Skip to content

Common principles


The keywords "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].

Throughout this protocol we use a list of terms. To avoid misunderstanding we clarify them here.

  1. Dataset - a set of neuroimaging and behavioral data acquired for a purpose of a particular study. A dataset consists of data acquired from one or more subjects, possibly from multiple sessions.

  2. Subject - a person or animal participating in the study.

  3. Session - a logical grouping of neuroimaging and behavioral data consistent across subjects. Session can (but doesn't have to) be synonymous to a visit in a longitudinal study. In general, subjects will stay in the scanner during one session. However, for example, if a subject has to leave the scanner room and then be re-positioned on the scanner bed, the set of MRI acquisitions will still be considered as a session and match sessions acquired in other subjects. Similarly, in situations where different data types are obtained over several visits (for example fMRI on one day followed by DWI the day after) those can be grouped in one session. Defining multiple sessions is appropriate when several identical or similar data acquisitions are planned and performed on all -or most- subjects, often in the case of some intervention between sessions (e.g., training).

  4. Data acquisition - a continuous uninterrupted block of time during which a brain scanning instrument was acquiring data according to particular scanning sequence/protocol.

  5. Data type - a functional group of different types of data. In BIDS we define six data types: func (task based and resting state functional MRI), dwi (diffusion weighted imaging), fmap (field inhomogeneity mapping data such as field maps), anat (structural imaging such as T1, T2, etc.), meg (magnetoencephalography), beh (behavioral).

  6. Task - a set of structured activities performed by the participant. Tasks are usually accompanied by stimuli and responses, and can greatly vary in complexity. For the purpose of this protocol we consider the so-called “resting state” a task. In the context of brain scanning, a task is always tied to one data acquisition. Therefore, even if during one acquisition the subject performed multiple conceptually different behaviors (with different sets of instructions) they will be considered one (combined) task.

  7. Event - a stimulus or subject response recorded during a task. Each event has an onset time and duration. Note that not all tasks will have recorded events (e.g., resting state).

  8. Run - an uninterrupted repetition of data acquisition that has the same acquisition parameters and task (however events can change from run to run due to different subject response or randomized nature of the stimuli). Run is a synonym of a data acquisition.

Compulsory, optional, and additional data and metadata

The following standard describes a way of arranging data and writing down metadata for a subset of neuroimaging experiments. Some aspects of the standard are compulsory. For example a particular file name format is required when storing structural scans. Some aspects are regulated but optional. For example a T2 volume does not need to be included, but when it is available it should be saved under a particular file name specified in the standard. This standard aspires to describe a majority of datasets, but acknowledges that there will be cases that do not fit. In such cases one can include additional files and subfolders to the existing folder structure following common sense. For example one may want to include eye tracking data in a vendor specific format that is not covered by this standard. The most sensible place to put it is next to the continuous recording file with the same naming scheme but different extensions. The solutions will change from case to case and publicly available datasets will be reviewed to include common data types in the future releases of the BIDS spec.

Source vs. raw vs. derived data

BIDS in its current form is designed to harmonize and describe raw (unprocessed or minimally processed due to file format conversion) data. During analysis such data will be transformed and partial as well as final results will be saved. Derivatives of the raw data (other than products of DICOM to NIfTI conversion) MUST be kept separate from the raw data. This way one can protect the raw data from accidental changes by file permissions. In addition it is easy to distinguish partial results from the raw data and share the latter. Similar rules apply to source data which is defined as data before harmonization and/or file format conversion (for example E-Prime event logs or DICOM files).

This specification currently does not go into details of recommending a particular naming scheme for including different types of source data (raw event logs, parameter files, etc. before conversion to BIDS) and data derivatives (correlation maps, brain masks, contrasts maps, etc.). However, in the case that these data are to be included:

  1. These data MUST be kept in separate sourcedata and derivatives folders each with a similar folder structure as presented below for the BIDS-managed data. For example: derivatives/fmriprep/sub-01/ses-pre/sub-01_ses-pre_mask.nii.gz or sourcedata/sub-01/ses-pre/func/sub-01_ses-pre_task-rest_bold.dicom.tgz or sourcedata/sub-01/ses-pre/func/MyEvent.sce.

