This repository contains code examples corresponding from the paper Protocol Update: The Normative Modelling Paradigm for Computational Psychiatry.
The notebooks include the main normative modelling workflow, longitudinal modelling, model comparison, data synthesis, data harmonization, federated learning workflows (model transfer, model extension, model merging) and a real-life scenario where we apply model transfer using pre-trained models from Rutherford et al. (2022)
Note that you need to run 1_main_workflow.ipynb first to get the trained model used by the other notebooks. The exceptions are 2.1_longitudinal_modelling.ipynb and 4.1_model_transfer_lifespan_reference.ipynb, which use their own data.