This repository includes feature preparation part, for ABIDE-I resting state fMRI data.
The repo includes three files:
- 1. DownloadData.py : just an automated script to download ABIDE-I preprocessed, for some strategy and atlases specified.
- 2. CalcCorr.ipynb : used to convert raw files to csv files, for easier pandas-based analysis. The file creates:
- 2.1. raw time signals per area, similar as input but csv but assigns each brain names to each column instead of numberic labels)
- 2.2. functional connectivity matrix, calculated with pearson coefficient, for each couple of brain regions. The correlation is labeled by area names accrdingly
- 2.3. Dist feature matrix, calculated as the euclidean distance between each two brain signals. the correlation is labeled by area names accrdingly.
- 2.1. raw time signals per area, similar as input but csv but assigns each brain names to each column instead of numberic labels)
- 3. CreateDynamicFeature.py dynamic functional connectivity, calculated with for a sliding windowed signals. window type (gaussian), width, and sliding step can be controlled. The output is the percantage of strong correlations, as well as weak/no correlation per pair of areas
This code was used in a recently submitted journal paper, Nov22.
-Please use citation the provided citation if used.
--For any questions, suggestions, feel free to contact me.