Functional connectivity analysis in magnetoencephalography

The anatomical structure of the human brain provides ample evidence for the existence of specialized local neural assemblies that are hierarchically organized at different spatial scales. The pathways connecting local networks form the communication channels that allow for highly dynamic interactions between the specialized networks. The proper functioning of the human brain is associated with high computational demands that can only be met by robust, flexible, and efficient communication within and between these networks. These constraints force the multitude of single neurons to dynamically form functional assemblies the core computational and representational units. Neurons may identify themselves as members of an assembly by synchronization of their filing pattern. The often oscillatory pattern of assembly dynamics governs the interactions with other assemblies. A number of measures have been introduced to identify and characterize these interactions among which coherence and phase synchronization are most commonly used. Recently, the necessity of performing neuromagnetic connectivity studies based on brain areas rather than at the sensor level has been recognized and evoked rapid methodological developments. These developments in combination with techniques from signal processing, system identification, and statistics offer the exciting possibility of studying dynamically neural communication processes under physiological and pathological conditions.