The detection of epileptic seizures with high temporal accuracyis still a challenge in the state-of-the-art. The proposedmethod exploits multi-resolution sample entropy for bothseizure detection and fingerprinting. First, a SVM classifieris used to detect the seizures’ onset, then the seizuresfingerprints across the subband structure are derived exploitingsample entropy non stationarity. Over 8 hours of EEGdata recordings from patients suffering from temporal lobeepilepsy were used for training and testing the system, andvalidation was performed based on annotation by one expertneurophysiologist. All the seizures were successfullydetected and the time-scale fingerprinting of their evolutionis obtained from the sample entropy variations in the subbands.A prominent impact in high () frequency band wasobserved whose neurophysiological ground is currently underinvestigation.
Multiscale sample entropy for time resolved epileptic seizure detection and fingerprinting
CONIGLIARO, Davide;MANGANOTTI, Paolo;MENEGAZ, Gloria
2014-01-01
Abstract
The detection of epileptic seizures with high temporal accuracyis still a challenge in the state-of-the-art. The proposedmethod exploits multi-resolution sample entropy for bothseizure detection and fingerprinting. First, a SVM classifieris used to detect the seizures’ onset, then the seizuresfingerprints across the subband structure are derived exploitingsample entropy non stationarity. Over 8 hours of EEGdata recordings from patients suffering from temporal lobeepilepsy were used for training and testing the system, andvalidation was performed based on annotation by one expertneurophysiologist. All the seizures were successfullydetected and the time-scale fingerprinting of their evolutionis obtained from the sample entropy variations in the subbands.A prominent impact in high () frequency band wasobserved whose neurophysiological ground is currently underinvestigation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.