FAQ

General

Is AutoHS a BIDS App?

Yes. AutoHS follows the BIDS Apps command-line convention: autohs bids_dir output_dir participant [options].

Does AutoHS support group-level analysis?

No. AutoHS runs at the participant level only. See Usage.

FreeSurfer or FastSurfer?

  • FreeSurfer (default): conventional subcortical segmentation; requires a license; ~2–4 h/subject on CPU.

  • FastSurfer (--fastsurfer): deep-learning segmentation; no FreeSurfer license; ~30–60 min/subject.

Both backends use the same AI-compute step and HS thresholds. Details: Background and theory.

Data requirements

What inputs does AutoHS need?

At least one BIDS T1w anatomical scan per subject or session. Other modalities are ignored.

Can I process only some subjects or sessions?

Use --participant-label and --session-label, or --bids-filter-file — see Usage.

Outputs

Where are results written?

Under <output_dir>/autohs/. File layout and thresholds: Outputs and Background and theory.

Can I regenerate reports without re-running segmentation?

Yes, with --reports-only when the -w work directory contains prior segmentation.

Containers and HPC

Docker image not found?

Build locally (docker/Dockerfile.bidsapp), use a GitHub Release .sif, or wait for Docker Hub publish — see Installation.

How do I run on Slurm / Apptainer?

Set FREESURFER_SIF / FASTSURFER_SIF and --runtime apptainer. See Installation.

Something failed during a run?

See Troubleshooting.

Citations

What should I cite?

AutoHS software, the Brain Communications asymmetry paper, and any datasets or tools you used. See Citation.

Methods text for my manuscript?

python run.py --md-only-boilerplate — see Methods boilerplate.