API Reference

MRSIPrep is structured as a library of workflow-orchestration functions, one per pipeline stage, invoked in sequence by mrsiprep.workflows.participant.run_participant_workflow() -- the same top-level entry point the CLI calls. Each stage function takes the run's config object plus the outputs of earlier stages and returns a small dataclass result that later stages consume.

This page documents the internal architecture for anyone extending MRSIPrep or calling its stages directly from Python (import mrsiprep) rather than through the CLI. For end-user usage and the full CLI flag reference, see Basic Usage.

Configuration

`mrsiprep.config.settings`	Runtime configuration objects.
`mrsiprep.config.defaults`	Default values for MRSIPrep.

Top-level orchestration

`mrsiprep.workflows.participant`	Participant workflow orchestration.
`mrsiprep.workflows.base`	Workflow base helpers.

Pipeline stage workflows

Each module below wraps one pipeline stage (see the Workflow Architecture section on the home page for the stage order) behind a single run_*_workflow entry point.

`mrsiprep.workflows.anatomical`	Anatomical preparation workflow.
`mrsiprep.workflows.mrsi`	MRSI preprocessing workflow.
`mrsiprep.workflows.tissue`	Tissue workflow.
`mrsiprep.workflows.registration`	Registration workflow.
`mrsiprep.workflows.parcellation`	Parcellation workflow.
`mrsiprep.workflows.connectivity`	Metabolic profile and connectivity workflow.
`mrsiprep.workflows.reports`	Report workflow.

Registration

`mrsiprep.registration.mrsi_to_t1`	MRSI-to-T1 registration.
`mrsiprep.registration.t1_to_mni`	T1-to-MNI registration.
`mrsiprep.registration.subject_template`	Subject-level longitudinal template construction and template-to-MNI registration.
`mrsiprep.registration.transforms`	Transform path and application helpers.

Backend interfaces

Thin wrappers around the external registration/segmentation tools MRSIPrep shells out to or calls via Python bindings.

`mrsiprep.interfaces.ants`	ANTs/antspyx interface.
`mrsiprep.interfaces.fsl`	FSL FLIRT/FNIRT registration interface.
`mrsiprep.interfaces.freesurfer`	FreeSurfer helpers.
`mrsiprep.interfaces.chimera`	Chimera parcellation wrapper.

MRSI processing

`mrsiprep.mrsi.filtering`	Biharmonic-style MRSI spike filtering.
`mrsiprep.mrsi.masks`	MRSI mask helpers.
`mrsiprep.mrsi.quality`	Voxelwise MRSI QC masks.
`mrsiprep.mrsi.reference`	MRSI reference image generation.
`mrsiprep.mrsi.resampling`	MRSI resampling helpers.
`mrsiprep.mrsi.pvc`	Partial-volume correction via PETPVC.

Tissue segmentation

`mrsiprep.tissue.synthseg_fast`	SynthSeg-constrained FSL FAST tissue backend.
`mrsiprep.tissue.fuzzy_cmeans`	MIDAS-style fuzzy c-means tissue segmentation.
`mrsiprep.tissue.fractions`	Tissue fraction map helpers.
`mrsiprep.tissue.psf`	MIDAS-style PSF-based tissue-fraction convolution.

Parcellation and connectivity

`mrsiprep.parcellation.synthseg`	Lightweight SynthSeg cortical and subcortical parcellation.
`mrsiprep.parcellation.chimera_native`	Chimera native-space parcellation workflow.
`mrsiprep.parcellation.mni_atlas`	MNI atlas parcellation workflow.
`mrsiprep.parcellation.extraction`	Regional metabolite extraction.
`mrsiprep.parcellation.tissue_regression`	MIDAS-style (Maudsley et al. 2006, Eq.
`mrsiprep.connectivity.export`	Connectivity export helpers.
`mrsiprep.connectivity.connectivity`	Perturbation-based metabolite connectivity.

I/O

`mrsiprep.io.bids`	Input layout discovery for MRSIPrep.
`mrsiprep.io.loaders`	MRSI input loading helpers.
`mrsiprep.io.naming`	BIDS-like naming helpers.
`mrsiprep.io.derivatives`	Derivative directory helpers.
`mrsiprep.io.validators`	Input validation.
`mrsiprep.io.mrsinmrs`	MRSinMRS (Lin et al. 2021, PMID 33559967) sequence-parameter reporting.