Dynamic Causal Models for Human Electrophysiology: EEG, MEG, and LFPs

Abstract

Dynamic causal modeling (DCM) is a Bayesian model inversion and selection framework for identifying the neurobiological mechanisms that generate electroencephalographic (EEG), magnetoencephalographic (MEG), and local field potential (LFP) recordings. The principal idea is to use generative models – based upon interconnected neural populations – to explain observed data, where these models embody competing hypotheses about how the data were generated. Dynamic causal models are parameterized in terms of effective connectivity or synaptic time constants and entail particular forms of neuronal dynamics that can be assessed using Bayesian model selection and parameter estimation.