Nonlinear Seizure Prediction by Surface EEGs
Abstract number :
2.054
Submission category :
Year :
2000
Submission ID :
3213
Source :
www.aesnet.org
Presentation date :
12/2/2000 12:00:00 AM
Published date :
Dec 1, 2000, 06:00 AM
Authors :
Nitza Thomasson, Thomas J Hoeppner, Joseph P Zbilut, Rush Medical Ctr, Chicago, IL.
RATIONALE: It has been reported that onset of seizures may be predicted by nonlinear dimensional and Lyapunov analysis of intracranial EEGs. Besides being invasive, the methods are equivocal due to nonstationarity. A computationally efficient method obviating these difficulties, is recurrence quantification analysis (RQA). METHODS: A Telefactor beehive was used to record 10 seizures. After EEG lead norming, windowed RQA analysis was applied to EEG and ECG signals. RESULTS: All cases showed significant divergences from baseline EEG RQA variables (p<.01) prior to seizures [mean time preceding=14.55s (SD=7.60)]. Changes were also seen in 8 ECGs. In the figure, a sharp divergence occurs at 9.5s (b-EEG) and at 11s (d-ECG) prior to seizure(a-arrow). CONCLUSIONS: RQA appears to detect EEG/ECG changes presaging seizures.