Performance Scaling for Structural MRI Surface Parcellations: A Machine Learning Analysis in the ABCD Study
This github repository contains READMEs dedicated primarily to explaining the actual usage and structure of this project's code - a separate dedicated project website can be found at https://sahahn.github.io/parc_scaling/ which acts as an online supplementary materials for the paper.
This project is setup with a few different directories, which if necessary contain their own more detailed READMEs.
-
This folder contains the code for processing, analyzing and plotting the results.
-
This configuration file is used across the project to specify different key shared parameters.
-
data/
This folder stores the processed data as generated from setup.
-
docs/
This include the markdown pages for the project website.
-
exp/
This folder includes all of the code used to perform the ML expiriments. See exp/ README.md for more details.
-
This folder contains various misc. notebooks for making brain figures and other misc. side-analyses.
-
This folder is where the random parcellations used for the multiple parcellation strategies are stored. This folder is created and filled by a script in setup.
-
This folder contains all of the processed and numpy saved versions of main parcellations used in this project.
-
raw/
This folder contains all of the raw data, parcellations and input data etc..., used in the project. See raw/ README.md for more details.
-
This folder contains the code used to setup the rest of the expiriments, including processing input data and parcellations for later ML. See setup/ README.md for more details.
Before cloning this repository (if you desire to run any of the analysis scripts) make sure you have git-lfs installed first, as some large files are hosted this way. Next after cloning i.e., git clone git clone https://github.com/sahahn/parc_scaling then make sure to tar extract the results. This can be done by first navigating to subfolder exp/, then running command tar -xf results.tar.gz.
