(go to slide show)

ANALYSIS OF COHERENCE IN MULTIPLE EPILEPTIC SEIZURES

Piotr J. Franaszczuk, Gregory K. Bergey
 

Department of Neurology, Johns Hopkins University School of Medicine and Epilepsy Center,
Johns Hopkins Hospital , Baltimore, Maryland U.S.A.
In a given patient, seizures originating from the same focus often produce very similar EEG recordings, particularly when
intracranial recordings of multiple seizures are compared. This allows for improvement in statistical properties of estimators of
coherence and directed transfer function (DTF) computed from an autoregressive model.

Intracranial recordings of multiple seizures (five or more) from three patients were analyzed.  Cross-correlations for each
pair of seizures were computed and used to determine the common time scale.

The peak value of the cross-correlation function was also used to determine the level of similarity between seizures.  Only
seizures with a peak value of 0.7 or more were further analyzed.  Each seizure was treated as a realization of the same stochastic
process during subsequent multichannel autoregressive model analyses. The ordinary, partial, and multiple coherences, as well as
directed transfer functions (DTF), were computed using all available data.

Using more than one seizure to compute coherences and DTFs allows for analysis of shorter quasi-stationary epochs, thus
providing better time resolution in analysis of fast changing seizure intracranial EEG recordings.   This may also allow these methods
to be applied to greater portions of the seizure recordings.
 

(Supported by NIH grant NS 33732)


(go to slide show)
(go back)