SCVR

Scripts for Spinal Cord Vascular Reactivity (SCVR) delay mapping. The main functions are contained in groupdelay.py. Example run files are provided for group delay and individual subject delay (with multiple runs).

---

---

Info:

These scripts were designed to take outputs from higher-level/group-level modeling including:

• t-statistic maps (other statistical maps would likely work but were not tested)
• parameter estimate maps (i.e., SCVR)
• group mask (this is the consensus mask in which the group analyses were run in)

This assumes that first-level and higher-level modeling were already performed at each desired temporal shift (delay).

---

Main Files and Usage:

groupdelay.py: Main code file that defines functions used in the run files. Should not need to be edited.

Group SCVR

1. runGroupDelay.py: Run file for calculation of group delay map.

• Takes the group-level statistic maps (e.g., t-stat maps from randomise) for all temporal shifts and calculates a delay map across the group.
• Outputs:
  • delayMap.nii.gz: 3D voxel-wise delay map
  • tstatMapsAll.nii.gz: 4D collated t-stat map for each time shift, used to calculate voxelwise max

2. runGroupDelayCorr.py: Run file for group delay correction.

• Takes parameter estimate files (e.g., cope files from FEAT, warped to template space) at each temporal shift and delay map to calculate delay-corrected SCVR maps.
• Outputs:
  • <prefix>_delayCorrSCVR.nii.gz: delay-corrected SCVR map for each subject/run (3D)
  • <prefix>_peMapsAll.nii.gz: 4D collated parameter estimate (SCVR) maps for each subject/run

3. Collect all delay-corrected SCVR files across subjects/runs and run higher-level stats (e.g., randomise).

Individual SCVR

1. runIndivDelay.py: Run file for individual delay map (e.g., for highly-sampled subjects).

• Takes the higher-level statistic maps (e.g., t-stat maps from randomise) for all temporal shifts and calculates a delay map across the runs.
• Outputs:
  • delayMap.nii.gz: 3D voxel-wise delay map
  • tstatMapsAll.nii.gz: 4D collated t-stat map for each time shift, used to calculate voxel-wise max

2. runIndivDelayCorr.py: Run file for individual delay correction.

• Takes parameter estimate files (e.g., cope files from FEAT, warped to template space) at each temporal shift and delay map to calculate delay-corrected SCVR maps.
• Outputs:
  • <prefix>_delayCorrSCVR.nii.gz: delay-corrected SCVR map for each fMRI run (3D)
  • <prefix>_peMapsAll.nii.gz: 4D collated parameter estimate (SCVR) maps for each run

3. Collect all delay-corrected SCVR files across runs and run higher-level stats (e.g., randomise).

---

Notes:

• Randomise was used for higher-level modeling in the development of these methods
• This repository contains code to accompany the manuscript cited below, and has not been prepared as a general package for SCVR modeling and hemodynamic delay analysis.

---

Citation:

Hemmerling KJ, Hoggarth MA, Sandhu MS, Parrish TB, Bright MG. MRI mapping of hemodynamics in the human spinal cord. bioRxiv [Preprint]. 2024 Feb 27:2024.02.22.581606. doi: 10.1101/2024.02.22.581606.