Computerized seizure detection based on heart rate changes
O’Donovan C, Burgess R, Lüders H, and Turnbull J (1995) Computerized seizure detection based on heart rate changes. Epilepsia p. 7.
Abstract: Objective: To assess the usefulness of a computerized seizure detection method based on increases in patients heart rate during video-EEG monitoring (VEEG). Background: Computerized seizure detection methods currently used are based on changes in the EEG. These methods are affected by absence of clear ictal EEG patterns in some seizures, similarity of some physiological EEG changes to seizure patterns and difficulties due to artefacts. lctal tachycardia occurs frequently in partial seizures. Methods: The study included 20 patients undergoing inpatient VEEG where seizure detection by trained personnel was done by continuous observation of the patients together with EEG. Baseline heart rate values were calculated daily from the previous 24 hours. Computer detections were triggered when the patients heart rate was elevated for 15 seconds above the mean baseline heart rate plus two standard deviations. Two minutes of VEEG were saved with detections and subsequently analyzed for seizure activity. Results: The total number of events detected was 63, of which 32 (50%) were partial seizures. Of the partial seizures detected. 64% were detected by both computer method (CM)and trained personnel (TP). 27% were detected by the CM only, and 6% being detected by TP alone. Auras (n=4) and non-epileptic events(n=15) were not detected by the computer. False positives occurred on average at a rate of one every 10 hours. Conc1usion:Heart rate elevation is a useful parameter for use in computer seizure detection.This method may serve as an excellent complement to continuous observation by trained personnel. and result in a significant number of additional seizures being detected during VEEG especially when limited personnel are available.