Standard Article

System Models for Inference on Mechanisms of Neuronal Dynamics

Systems Biology

  1. Klaas E. Stephan1,2,
  2. Karl J. Friston2

Published Online: 15 MAY 2012

DOI: 10.1002/3527600906.mcb.201100042

Reviews in Cell Biology and Molecular Medicine

Reviews in Cell Biology and Molecular Medicine

How to Cite

Stephan, K. E. and Friston, K. J. 2012. System Models for Inference on Mechanisms of Neuronal Dynamics. Reviews in Cell Biology and Molecular Medicine. .

Author Information

  1. 1

    University of Zurich, Laboratory for Social and Neural Systems Research, Department of Economics, Zurich, Switzerland

  2. 2

    University College London, Wellcome Trust Centre for Neuroimaging, Institute of Neurology, London, UK

Publication History

  1. Published Online: 15 MAY 2012


Understanding processes that result from the interaction of multiple elements requires mathematical models of system dynamics. This increasingly important theme has led to the revitalization of “systems biology,” with a focus on molecular questions. However, system models have been used for much longer in computational neuroscience and neuroimaging to infer on causal mechanisms of neuronal dynamics, for example, effective connectivity. In this chapter, system models are reviewed for inferring effective connectivity from brain activity data obtained by functional magnetic resonance imaging or electroencephalography. Following an outline of general systems theory, a generic framework is provided for formalizing the description of systems; this guides the subsequent description of various models of effective connectivity. Attention is focused particularly on dynamic causal modeling (DCM), a Bayesian system identification approach which distinguishes between neuronal state equations and biophysical forward models, as well as methods for Bayesian model selection.


  • Synaptic plasticity;
  • Neuromodulation;
  • Dopamine;
  • Acetylcholine;
  • Dynamic causal modeling (DCM);
  • Granger causality;
  • Psychophysiological interactions;
  • Structural equation modeling;
  • Bayesian model selection