This repository contains analysis scripts related to the manuscript Interindividual variability and lateralization of μ-opioid receptors in the human brain by Tatu Kantonen, Tomi Karjalainen, Janne Isojärvi, Pirjo Nuutila, Jouni Tuisku, Juha Rinne, Jarmo Hietala, Valtteri Kaasinen, Kari Kalliokoski, Harry Scheinin, Jussi Hirvonen, Aki Vehtari, and Lauri Nummenmaa.
The PET data used in the manuscript were preprocessed and modeled with minimal user-intervention using Magia (https://github.com/tkkarjal/magia), an automated analysis pipeline for brain PET studies. Magia preprint is available from here.
- fit_and_compare_models_roi.R
- Fits seven different models to the ROI-level data and compares their predictive accuracy using Bayesian 10-fold cross-validation. Model fitting is done using brms (https://paul-buerkner.github.io/brms/), and the cross-validation is done using the R package LOO (https://CRAN.R-project.org/package=loo)
- build_roi_specific_posteriors.R
- Draws samples from the posterior distribution of Model 4 and sums them up to get ROI-specific posteriors for age, sex, and smoking
- visualize_results_roi.R
- Visualizes the results from the ROI-level analysis
- nonlinear_models.R
- Fits the nonlinear models
- create_clusters.m
- Takes in i) high-resolution mean BP-image over subjects, and ii) an anatomical ROI atlas of same resolution to produce functionally (BP-wise) homogenous clusters within clearly defined anatomical boundaries.
- fit_model_full.R
- Fits the Model 4 to the full-brain data (again, with brms)
- build_cluster_specific_posteriors.R
- Draws samples from the posterior distribution of Model 4 and sums them up to get cluster-specific posteriors for age, sex, and smoking
- visualize_cluster_posteriors.R
- Visualizes the cluster-specific posterior distributions
- fit_model_asymmetry.R
- Fits the model and builds the posterior distributions of interest