摘要
立体定向脑电(stereo-EEG,sEEG)的癫痫间期高频振荡(High Frequency Oscillations,HFOs)与癫痫灶高度相关,广泛用于难治性癫痫切除术前定位中,但HFOs易与高频伪迹等混淆,自动辨识精度低,临床上仍依赖人工辨识,长程sEEG数据量巨大,人工辨识耗时费力易出错,急需HFOs高精度自动识别方法。考虑sEEG具有非线性、非平稳以及多维sEEG之间具有一致相关性等特点,本文提出基于最小二乘-多维经验模态分解(Least Square-Multivariate Empirical Mode Decomposition,LS-MEMD)的HFOs快速自动识别方法。本文基于临床1680段HFOs和1720段高频伪迹测试了该算法的性能,且与小波变换、经验模态分解等方法比较,证明了所提方法具有更高的准确率和更低的误检率。
High Frequency Oscillations(HFOs)detected by stereo-EEG(sEEG)during interictal periods are highly correlated with epileptogenesis,and are widely used in the localization of refractory epilepsy before surgery.However,the detection of HFOs are challenging because they are easily confused with high frequency artifacts.Manual detection of HFOs with long-term sEEG recordings are time consuming,labor intensive and error prone.An automatic and accurate approach for HFO detection is urgently needed.Considering that sEEG has the characteristic of nonlinear,non-stationary,and there is uniform correlation between multi-dimensional sEEG,we propose a fast automatic HFOs recognition method based on Least-Square-Multivariate Empirical Mode Decomposition(LS-MEMD).The method is tested on1680 clinical HFO data and 1720 high-frequency artifacts.Compared with other state of the art methods,e.g.Wavelet Transformation,EMD and Ensemble EMD,LS-MEMD gives higher accuracy and lower false alarm rate for the detection of HFOs.
作者
刘燕
周渊峰
胡莹
郎恂
张龑囧
郑潜
张丽
汤继宏
戴亚康
LIU Yan;ZHOU Yuanfeng;HU Ying;LANG Xun;ZHANG Yanjiong;ZHENG Qian;ZHANG Li;TANG Jihong;DAI Yakang(Suzhou Institute of Biomedical Engineering and Technology,Chinese Academy of Sciences,Suzhou,Jiangsu 215163,China;Children's Hospital,Fudan University,Shanghai 200000,China;Children's Hospital,Soochow University,Suzhou,Jiangsu 215000,China;Yunnan University,Kunming,Yunnan 650000,China;Zhejiang University,Hangzhou,Zhejiang 310000,China;Suzhou Key Laboratory of Medical and Health Information Technology,Suzhou,Jiangsu 215163,China;Jinan Guoke Medical Engineering Technology Development Company,Ltd.,Jinan,Shandong 250000,China;Suzhou Guoke Health Medical Technology Company,Ltd.,Suzhou,Jiangsu 215163,China;Nanjing Brain Hospital,Nanjing,Jiangsu 210000,China)
出处
《中国体视学与图像分析》
2020年第2期183-191,共9页
Chinese Journal of Stereology and Image Analysis
基金
国家自然科学基金(No.61801476,61971413)
江苏省自然科学基金(No.BK20170387,BK20180221)
江苏省重点研发项目(No.BE2018610)
江苏省博士后项目(No.2018K030A)
苏州市科技计划项目(No.SS201866)
苏州市重点实验室(No.SZS201818)
济南市高校20条项目(No.2018GXRC017)和济南市5150人才项目。
关键词
立体脑电
高频痫样振荡
最小二乘
多维经验模态分解
stereo-EEG
high frequency oscillations
least-square
multivariate empirical mode decomposition
