Potentials and pitfalls of connectivity analyses in EEG data

König, Thomas (2 September 2016). Potentials and pitfalls of connectivity analyses in EEG data (Unpublished). In: 18th World Congress of the International Organization of Psychophysiology. Habana, Cuba. 31.08.-04.09.2016.

Using EEG to study brain connectivity seems to be something quite obvious, because EEG results from the transmission of electric signals that neurons use to communicate. At the same time, due to the well-known volume conduction problem, it is a non-trivial modelling task to identify the separate elements of that communication and isolate their dynamics and interdependencies. The EEG literature offers a series of different models that can accomplish such decomposition, at the price of introducing a-priori assumptions. I will focus on two particularly popular models for this type of problems, namely measures of lagged coherence, that putatively quantify causal interactions, and measures of non-lagged connectivity, that putatively quantify non-causal interactions among brain elements. These two approaches have mutually exclusive assumptions about what defines “connectivity”, and entail very complementary advantages and problems. As a result, depending on the choice of methodology or combinations thereof, the conclusions obtained in terms of “connectivity” may greatly vary, or even be meaningless. The author therefore proposes a framework for these different models of connectivity and discusses how they may be integrated to interpret the various observations made.

Item Type:

Conference or Workshop Item (Speech)


04 Faculty of Medicine > University Psychiatric Services > University Hospital of Psychiatry and Psychotherapy > Translational Research Center

UniBE Contributor:

König, Thomas


500 Science > 570 Life sciences; biology




Thomas König

Date Deposited:

09 Nov 2016 12:55

Last Modified:

05 Dec 2022 14:59



Actions (login required)

Edit item Edit item
Provide Feedback