Dynamic causal modeling (DCM) is a method to non-invasively assess effective connectivity between brain regions. 'Musicogenic epilepsy' is a rare reflex epilepsy syndrome in which seizures can be elicited by musical stimuli and thus represents a unique possibility to investigate complex human brain networks and test connectivity analysis tools. We investigated effective connectivity in a case of musicogenic epilepsy using DCM for fMRI, high-density (hd-) EEG and MEG and validated results with intracranial EEG recordings. A patient with musicogenic seizures was examined using hd-EEG/fMRI and simultaneous '256-channel hd-EEG'/ 'whole head MEG' to characterize the epileptogenic focus and propagation effects using source analysis techniques and DCM. Results were validated with invasive EEG recordings. We recorded one seizure with hd-EEG/fMRI and four auras with hd-EEG/MEG. During the seizures, increases of activity could be observed in the right mesial temporal region as well as bilateral mesial frontal regions. Effective connectivity analysis of fMRI and hd-EEG/MEG indicated that right mesial temporal neuronal activity drives changes in the frontal areas consistently in all three modalities, which was confirmed by the results of invasive EEG recordings. Seizures thus seem to originate in the right mesial temporal lobe and propagate to mesial frontal regions. Using DCM for fMRI, hd-EEG and MEG we were able to correctly localize focus and propagation of epileptic activity and thereby characterize the underlying epileptic network in a patient with musicogenic epilepsy. The concordance between all three functional modalities validated by invasive monitoring is noteworthy, both for epileptic activity spread as well as for effective connectivity analysis in general.© 2015 Elsevier Inc. All rights reserved.
IntroductionMany physiological and pathological processes of the human brain are driven by networks and connectivity between brain regions. Thus, analysis of these connections is of particular importance in neurosciences and several methods have been proposed to study connectivity in-vivo by means of functional imaging. One of these is dynamic causal modeling (DCM), a method to assess the effective connectivity between brain regions, i.e. the causal influence that one neuronal system exerts over others. It was first developed for fMRI with the aim to estimate the parameters of a neuronal system model from which a blood oxygenation level dependent (BOLD) signal can be predicted that corresponds as closely as possible to the measured BOLD time series (Friston et al., 2003). Later, DCM was extended to the analysis of effective connectivity in EEG and MEG data taking into account the very high temporal resolution of these techniques and thereby combining a spatial forward model with a biologically informed temporal forward model (Kiebel et al., 2008). DCM can be used to test which brain region drives which by constructing different models of interacting regions or nodes and identifying the best one by model comparison. Thi...