talk How is BIDS extended to EEG? What is specific for EEG? File formats Channels and electrodes Coordinates Other details Re-using BIDS EEG datasets Creating BIDS EEG datasets
- 15. Re-using BIDS EEG datasets …. just use your standard tools! EEG data in BIDS is stored in formats supported by all EEG software. Metadata is in human and machine-readable format (tsv, json). In the future it is likely that complete (semi)-automatic analysis pipelines will be implemented to process complete datasets.
- 16. Creating BIDS EEG datasets http://www.fieldtriptoolbox.org/example/bids/ Tools are also currently being developed in EEGLAB, BrainStorm, SPM, MNE-Python, etc. BrainVision is also extending their software to be more BIDS friendly.
- 17. Quick demo
- 18. Why share your EEG data BIDS? Sharing your data (externally) increases your research impact. Archiving your data properly (internally) increases its future value. Document for your “future self”, so that you don’t forget the details (and that others will know them as well). Plan your experiment, acquire the data, organize it in BIDS, and only then start analyzing it!
- 19. Share/publish your research output Peer-reviewed publication is the primary output! More details in your analysis than fits in your “Methods” section. Not possible to describe all details in human-oriented text. Raw data EEG, lab notes, etc Analysis details MATLAB, Python, R, SPSS, Bash, … Analysis results Output of the computations -> BIDS for raw EEG data -> BIDS derivatives (work-in-progress) -> possibly BIDS apps, otherwise github or similar
- 20. https://bids-specification.readthedocs.io/ http://bids.neuroimaging.io these slides will be shared on https://www.slideshare.net/RobertOostenveld
Editor's Notes
- #5: OSF is the service by the Center for Open Science
- #10: Imagine in 10 years from now, where quite some of us will be working at different places, we might not be using our current tools any more: we want to archive and publish data such that it remains accessible
- #12: A channel can be connected to two electrodes (to measure the potential difference between them), a magnetic field or magnetic gradient sensor, temperature sensor, accelerometer, etc.