Propagation of Electrical Activity: Nonlinear Associations and Time Delays between EEG Signals

Abstract

To assess how dynamic patterns of activity are established in the brain, the recording of EEG or MEG signals is important, since these signals give information about the activity of brain regions on a millisecond basis. Such assessment is of practical importance in a number of conditions, both in healthy subjects, for example, when the activity of brain regions during cognitive tasks or motor acts is investigated, and in patients, for example, when the propagation of epileptiform activity from a focus is to be determined. Here we present and discuss a methodology that enables us to analyze the propagation of electrical activity from one brain structure to other structures on the basis of intracranial EEG recordings. There are two problems with this methodology. The first problem is to determine whether the sources that generate the EEG signals are related to each other or are independently active. The second problem is to determine, in the case of related signals (or sources), whether the activity of one source occurs in a fixed time sequence in relation to that of the other (in which case a delay may be expected between the signals due to the finite propagation velocity of electrical activity in the brain).If a signal from a source is related to another signal and yet an unknown delay exists between them, any method of analysis used to determine the relationship between the signals (sources) ought to take into account not only the (unknown) delay but also the type of relationship. These two aspects are interrelated. Moreover, to obtain a good understanding of the process that take place in the brain during propagation of electrical activity, knowledge of both the strenght of relationship and of the delay between the signals (sources) is necessary. This point is elaborated upon in the present article. We developed a method by means of which relationships between signals can be estimated with a minimum number of restricted conditions on the nature of such a relationship.KeywordsElectrical ActivityAssociation FunctionRegression CurveTransmission DelayEpileptiform ActivityThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.