Quick Start

Before you can use QSIRecon, you must have some preprocessed dMRI data. See Input Data for QSIRecon

The next step is to get a containerized version of QSIRecon. This can be done with Singularity, Apptainer or Docker. Most users run QSIRecon on a high performance computing cluster, so we will assume Apptainer is being used throughout this documentation. See Installation on how to create a sif file or pull the image with Docker.

Next, you need to decide which workflow you’d like to run. You can pick from any of the Built-In Reconstruction Workflows or Custom Reconstruction Workflows. Here we’ll pick the dsi_studio_autotrack workflow.

Finally, you’ll need to craft a command to set up your QSIRecon run. Suppose you’re in a directory where there are some qsiprep results in inputs/qsiprep. You’d like to save QSIRecon outputs in results. You have access to 8 cpus. To run the from qsirecon-latest.sif you could use:

apptainer run \
    --containall \
    --writable-tmpfs \
    -B "${PWD}" \
    qsirecon-latest.sif \
    "${PWD}/inputs/qsiprep" \
    "${PWD}/results/qsirecon" \
    participant \
    -w "${PWD}/work" \
    --nthreads 8 \
    --omp-nthreads 8 \
    --recon-spec dsi_studio_autotrack \
    -v -v

Once this completes you will see a number of new directories written to results. You will find errors (if any occurred) and configuration files for each subject directly under results/sub-*. Each analysis also creates its own directory that contains results per subject. In the case of dsi_studio_autotrack we will see results/qsirecon-DSIStudio/sub-* containing the outputs from the ss3t_autotrack workflow. Some workflows produce multiple directories, particularly when multiple models are fit.

Command-Line Arguments

Troubleshooting

Logs and crashfiles are outputted into the <output dir>/qsirecon/sub-<participant_label>/log directory. Information on how to customize and understand these files can be found on the nipype debugging page.