In this project, you will work on expanding an existing package for EEG analysis and machine learning (https://github.com/onedeeper/EEGLearn) to incorporate multiscale entropy and connectivity features, and benchmark these features on an established EEG dataset for machine learning related to psychiatry.

Machine Learning algorithms than can be explored are traditional classifiers, as well as Graph Neural Networks, with or without Self-Supervised pre-training

Supervision will be provided by members of the BOOST consortium (dr. van Wingerden, dr. Pöttkamper) and the author of EEGlearn, Udesh (UMC Amsterdam).

required proficiency:

intermediate python

version control (or willing to learn)

interest in software engineering principles of reproducibility