ABCD Sociocultural PLSC Analysis

This repository contains R scripts and Python notebooks that use partial least squares correlation (PLSC) (Krishnan et al., 2011; McIntosh and Lobaugh, 2004) to identify latent, multivariate patterns of associations between the adolescent sociocultural environment (at the individual, family, and community levels) and resting state functional connectivity (rsFC). rsFC was chosen as a neurobiological metric of interest given (i) prior work linking resting state fluctuations and cultural variables (Constante et al., 2023; Luo et al., 2022; Meca et al., 2023), (ii) the known functional relevance of rsFC to self-referential and affiliative processing (Pintos Lobo et al., 2020; Raichle, 2015), and (iii) the flexibility of rsFC to examine the full range of brain systems in absence of external stimuli and/or task demands (Smith et al., 2009).

Additionally, for each significant latent dimension derived from PLSC, we conducted a regression analysis to determine how the identified sociocultural brain patterns may relate to behavioral risk/resilience factors, which will include measures of both physical health (e.g., physical inactivity, weight/obesity, sleep) and mental health (e.g., internalizing/externalizing symptoms).

Repository Structure

Analysis Scripts

1. clean-vars.ipynb

Purpose: Data preprocessing and variable name standardization
Description: This Jupyter notebook loads raw ABCD data files and standardizes variable names for clarity and brevity. It performs two main operations:

• Sociocultural variables: Renames measures to short codes (e.g., meim_ss_exp → EIE-Y, via_ss_hc → HA-Y)
• RSFC variables: Converts network connectivity column names from ABCD format to simplified network pairs (e.g., rsfmri_c_ngd_cgc_ngd_dt → CON-DN)

Outputs:

• derivatives/none-reduced/clean-socio.csv - Cleaned sociocultural data
• derivatives/none-reduced/clean-rsfc.csv - Cleaned RSFC data

2. clean-plsc.Rmd

Purpose: Main PLSC analysis pipeline
Description: This R Markdown document performs the complete PLSC analysis workflow:

• Residualization: Removes effects of covariates (age, sex, scanner site, motion) from both RSFC and sociocultural data using multiple linear regression
• PLSC Analysis: Runs tepPLS to identify latent dimensions linking sociocultural measures to brain connectivity patterns
• Permutation Testing: Tests significance of dimensions using 10,000 permutations via perm4PLSC
• Bootstrap Analysis: Calculates bootstrap ratios to assess variable stability using Boot4PLSC
• Variance Calculation: Computes percent variance explained by each component
• Visualization: Creates scree plots, correlation plots, and bootstrap ratio bar plots

Key Outputs:

• Scree plot showing significant dimensions
• Correlation matrix heatmap
• Bootstrap ratio plots for each significant dimension
• CSV files with loadings and bootstrap ratios for both data blocks
• Latent variable scores (lx, ly)

Required R Packages:

• TExPosition, PTCA4CATA, data4PCCAR - PLSC functionality
• tidyverse, dplyr - Data manipulation
• ggplot2, corrplot, cowplot - Visualization

3. plsc-figures.ipynb

Purpose: Brain visualization and surface plotting
Description: This Python notebook creates publication-ready brain visualizations of significant PLSC results:

• Custom Colormaps: Defines network-specific color schemes for each Gordon network
• Individual Networks: Creates surface plots for each significant network (Dimensions 1 & 3)
• Combined Networks: Generates overlapping network visualizations showing spatial relationships between network pairs
• Composite Figures: Assembles multi-panel figures for dimensions 1 and 3

Key Features:

• Uses surfplot and neuromaps for high-quality surface rendering on fsLR 32k surface
• Transforms Gordon parcellation from volumetric to surface space
• Creates both individual network plots and network pair combinations
• Generates publication-ready figures at 300 DPI

Outputs:

• Individual network surface plots (e.g., CinguloOpercular_surface_plot.png)
• Combined network plots (e.g., CinguloOpercular_Default_combined_surface_plot.png)
• Dimension-specific composite figures (networks-dim1.png, networks-dim3.png)
• All outputs saved to derivatives/none-reduced/figures/

4. plsc-reg-corr.R

Purpose: Regression analysis of PLSC network connectivity and health outcomes
Description: This R script examines relationships between specific network connectivity pairs (identified from PLSC) and physical/mental health variables using linear mixed-effects models (LME). It tests associations while controlling for demographic covariates and accounting for nested data structure (participants within families within sites).

