This repository contains all of the code used in the analyses described in the manuscript "The positive-negative mode link between brain connectivity, demographics, and behavior: A pre-registered replication of Smith et al. (2015)" currently in press at Royal Society Open Science.

Software used in the code include:

• Matlab
• Connectome Workbench 1.4.2
• FSL version 6.0.1
• ICA+FIX version 1.06.15
• MATLAB 2017b (MCR v93)
• MATLAB 2020a (MCR v98)
• R version 4.0.0
• Python version 3.8.1
• FSLNets 0.6.3
• PALM (Alpha version 116)
• PWLING v1.2
• libsvm-3.24
• L1precision

All software was run on the NIH High Performance Computing Cluster where all nodes run CentOS 7 and Slurm is used for job scheduling

Installing

1. Clone the repo
2. Install the conda environment

conda env create -f environment.yml

Data needed from the NIMH Data Archive

• The NDA RDS file for ABCD Release 2.0.1 is needed here: https://nda.nih.gov/study.html?id=796
• The BIDS formatted release of the ABCD data available in Collection 3165

Running the pipeline

1. Run /create_config.sh/ and provide the absolute paths to:
   • the main abcd_bids folder (On the NIH HPC, /data/ABCD_MBDU/abcd_bids/bids/)
   • the NDA RDS file
   • the location of the ABCD data reprocessed with the DCAN pipeline
   • the absolute path to the conda environment install of python
2. Navigate to abcd_cca_replication/data_prep/ and run the scripts in this order (and follow the intermediate instructions provided by each script:
   • prep_stage_0.sh
   • prep_stage_1.sh
   • prep_stage_2.sh
   • prep_stage_3.sh
   • prep_stage_4.sh

Run Notes:

1. Stage 0 can take up to 24 hours to run completely.
2. When running FSL's dual_regression, jobs will need at least 32gb memory