[STORY] segment T1-weighted anatomical MRI images

[arkiev]

Metadata

Epic: #17
Feature: #79
Feature Release:
Required knowledge: mid-level

Description

As a researcher, I want to segment T1-weighted anatomical MRI images in order to
(1) quantify brain volume (as well as GM, WM & CSF volume),
(2) derive an anatomical brainmask,
(3) derive a mask for cortical/sub-cortical structures, and/or,
(3) use segmentation image for subsequent analysis (i.e., anatomically-constrained diffusion tractography).

Notes

This process should include 3 sub-processes:
(1) FreeSurfer/FastSurfer recon-all
(2) 5-tissue-type generation* 5ttgen hsvs
(3) Desikan-Killiany (DK) parcellation*

*using mrtrix3 commands in Pydra

Acceptance Criteria

Given a T1-weighted anatomical MRI image, when this pipeline is executed, the following outputs should be generated:

• standard FreeSurfer/FastSurfer outputs directory
• 5-tissue-type image
• 5-tissue-type image visualisation image (vis image)
• DK parcellation image (indexed from 1 to 84)

Bugs to fix

Preview Give feedback

1. magic number error with .mgz file format #241
   +0
2. labelconvert_.py has repeating arguments in the input_spec #240
   +0
3. zip .mif files #243
   +0
4. expand FastSurfer task outputs to include images from /mri directory #244
   +0