  2. A README file SHOULD be found at the root of the sourcedata or the derivatives folder (or both). This file should describe the nature of the raw data or the derived data. In the case of the existence of a derivatives folder, we RECOMMEND including details about the software stack and settings used to generate the results. Inclusion of non-imaging objects that improve reproducibility are encouraged (scripts, settings files, etc.).

  3. We RECOMMEND including the PDF print-out with the actual sequence parameters generated by the scanner in the sourcedata folder.

The Inheritance Principle

Any metadata file (.json, .bvec, .tsv, etc.) may be defined at any directory level, but no more than one applicable file may be defined at a given level (Example 1). The values from the top level are inherited by all lower levels unless they are overridden by a file at the lower level. For example, sub-*_task-rest_bold.json may be specified at the participant level, setting TR to a specific value. If one of the runs has a different TR than the one specified in that file, another sub-*_task-rest_bold.json file can be placed within that specific series directory specifying the TR for that specific run. There is no notion of "unsetting" a key/value pair. For example if there is a JSON file corresponding to particular participant/run defining a key/value and there is a JSON file on the root level of the dataset that does not define this key/value it will not be "unset" for all subjects/runs. Files for a particular participant can exist only at participant level directory, i.e /dataset/sub-*[/ses-*]/sub-*_T1w.json. Similarly, any file that is not specific to a participant is to be declared only at top level of dataset for eg: task-sist_bold.json must be placed under /dataset/task-sist_bold.json

Example 1: Two JSON files that are erroneously at the same level.


In the above example, two JSON files are listed under sub-01/ses-test/, which are each applicable to sub-01_ses-test_task-overtverbgeneration_run-2_bold.nii.gz, violating the constraint that no more than one file may be defined at a given level of the directory structure. Instead sub-01_ses-test_task-overtverbgeneration_run-2_bold.json should have been under sub-01/ses-test/func/.

Example 2: Multiple run and rec with same acquisition (acq) parameters acq-test1


For the above example, all NIfTI files are acquired with same scanning parameters (acq-test1). Hence a JSON file describing the acq parameters will apply to different runs and rec files. Also if the JSON file (task-xyz_acq-test1_bold.json) is defined at dataset top level directory, it will be applicable to all task runs with test1 acquisition parameter.

Example 3: Multiple json files at different levels for same task and acquisition parameters


In the above example, the fields from task-xyz_acq-test1_bold.json file will apply to all bold runs. However, if there is a key with different value in sub-01/func/sub-01_task-xyz_acq-test1_run-1_bold.json, the new value will be applicable for that particular run/task NIfTI file/s.

Good practice recommendations

Try to avoid excessive amount of overrides. Do not specify a field value in the upper levels if lower levels have more or less even distribution of multiple possible values. E.g., if a field X has one value for all ses-01/ and another for all ses-02/ it better not to be defined at all in the .json at the upper level.

File Formation specification

Imaging files

All imaging data MUST be stored using the NIfTI file format. We RECOMMEND using compressed NIfTI files (.nii.gz), either version 1.0 or 2.0. Imaging data SHOULD be converted to the NIfTI format using a tool that provides as much of the NIfTI header information (such as orientation and slice timing information) as possible. Since the NIfTI standard offers limited support for the various image acquisition parameters available in DICOM files, we RECOMMEND that users provide additional meta information extracted from DICOM files in a sidecar JSON file (with the same filename as the .nii[.gz] file, but with a .json extension). Extraction of BIDS compatible metadata can be performed using dcm2nii and dicm2nii DICOM to NIfTI converters. A provided validator will check for conflicts between the JSON file and the data recorded in the NIfTI header.

