mrsiprep.tissue.fractions

Tissue fraction map helpers.

Functions

copy_tissue_to_derivatives(config, subject, ...)

Copy T1w-space tissue-probability maps into the MRSIPrep derivatives tree.

load_existing_cat12(config, subject, session)

Locate pre-computed CAT12 GM/WM/CSF tissue-probability maps for a recording.

resample_tissue_to_mrsi(config, subject, ...)

Resample T1w-space tissue-probability maps onto the native MRSI grid.
mrsiprep.tissue.fractions.copy_tissue_to_derivatives(config, subject, session, tissue_t1)[source]

Copy T1w-space tissue-probability maps into the MRSIPrep derivatives tree.

Used with the existing tissue backend to bring CAT12 maps (found wherever load_existing_cat12() located them) under MRSIPrep's own confounds/ output naming, alongside SynthSeg+FAST-produced maps.

Parameters:

  • config -- Run-wide mrsiprep.config.settings.MRSIPrepConfig.

  • subject (str) -- BIDS subject label, without the sub- prefix.

  • session (str | None) -- BIDS session label without the ses- prefix, or None for session-less datasets.

  • tissue_t1 (dict[str, Path]) -- {label: path} of T1w-space tissue-probability maps to copy, keyed by tissue label (e.g. "GM", "WM", "CSF").

Returns:

{label: path} of the copied files' derivative paths. Recordings already copied are skipped (returned as-is) unless config.overwrite_seg/config.overwrite is set.

Return type:

dict[str, Path]

mrsiprep.tissue.fractions.load_existing_cat12(config, subject, session)[source]

Locate pre-computed CAT12 GM/WM/CSF tissue-probability maps for a recording.

Used by the existing tissue_backend, which reuses tissue segmentation already produced outside MRSIPrep instead of running SynthSeg+FAST.

Parameters:

  • config -- Run-wide mrsiprep.config.settings.MRSIPrepConfig.

  • subject (str) -- BIDS subject label, without the sub- prefix.

  • session (str | None) -- BIDS session label without the ses- prefix, or None for session-less datasets.

Returns:

{"GM": path, "WM": path, "CSF": path}, the CAT12 p1/p2/p3 probability-segmentation files found in the BIDS layout.

Raises:

FileNotFoundError -- If any of the three maps is missing.

Return type:

dict[str, Path]

mrsiprep.tissue.fractions.resample_tissue_to_mrsi(config, subject, session, tissue_t1, mrsi_reference, t1_to_mrsi_transforms)[source]

Resample T1w-space tissue-probability maps onto the native MRSI grid.

Plain linear-interpolation resample, applying the (inverse) MRSI-to-T1w transform to bring each map back onto mrsi_reference's grid for use in partial-volume correction. For the MIDAS-mode PSF-aware alternative (which convolves with the MRSI spatial response before resampling instead of a bare linear interpolation), see mrsiprep.tissue.psf.resample_tissue_to_mrsi_psf().

Parameters:

  • config -- Run-wide mrsiprep.config.settings.MRSIPrepConfig.

  • subject (str) -- BIDS subject label, without the sub- prefix.

  • session (str | None) -- BIDS session label without the ses- prefix, or None for session-less datasets.

  • tissue_t1 (dict[str, Path]) -- {label: path} of T1w-space tissue-probability maps to resample.

  • mrsi_reference (Path) -- MRSI reference image defining the target grid.

  • t1_to_mrsi_transforms (list[Path]) -- Ordered ANTs/FSL transform file(s) mapping T1w space to MRSI space (i.e. the inverse of the MRSI-to-T1w registration).

Returns:

{label: path} of the resampled, MRSI-grid tissue-probability maps. Recordings already resampled are skipped (returned as-is) unless config.overwrite_seg/config.overwrite is set.

Return type:

dict[str, Path]