Quick Start
This page shows example commands to run AutoHS. Install options (Docker, Apptainer, source) are on Installation.
Validate your dataset
AutoHS expects a BIDS-valid dataset with at least one T1-weighted anatomical scan per subject. We recommend validating with the BIDS Validator.
bids-validator /path/to/bids_dataset
IDEAS sample (recommended first run)
Download two public IDEAS subjects (OpenNeuro ds005602), then run AutoHS:
./scripts/download_ideas_sample.sh
python run.py sample_data/ideas_bids bids_output participant \
--participant-label 1 2 \
--fastsurfer \
--runtime apptainer \
-w bids_output/work
See Sample data and Citation for required citations. Outputs:
bids_output/autohs/sub-*/.
Docker example (one subject)
Requires FreeSurfer license.txt when not using --fastsurfer.
docker run --rm \
-v /path/to/bids_dataset:/data:ro \
-v /path/to/output:/out \
-v /path/to/license.txt:/license.txt:ro \
autohs/autohs:latest \
/data /out participant \
--participant-label 001 \
--fs-license-file /license.txt
Apptainer example (HPC)
Set FREESURFER_SIF / FASTSURFER_SIF before running (see Installation).
export AUTOHS_RUNTIME=apptainer
export FREESURFER_SIF=/path/to/freesurfer_7.4.1.sif
apptainer exec --bind /path/to/bids_dataset:/data:ro \
--bind /path/to/output:/out \
--bind /path/to/license.txt:/license.txt:ro \
autohs.sif \
/data /out participant \
--participant-label 001 \
--fs-license-file /license.txt
FastSurfer (no FreeSurfer license)
python run.py /path/to/bids_dataset /path/to/output participant \
--participant-label 001 \
--fastsurfer \
--runtime apptainer
FastSurfer is typically 5–10× faster than FreeSurfer on CPU. CLI reference: Usage.