Physiological data alongside neuroimaging data¶

This guide explains how to create BIDS-compliant physiological recordings (*_physio.tsv.gz + *_physio.json) together with neural recordings, following the BIDS specification.

1. Create the TSV file (compressed)¶

• Save your physio data as a tab-separated values file.
• Do not include a header line.
• Each row = one time point, each column = one channel (pulse, respiration, trigger, etc.).
• Compress the file with gzip:

gzip sub-01_task-rest_physio.tsv

2. Create the JSON sidecar (required fields)¶

Each *_physio.tsv.gz must have a matching *_physio.json with:

• SamplingFrequency: sampling frequency in Hz — applies to all columns
• StartTime: time in seconds relative to the start of acquisition of the first neural data sample (can be negative)
• Columns: array listing column names in the same order as the TSV columns

{
  "SamplingFrequency": 100.0,
  "StartTime": -22.345,
  "Columns": ["cardiac", "respiratory", "trigger"]
}

Recommended field:

• PhysioType: type of physiological recording. Defaults to "generic"; use "eyetrack" for eye-tracking data.

Optional hardware fields: Manufacturer, ManufacturersModelName, DeviceSerialNumber, SoftwareVersions

3. Add descriptions and units for each column¶

Each column listed in Columns can have its own metadata block in the JSON:

{
  "SamplingFrequency": 100.0,
  "StartTime": -22.345,
  "Columns": ["cardiac", "respiratory", "trigger"],
  "cardiac": {
    "Description": "continuous pulse measurement",
    "Units": "mV"
  },
  "respiratory": {
    "Description": "respiration belt signal",
    "Units": "mV"
  },
  "trigger": {
    "Description": "scanner trigger signal",
    "Units": "a.u."
  }
}

4. Physiological events (_physioevents)¶

If your recording contains discrete events (e.g. blink onsets, artefact markers), you can store them separately as:

• *_physioevents.tsv.gz — event onsets and durations within the physiological recording
• *_physioevents.json — sidecar with column descriptions

5. Handle multiple sampling frequencies or recording devices¶

Split them into separate files and distinguish with _recording-<label> in the filename:

sub-01_task-rest_recording-cardiac_physio.tsv.gz
sub-01_task-rest_recording-respiratory_physio.tsv.gz

6. Place your files in the correct folder¶

Physiological files go in the same directory as the matching imaging data. A session directory is mandatory in the ANC, even for single-session studies.

sub-<label>/
└── ses-<label>/
    └── <datatype>/
        ├── <matches>[_recording-<label>]_physio.tsv.gz
        ├── <matches>[_recording-<label>]_physio.json
        ├── <matches>[_recording-<label>]_physioevents.tsv.gz
        └── <matches>[_recording-<label>]_physioevents.json

• <datatype>: func, anat, meg, eeg, ieeg, dwi, beh, etc.
• <matches>: filename entities matching the imaging file (e.g. sub-01_ses-01_task-rest_run-1).

Checklist¶

• File is *_physio.tsv.gz (gzipped, no header)
• Matching *_physio.json exists
• JSON includes SamplingFrequency, StartTime, Columns
• Column order in Columns matches TSV columns exactly
• Per-column Description and Units included
• StartTime is aligned to the neural recording onset
• Different sampling rates or devices → stored in separate files with _recording-<label>