Automatic Seizure Detection in Rats Using Laplacian EEG and Verification with Human Seizure Signals

• Published in: Annals of biomedical engineering 2013-03, Vol.41 (3), p.645-654
• Main Authors: Feltane, Amal, Faye Boudreaux-Bartels, G., Besio, Walter
• Format: Article
• Language: English
• Subjects: Algorithms; Animals; Biochemistry; Biological and Medical Physics; Biomedical and Life Sciences; Biomedical Engineering and Bioengineering; Biomedicine; Biophysics; Classical Mechanics; Diagnosis, Computer-Assisted - methods; Diagnosis, Computer-Assisted - statistics & numerical data; Electroencephalography - methods; Electroencephalography - statistics & numerical data; Humans; Male; Rats; Rats, Sprague-Dawley; Seizures - diagnosis; Signal Processing, Computer-Assisted; Support Vector Machine
• ISSN: 0090-6964; 1573-9686; 1521-6047
• DOI: 10.1007/s10439-012-0675-4

Automated detection of seizures is still a challenging problem. This study presents an approach to detect seizure segments in Laplacian electroencephalography (tEEG) recorded from rats using the tripolar concentric ring electrode (TCRE) configuration. Three features, namely, median absolute deviation, approximate entropy, and maximum singular value were calculated and used as inputs into two different classifiers: support vector machines and adaptive boosting. The relative performance of the extracted features on TCRE tEEG was examined. Results are obtained with an overall accuracy between 84.81 and 96.51%. In addition to using TCRE tEEG data, the seizure detection algorithm was also applied to the recorded EEG signals from Andrzejak et al . database to show the efficiency of the proposed method for seizure detection.