is it possibel to take the TC.mat as inputs and how to undersnad the expectes d csv/.pt formats?

[albertleemon]

Hi all,

Quick question about input preparation. In the paper (DOI: 10.1109/TMI.2025.3617310), my understanding is that the model ultimately uses ROI time series to build dynamic FC graphs via a sliding window (e.g., ROI time-course matrix
P → windowed correlation matrices Xt → thresholded adjacency At).

In this repo, I’m trying to understand how that maps to the expected files:

1. Is it possible to use the TC.mat file generated by pNet as input, if it contains ROI-wise time courses?
2. For the repo’s csv / .pt and pth files:

• Does .pt store the extracted ROI time courses (the P matrix), or does it store derived objects (e.g., windowed FC matrices / graphs)?
• What is the recommended pipeline to go from preprocessed fMRI to the repo’s expected csv/.pt (and pth, if applicable) formats?

Thanks a lot for any clarification or pointers to an example script!
Best regards
Meng