10. An introduction to BIDSPath#

BIDSPath is MNE-BIDS’s working horse when it comes to file and folder operations. Learn here how to use it.

# Authors: Richard Höchenberger <richard.hoechenberger@gmail.com>
#          Stefan Appelhoff <stefan.appelhoff@mailbox.org>
# License: BSD-3-Clause

Obviously, to start exploring BIDSPath, we first need to import it.

from pathlib import Path

import mne_bids
from mne_bids import BIDSPath

Now let’s discuss a little bit of background on the BIDS file and folder naming scheme. The first term we are going to introduce is the BIDS root. The BIDS root is simply the root folder of your BIDS dataset. For example, if the BIDS data of one of your studies is stored in /Users/me/Studies/Study_01, then this will be the BIDS root.

Similarly, if you have no BIDS dataset to begin with, you need to consider where to store your data upon BIDS conversion. Again, the intended target folder will be the BIDS root of your data.

For the purpose of this demonstration, let’s pick the tiny_bids example dataset that ships with the MNE-BIDS test suite.

# We are using a pathlib.Path object for convenience, but you could just use
# a string to specify ``bids_root`` here.
bids_root = Path(mne_bids.__file__).parent / "tests" / "data" / "tiny_bids"

This refers to a folder named my_bids_root in the current working directory. Finally, let is create a BIDSPath, and tell it about our BIDS root. We can then also query the BIDSPath for its root.


Great! But not really useful so far. BIDS also asks us to specify subject identifiers. We can either create a new BIDSPath, or update our existing one. The value can be retrieved via the .subject attribute.

subject = "01"

# Option 1: Create an entirely new BIDSPath.
bids_path_new = BIDSPath(subject=subject, root=bids_root)

# Option 2: Update the existing BIDSPath in-place.

In this example, we are going to update the existing BIDSPath using its update() method. But note that all parameters we pass to this method can also be used when creating a BIDSPath.

Many studies consist of multiple sessions. As you may have guessed, BIDS specifies how to store data for each session, and consequently, BIDSPath handles this for you too! Let’s update our BIDSPath with information on our experimental session, and try to retrieve it again via .session.

session = "eeg"

Now that was easy! We’re almost there! We also need to specify a data type, i.e., meg for MEG data, eeg and ieeg for EEG and iEEG data, or anat for anatomical MRI scans. Typically, MNE-BIDS will infer the data type of your data automatically, for example when writing data using mne_bids.write_raw_bids. For the sake of this example, however, we are going to specify the data type explicitly.


Excellent! Let’s have a look at the path we have constructed!


As you can see, BIDSPath automatically arranged all the information we provided such that it creates a valid BIDS folder structure. You can also retrieve a pathlib.Path object of this path:


Let’s have a closer look at the components of our BIDSPath again.

root: C:/Users/stefan/Desktop/git-repos/mne-bids/mne_bids/tests/data/tiny_bids
datatype: eeg
basename: sub-01_ses-eeg)

The most interesting thing here is probably the basename. It’s what MNE-BIDS uses to name individual files. The basename consists of a set of so-called entities, which are concatenated using underscores. You can access it directly:


The two entities you can see here are the subject entity (sub) and the session entity (ses). Each entity name also has a value; for sub, this is 01, and for ses, it is eeg in our example. Entity names (or “keys”) and values are separated via hyphens. BIDS knows a much larger number of entities, and MNE-BIDS allows you to make use of them. To get a list of all supported entities, use:

{'subject': '01', 'session': 'eeg', 'task': None, 'acquisition': None, 'run': None, 'processing': None, 'space': None, 'recording': None, 'split': None, 'description': None}

As you can see, most entity keys are set to None, which is the default and implies that no value has been set. Let us add a run entity, and remove the session:

run = "01"
session = None
bids_path.update(run=run, session=session)
root: C:/Users/stefan/Desktop/git-repos/mne-bids/mne_bids/tests/data/tiny_bids
datatype: eeg
basename: sub-01_run-01)

As you can see, the basename has been updated. In fact, the entire path has been updated, and the ses-eeg folder has been dropped from the path:


However, for the sake of the present examples, let’s revert to the iteration of our bids_path instance with no run and session="eeg", as that file exists in the tiny_bids dataset.

bids_path.update(run=None, session="eeg")

Awesome! We’re almost done! Two important things are still missing, though: the so-called suffix and the filename extension. Sometimes these terms are used interchangeably, but in BIDS, they have a very specific and different meaning!

The suffix is the last part of a BIDS filename before the extension. It is the same as the datatype for MEG, EEG, and iEEG recordings (i.e. meg, eeg, and ieeg, respectively) and T1w for T1-weighted MRI scans. But the suffix is also used to create the names of sidecar files like *_events.tsv.

Which brings us directly to the extension: the very last part of a filename. In MNE-BIDS, the extension contains a leading period, e.g. .tsv. Let’s put our new knowledge to use!

bids_path.update(suffix="eeg", extension=".vhdr")

root: C:/Users/stefan/Desktop/git-repos/mne-bids/mne_bids/tests/data/tiny_bids
datatype: eeg
basename: sub-01_ses-eeg_eeg.vhdr)

By default, most MNE-BIDS functions will try to infer to correct suffix and extension for your data, and you don’t need to specify them manually.

Total running time of the script: (0 minutes 0.012 seconds)

Gallery generated by Sphinx-Gallery