The Think-Tank Seminar Series

Computational Modelling of Mood Disorders

Facilitator

Dr Arno Steinacher

Date and time

12:00 - 1:00pm, 5 October 2012

Abstract

Does Psychology need computational modeling? What is the use of this approach and its benefits as opposed to descriptive statistical models? Having a background in Computational Systems Biology, I would like to highlight some of the common techniques used in dynamical systems theory and prominent examples of their application in biology. As an example for how dynamical modeling could be used in psychological research, a computational model of bipolar disorder will be presented, which has been developed in a collaboration with Kim White. Based on this, I would then like to invite the audience to discuss potential applications in their research and foster future interdisciplinary collaborations.

Background: Dynamical Systems Theory, a multidisciplinary approach stemming from physics and applied mathematics, has over the last years brought forward the highly successful area of Systems Biology in the domain of Biosciences. Aiming at mathematical formulation and simulation of biological entities of interest across many scales, it has become an integrate part of research. Its methods enable us to more rigorously test hypotheses stemming from empirical data collections as well as strengthen the predictions of our hypotheses and suggesting new experiments based on those predictions. Dynamical models are suited well to capture the highly nonlinear architecture of most biological systems and their behaviour over time, such as in the case of gene regulation networks or neuronal networks. They are also critical to understand emergent behaviours and thus to integrate research across different scales.
While increasingly used also in social sciences and economy, this methodological approach seems to be under-represented in psychological research. This is surprising, given the highly dynamic and complex nature of its subjects. While statistical models can make sense of data and their linear correlations, they fail to give us hints on many aspects of complex systems, such as equilibria, tipping points, sensitivity or robustness. In investigating such questions, dynamical modeling can be of immense help.

Literature: For everyone interested in attending this seminar, I would like to suggest the following paper as a good starting point into the subject:

  • Huys QJ, Moutoussis M, Williams J. (2011) Are computational models of any use to psychiatry? Neural Networks 24(6):544-51.