Disputation: Jonas Verhellen

PhD candidate Jonas Verhellen at the Department of Biosciences will be defending the thesis "Computational Diversity in Neuroscience and Drug Design: Developing Novel Computational Tools Supporting The Study and Treatment of Severe Mental Disorders" for the degree of PhD.

Profile picture of Jonas Verhellen

Jonas Verhellen: private.

The trial lecture is: "AI in science. Should it be explainable?".

Time and place: June 16, 2023 10:15 AM, Zoom and Nucleus, Bikuben, The Kristine Bonnevie building.

The events will also be live streamed using Zoom. The host of the session will moderate the technicalities while the chair of the defence will moderate the events.

The events opens for participation just before they start, and closes for new participants approximately 15 minutes after it has begun.

Click here to join the events

Main research findings

The study explores the molecular basis of severe mental disorders like schizophrenia and bipolar disorder and emphasizes the role of computational simulations in advancing their treatment. The dissertation presents two novel computational tools for generating drug-like molecules and one novel technique accelerating the simulation of biophysically-detailed neuron models. These tools leverage recent advancements in quality-diversity algorithms and multi-task deep learning.

Current pharmacological treatments for severe mental disorders are limited in efficacy, leaving room for the development of novel medications. Deep learning algorithms have been explored for generating drug molecules, but their performance hasn't surpassed classical genetic algorithms. The dissertation improves upon genetic algorithms by incorporating chemical diversity through quality-diversity techniques borrowed from soft robotics.

To enable larger and better simulations of biophysically-detailed neuron networks, deep learning is employed to distill the complex equations governing neuron models into artificial neural networks. This approach accelerates network simulations significantly. Our work aims to extend predictions to all compartments of a neuron model using multi-task deep learning architectures. Future research in this area is expected to further advance our understanding and utilization of computational tools in EEG analysis.

Adjudication committee

Professor Jan Halborg Jensen, Copenhagen University

Associate Professor Benjamin Adric Dunn, NTNU

Associate Professor Jonas Paulsen, University of Oslo

Chair of defence

Professor Cinzia Anita Maria Progida, University of Oslo

Supervisors

Professor Gaute T. Einevoll, Department of Physics, University of Oslo and Norwegian University of Life Sciences (NBMU) / CINPLA

Professor Marianne Hafting Fyhn, Department of Biosciences / CINPLA

 

Published June 2, 2023 9:54 AM - Last modified June 2, 2023 9:54 AM