Tabular files

Tabular data MUST be saved as tab delimited values (.tsv) files, i.e. csv files where commas are replaced by tabs. Tabs MUST be true tab characters and MUST NOT be a series of space characters. Each TSV file MUST start with a header line listing the names of all columns (with the exception of physiological and other continuous acquisition data - see below for details). Names MUST be separated with tabs. String values containing tabs MUST be escaped using double quotes. Missing and non-applicable values MUST be coded as n/a. TSV files MUST be in UTF-8 encoding.


onset duration  response_time correct stop_trial  go_trial
200 200 0 n/a n/a n/a

Tabular files MAY be optionally accompanied by a simple data dictionary in a JSON format (see below). The data dictionaries MUST have the same name as their corresponding tabular files but with .json extensions. If a JSON file is provided, it MAY contain one or more fields describing the columns found in the TSV file (in addition to any other metadata one wishes to include that describe the file as a whole). Note that if a field name included in the JSON sidecar matches a column name in the TSV file, then that field MUST contain a description of the corresponding column, using an object containing the following fields:

Field name Definition
LongName Long (unabbreviated) name of the column.
Description Description of the column.
Levels For categorical variables: a dictionary of possible values (keys) and their descriptions (values).
Units Measurement units. [<prefix symbol>] <unit symbol> format following the SI standard is RECOMMENDED (see Appendix V).
TermURL URL pointing to a formal definition of this type of data in an ontology available on the web.


  "test": {
    "LongName": "Education level",
    "Description": "Education level, self-rated by participant",
    "Levels": {
      "1": "Finished primary school",
      "2": "Finished secondary school",
      "3": "Student at university",
      "4": "Has degree from university"
  "bmi": {
    "LongName": "Body mass index",
    "Units": "kilograms per squared meters",
    "TermURL": ""

Key/value files (dictionaries)

JavaScript Object Notation (JSON) files MUST be used for storing key/value pairs. JSON files MUST be in UTF-8 encoding. Extensive documentation of the format can be found here: Several editors have built-in support for JSON syntax highlighting that aids manual creation of such files. An online editor for JSON with built-in validation is available at:


  "RepetitionTime": 3,
  "Instruction": "Lie still and keep your eyes open"

Participant names and other labels

BIDS uses custom user-defined labels in several situations (naming of participants, sessions, acquisition schemes, etc.) Labels are strings and MUST only consist of letters (lower or upper case) and/or numbers. If numbers are used we RECOMMEND zero padding (e.g., 01 instead of 1 if you have more than nine subjects) to make alphabetical sorting more intuitive. Please note that the sub- prefix is not part of the subject label, but must be included in file names (similarly to other key names). In contrast to other labels, run and echo labels MUST be integers. Those labels MAY include zero padding, but this is NOT RECOMMENDED to maintain their uniqueness.


All units SHOULD be specified as per International System of Units (abbreviated as SI, from the French Système international (d'unités)) and can be SI units or SI derived units. In case there are valid reasons to deviate from SI units or SI derived units, the units MUST be specified in the sidecar JSON file. In case data is expressed in SI units or SI derived units, the units MAY be specified in the sidecar JSON file. In case prefixes are added to SI or non-SI units (e.g. mm), the prefixed units MUST be specified in the JSON file (see Appendix V: Units). In particular:

  • Elapsed time SHOULD be expressed in seconds. Please note that some DICOM parameters have been traditionally expressed in milliseconds. Those need to be converted to seconds.

  • Frequency SHOULD be expressed in Hertz.

Describing dates and timestamps:

  • Date time information MUST be expressed in the following format YYYY-MM-DDThh:mm:ss (one of the ISO8601 date-time formats). For example: 2009-06-15T13:45:30

  • Time stamp information MUST be expressed in the following format: 13:45:30

  • Dates can be shifted by a random number of days for privacy protection reasons. To distinguish real dates from shifted dates always use year 1900 or earlier when including shifted years. For longitudinal studies please remember to shift dates within one subject by the same number of days to maintain the interval information. Example: 1867-06-15T13:45:30

  • Age SHOULD be given as the number of years since birth at the time of scanning (or first scan in case of multi session datasets). Using higher accuracy (weeks) should in general be avoided due to privacy protection, unless when appropriate given the study goals, e.g., when scanning babies.

Directory structure

Single session example

This is an example of the folder and file structure. Because there is only one session, the session level is not required by the format. For details on individual files see descriptions in the next section:



Additional files and folders containing raw data may be added as needed for special cases. They should be named using all lowercase with a name that reflects the nature of the scan (e.g., calibration). Naming of files within the directory should follow the same scheme as above (e.g., sub-control01_calibration_Xcalibration.nii.gz)