Predicting psychotic disorder with machine learning

Overview

This research project deals with predicting psychosis disorder using machine learning based on different MRI metrics and is conducted in the MSc05 course in the Neurocognitive Psychology lab at Goethe University Frankfurt within the psychology master.

The aim of the project is to use different brain modality data, ranging from macro-structural MRI, micro-structural MRI, fMRI, functional MRI connectivity and network analysis to structural connectivity and DWI tractography to predict which of the brain modalities provide more accuracy in classifying participants either as control (and therefore healthy) or patient with psychotic disorder.

The data at hand originates from the paper "Functional MRI connectivity accurately distinguishes cases with psychotic disorders from healthy controls, based on cortical features associated with brain network development" by Morgan et al. (2020) and is available on figshare. The github repository for the study can be accessed here.

Structure and content

The theoretical background and derivation of the research question can be found in the content/general_information folder. Exploration and analysis of the dataset can be found in the content/code folder and the respective data in the content/data folder.

The project is also available as a Jupyter Book and can be accessed here. In the content/static folder, graphics and images used in the Jupyter Book are stored.

If something remains unclear about the repository, please open an issue on github.