Quantifying brain-functional dynamics using deep dynamical systems: Technical considerations

Jiarui Chen; Anastasia Benedyk; Alexander Moldavski; Heike Tost; Andreas Meyer-Lindenberg; Urs Braun; Daniel Durstewitz; Georgia Koppe; Emanuel Schwarz

doi:10.1016/j.isci.2024.110545

Quantifying brain-functional dynamics using deep dynamical systems: Technical considerations

iScience. 2024 Jul 22;27(8):110545. doi: 10.1016/j.isci.2024.110545. eCollection 2024 Aug 16.

Authors

Jiarui Chen^{1

2}, Anastasia Benedyk^{2

3}, Alexander Moldavski^{2

3}, Heike Tost^{2

3}, Andreas Meyer-Lindenberg^{1

2

3}, Urs Braun^{1

2

3}, Daniel Durstewitz^{4

5

6}, Georgia Koppe^{1

2

4

5}, Emanuel Schwarz^{1

2

3}

Affiliations

¹ Hector Institute for Artificial Intelligence in Psychiatry, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, M7, 68161 Mannheim, Germany.
² Department of Psychiatry and Psychotherapy, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, J5, 68159 Mannheim, Germany.
³ German Center for Mental Health (DZPG), Partner Site Mannheim, 68159 Mannheim, Germany.
⁴ Department of Theoretical Neuroscience, Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, J5, 68159 Mannheim, Germany.
⁵ Interdisciplinary Center for Scientific Computing, Heidelberg University, J5, 68159 Mannheim, Germany.
⁶ Faculty of Physics and Astronomy, Heidelberg University, J5, 68159 Mannheim, Germany.

Abstract

Both mental health and mental illness unfold in complex and unpredictable ways. Novel artificial intelligence approaches from the area of dynamical systems reconstruction can characterize such dynamics and help understand the underlying brain mechanisms, which can also be used as potential biomarkers. However, applying deep learning to model dynamical systems at the individual level must overcome numerous computational challenges to be reproducible and clinically useful. In this study, we performed an extensive analysis of these challenges using generative modeling of brain dynamics from fMRI data as an example and demonstrated their impact on classifying patients with schizophrenia and major depression. This study highlights the tendency of deep learning models to identify functionally unique solutions during parameter optimization, which severely impacts the reproducibility of downstream predictions. We hope this study guides the future development of individual-level generative models and similar machine learning approaches aimed at identifying reproducible biomarkers of mental illness.

Keywords: Artificial intelligence; Psychiatry.