A four-stage localization method for epileptic seizure onset zones*. Issue 1 (July 2017)

• Record Type: Journal Article
• Authors: Wan, Ting; Wu, Min; Lai, Xuzhi; Wan, Xiongbo; She, Jinhua; Du, Yuxiao
• Is Part Of: IFAC-PapersOnLine. Volume 50:Issue 1(2017)
• Page Start: 4412
• Page End: 4417
• Publication Date: 2017-07
• Subjects: Epilepsy -- Seizure onset zone (SOZ) -- High-frequency oscillation (HFO) -- Four-stage localization method; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2017.08.915
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: This paper presents a four-stage localization method for epileptic seizure onset zones (SOZs). This method combines the advantages of both the Shannon-entropy-based complex Morlet wavelet transform (SE-CMWT) and adaptive-genetic-algorithm-based matching pursuit (AGA-MP) methods to achieve satisfactory localization performance. The Hilbert transform method is used to identify events of interest. The SE-CMWT-based power spectral density method is applied to acquire channels of interest (CoIs). By using the Morlet wavelets as the dictionary elements, the AGA-MP method combined with Morlet wavelets is developed to detect high-frequency oscillations (HFOs) on CoIs with faster speed and higher accuracy than the AGA-MP method. And the characteristics of HFOs are used to localize SOZs based on the half-maximum method. A comparison shows that our method has higher sensitivity and specificity than four existing methods.