摘要
目的利用一种多重分形去趋势波动分析方法探索脑电信号的波动和不规则信息中隐藏的规律。方法数据来源于德国波恩大学医学中心的数据库,分别对睁眼、闭眼、正常人、癫痫未发作期和发作期患者的脑电信号进行计算,求出各自的奇异谱宽度△α和Hurst指数,并对各组数据的均值和方差做统计分析。结果奇异谱宽度△α可以有效区分不同的脑电状态。结论脑电信号的多重分形去趋势波动分析方法有助于癫痫疾病的早期诊断。
Objective To explore the significant information deeply hidden in the irregular electroencephalography (EEG) signals with muhifractaldetrended fluctuation analysis. Methods Multifractal singular spectrum and Hurst index of EEG signals in different groups were calculated based on the data from the database in Medical Center of University of Bonn. Results Muhifractal singular spectrum discriminated different states of EEG effectively. Conclusions Multifractaldetrended fluctuation analysis is helpful to make an early diagnosis for epilepsy.
出处
《北京生物医学工程》
2013年第3期226-229,共4页
Beijing Biomedical Engineering
基金
中央高校基本科研业务费(QN201108)
江苏省自然科学基金(BK2011565)资助