Network Pairs Analyzed:

• CON-DN (Cingulo-Opercular - Default Network)
• DN-DAN (Default - Dorsal Attention Network)
• DN-DN (Default - Default Network)
• DN-VN (Default - Visual Network)
• VN-VN (Visual - Visual Network)

Health Variables Tested:

• Sleep measures (sleep duration, sleep corrected)
• Respiratory health (wheeze, cough, diagnosis, bronchitis)
• Blood pressure (systolic, diastolic)
• Physical activity
• Behavioral problems (internalizing, externalizing symptoms)

Statistical Model:

rsfc ~ health_variable + covariates + (1|site_id_l/rel_family_id)

Key Features:

• Controls for age, sex, parental demographics, income, scanner manufacturer, and head motion
• Accounts for nested structure with random effects for site and family
• Scales all continuous predictors for standardized coefficients
• Loops through all network pairs and health variables systematically

5. plsc-reg-boot.R

Purpose: Regression analysis of PLSC latent dimension scores and health outcomes
Description: Similar to plsc-reg-corr.R, but analyzes the latent variable scores (composite brain-behavior patterns) from PLSC dimensions rather than individual network pairs. Tests how the overall brain-behavior patterns (Dimension 1 and Dimension 3) relate to health outcomes.

Dimensions Analyzed:

• Dimension 1: Primary latent pattern from PLSC
• Dimension 3: Secondary latent pattern from PLSC

Statistical Model:

dimension_score ~ health_variable + covariates + (1|site_id_l/rel_family_id)

Additional Features:

• Creates visualizations for significant associations
• Uses ggplot2 for publication-quality plots
• Focuses on significant variables (e.g., wheeze) for detailed plotting
• Provides holistic view of how brain-behavior patterns relate to health

Comparison with plsc-reg-corr.R:

• plsc-reg-corr.R: Tests specific network connectivity pairs
• plsc-reg-boot.R: Tests overall latent dimension scores (composite patterns)

Environment Setup

Python Environment

For the Python analysis components (including brain surface plotting and visualization), you will need:

• Python 3.9.6
• Gradec: Install first using:

  pip install git+https://github.com/JulioAPeraza/gradec.git

After installing Gradec, install the required neuroimaging packages:

pip install nilearn surfplot neuromaps nibabel matplotlib seaborn pillow

R Environment

For the PLSC analysis in R:

• R version 4.5.1 (or 4.3.x/4.4.x)

• Required packages (install from CRAN):

  install.packages(c("tidyverse", "dplyr", "ggplot2", "corrplot", 
                     "cowplot", "readr", "gridExtra", "psych", 
                     "car", "kableExtra", "ExPosition"))

• GitHub packages (install using devtools):

  library(devtools)
  install_github("HerveAbdi/TExPosition")
  install_github("HerveAbdi/PTCA4CATA")
  install_github("HerveAbdi/data4PCCAR")

Workflow

1. Data Preparation (clean-vars.ipynb)

   • Load raw ABCD data
   • Standardize variable names
   • Save cleaned datasets

2. PLSC Analysis (clean-plsc.Rmd)

   • Residualize covariates
   • Run PLSC with permutation testing
   • Calculate bootstrap ratios
   • Generate statistical visualizations
   • Export latent variable scores and loadings

3. Brain Visualization (plsc-figures.ipynb)

   • Create custom network colormaps
   • Generate individual network surface plots
   • Create network pair combinations
   • Assemble multi-panel figures

4. Health Outcomes Analysis (R scripts)

   • Network-level (plsc-reg-corr.R): Test specific network connectivity pairs against health variables
   • Pattern-level (plsc-reg-boot.R): Test overall PLSC dimension scores against health variables
   • Control for demographic and technical covariates
   • Account for nested data structure (participants within families within sites)

Variables

Variables that were included in the PLSC analysis:

Sociocultural Measures	Gordon Networks
MEIM-R youth	Auditory Network
VIA youth	Cingulo-opercular Network
MACV youth	Cingulo-parietal Network
MEIM-R caregiver	Default Network
VIA caregiver	Dorsal Attention Network
MACV caregiver	Fronto-parietal Network
Perceived discrimination	Retrosplenial Temporal Network
Community cohesion	Sensorimotor Hand Network
LED: Child opportunity index	Sensorimotor Mouth Network
LED: NaNDA disadvantage index	Salience Network
LED: Getis-Ord Gi* Statistic	Ventral Attention Network
	Visual Network