Gradientography pipeline for task fMRI

Here we propose to apply gradientography to study task fMRI. Gradientography is a pipeline that uses resting-state functional MRI (fMRI) to map the complex topographic organization of the human subcortex, enabling the characterization of cortico–subcortical connectivity. We adapted the pipeline using psychophysiological interactions (PPI) to study the task-induced changes in connectivity.

Most of the Matlab scripts described below are adaptations of the gradientography pipeline available on github. Results visualizations are done in Python, using Nilearn.

More information on the pipeline is accessible in our paper "Functional re-organization of hippocampal-cortical gradients during naturalistic memory processes." (Borne et al. 2023). Our reviewer response is accessible here.

preprocessing.m

This script do a few additional preprocessing steps on top of HCP minimal preprocessing pipeline (Glasser et al 2013).

similarity_matrix.m

This script compute the similarity matrix, using PPI, from the preprocessed fMRI scan.

gradients.m

This script compute the gradients for each task and each group from the similarity matrices.

The result folder has the following organisation:

result
└───tasks
│   └───naive
│       └───cohorts
|           └───hc
│               │   subjects.txt
│               │   savg.mat
│               │   Vn2_eigenvector.nii
│               │   Vn2_magnitude.nii
│               │   ...
|           └───cc
│               │   ...
│   └───continuing
│       └───cohorts
|           └───hc
│               │   ...
|           └───cc
│               │   ...

permutation.m

This script do a permutation test between cohorts (for a fixed task) or between tasks (for a given cohort).

For a cohort permutation (for the naive task), the result folder needs to have the following organisation:

result
└───tasks
│   └───naive
│       └───cohorts
|           └───hc
│               │   subjects.txt
|           └───cc
│               │   subjects.txt

For a task permutation (for the healthy cohort), the result folder needs to have the following organisation:

result
└───cohorts
│   └───hc
|       │   subjects.txt

The script will create a permutation folder in each group/task permuted. The output files are used for the visualisation made by the Python scripts below.

figures.py

This python script was used to create the article figures. It includes:

• eigenmap and magnitude projection on the hippocampus or subcortex mesh
• graph with the variance explained (need to run variance_explained.m first)
• cohort and task magnitude comparison (need to run compute_zdiff.m first)
• eigenmap projection on the cortex (need to run compute_cortical_projection.m first)