Epileptic Seizure Identification based on EEG Rhythm Decomposition

摘要：

　　Discrimination of epileptic seizure can be carried out based on the four rhythms (δ,θ,α,and β) or frequency bands characterizing each patient’s clinical condition.Since the concept of correct interpretation and physiological or clinical meaning becomes critical,there is the need for finding the frequency bands providing more information within epileptic seizure framework.The present study develops a rhythmic component extraction method,for extracting a signal oscillation modes at a certain frequency.The feature set evaluates power spectral density by using the responses of two parametric models.Attained outcomes for k-nn classifier over 500 epilepsy records,reach a performed accuracy as high as 100%.The advantage of the proposed methodology is the rhythm interpretation suitable for detection of epileptic seizures.

关键词： Rhythm Decomposition epileptic seizure Time-varying autoregressive model Exponentially damped sinusoidal model

作者: L. Duque-Mu(n)oz J. Espinoza-Oviedo C. G. Castellanos-Domínguez

作者单位: Universidad Nacional de Colombia

会议类型: 国际会议

会议名称: World Congress on Medical Physics and Biomedical Engineering (2012年医学物理及生物医学工程国际会议(IFMBE))

会议地点: 北京

会议语种:英文

页码: 363-366

在线出版日期: 2012-05-26（万方平台首次上网日期，不代表论文的发表时间）