Data and Code for Longitudinal Network Temperature Analyses of Depressive Symptoms, Perceived Stress, and Stress-Related Biomarkers
Description
This repository contains the data and R code used for the longitudinal network-temperature analyses of depressive symptoms and perceived stress across six assessment waves. The primary analyses estimate network temperature for the PHQ-9 and PSS using multigroup Ising models and examine the robustness of longitudinal temperature trajectories using available-case, complete-longitudinal-case, and inverse-probability-weighted analyses. The repository also includes code for examining physiological correspondence with network temperature. Hair LC–MS/MS measures include cortisol and cortisone, with additional analyses of the cortisol-to-cortisone ratio. Hair lipid measures include anandamide (AEA), 2-arachidonoylglycerol (2-AG/AG), palmitoylethanolamide (PEA), oleoylethanolamide (OEA), and stearoylethanolamide (SEA). Analyses include wave-level Pearson and Spearman correlations, adjacent-wave change correlations, and individual-level linear mixed-effects models separating within-person and between-person variation. Exploratory analyses further examine within-system and cross-system biomarker ratios, including ratios among endocannabinoids and related N-acylethanolamines and ratios between these measures and glucocorticoids. The code also includes covariate-adjusted sensitivity analyses and scripts for generating the corresponding summary tables and longitudinal trajectory figures.