PRIMAS: Precision Functional Imaging in Autism

The PRISMA dataset is organised according to the Brain Imaging Data Structure
(BIDS) specification version 1.10.1 and is publicly available at doi:10.18112/openneuro.ds007182.v1.0.0
under accession number ds007182.

### Key naming conventions:

- Participant labels: sub-001 to sub-066
- Session labels: ses-01, ses-02, ses-03
- Task labels: rest, news, realitytv
- Run labels: run-01, run-02, run-03, ...
- Direction labels (dwi/fmap): AP (anterior-posterior), PA (posterior-anterior)

### File Formats

All neuroimaging data are stored as compressed NIfTI files (.nii.gz) with
accompanying JSON sidecar files containing acquisition parameters. Physiological
recordings are stored as compressed tab-separated value files (.tsv.gz) following
BIDS specification for physiological recordings. Behavioural and phenotypic data
are provided as tab-separated value files (.tsv) with JSON metadata files.

### Metadata Files

- `dataset_description.json`: Dataset-level metadata including BIDS version,
dataset name, authors, and acknowledgements
- `participants.tsv`: Demographic information for all participants (age, sex,
handedness, group assignment, education level, time between sessions)
- `participants.json`: Data dictionary for participants.tsv
- `README`: Detailed dataset description, experimental procedures, and data
collection protocols
- `CHANGES`: Version history and dataset updates
- `*_scans.tsv`: Scan-specific metadata (acquisition time, quality notes)

### Derivatives

Preprocessed data are provided in `derivatives/` following BIDS Derivatives
specification:

- `fmriprep/`: Functional MRI data preprocessed with fMRIPrep version X.X.X,
including motion-corrected, spatially normalised, and surface-projected data
- `qsiprep/`: Diffusion MRI data preprocessed with QSIPrep version X.X.X,
including denoised, distortion-corrected, and eddy-current-corrected data
- `qsirecon/`: Structural connectomes derived using mrtrix_singleshell_ss3t
pipeline with anatomically constrained tractography
- `physio/`: Physiological data for peak detection, heart rate
variability, and respiratory volume per time calculations

Each derivative directory contains a `dataset_description.json` file specifying
the generating software and version.

### Defacing and Privacy

All structural images have been defaced using MiDeFace 2, which is part of
FreeSurfer v7.4.1 Docker image. The de-facing mask was derived using the
T1-weighted images and applied to the T2-weighted images after corregistration
using the ‘mri_coreg’ function of FreeSurfer.

### Missing Data

Not all participants have complete data across all modalities and sessions.
Data availability is documented in:
- `participants.tsv`: Indicates which sessions each participant completed
- `sub-<label>_sessions.tsv`: Session-specific completion status
- Individual `*_scans.tsv` files: Scan-level quality notes and usability flags

Known missing data patterns:
- T1w images: 65/65 participants (100%)
- T2w images: 62/65 participants (95.4%)
- DWI: 61/65 participants (94.8%)
- Functional scans: See Scientific Data publication for detailed breakdown
- Physiological recordings: See Scientific Data publication for detailed breakdown

### Data Access and Licensing

The dataset is available on OpenNeuro under Creative Commons CC0. Users must cite this data
descriptor publication when using the dataset.

DOI: [insert DOI once available]
Repository: [insert repository URL]
Accession number: [insert accession number]

### Data Use Agreement

All participants provided written informed consent for public data sharing.
Users are required to:
1. Cite this data descriptor in any publications using the dataset
2. Respect participant privacy by not attempting re-identification
3. Follow repository-specific terms of use

Technical Validation data and analysis code are available at [GitHub/OSF URL].