In order to explore the correlation between the adjacent segments of a long term EEG, an improved principal component analysis(PCA) method based on mutual information algorithm is proposed. A one-dimension EEG time se...In order to explore the correlation between the adjacent segments of a long term EEG, an improved principal component analysis(PCA) method based on mutual information algorithm is proposed. A one-dimension EEG time series is divided equally into many segments, so that each segment can be regarded as an independent variables and multi-segmented EEG can be expressed as a data matrix. Then, we substitute mutual information matrix for covariance matrix in PCA and conduct the relevance analysis of segmented EEG. The experimental results show that the contribution rate of first principal component(FPC) of segmented EEG is more larger than others, which can effectively reflect the difference of epileptic EEG and normal EEG with the change of segment number. In addition, the evolution of FPC conduce to identify the time-segment locations of abnormal dynamic processes of brain activities,these conclusions are helpful for the clinical analysis of EEG.展开更多
A new wavelet variance analysis method based on window function is proposed to investigate the dynamical features of electroencephalogram(EEG).The exprienmental results show that the wavelet energy of epileptic EEGs a...A new wavelet variance analysis method based on window function is proposed to investigate the dynamical features of electroencephalogram(EEG).The exprienmental results show that the wavelet energy of epileptic EEGs are more discrete than normal EEGs, and the variation of wavelet variance is different between epileptic and normal EEGs with the increase of time-window width. Furthermore, it is found that the wavelet subband entropy (WSE) of the epileptic EEGs are lower than the normal EEGs.展开更多
基金Natural Science Foundatoin of Fujian Province of Chinagrant number:2010J01210,2012J01280
文摘In order to explore the correlation between the adjacent segments of a long term EEG, an improved principal component analysis(PCA) method based on mutual information algorithm is proposed. A one-dimension EEG time series is divided equally into many segments, so that each segment can be regarded as an independent variables and multi-segmented EEG can be expressed as a data matrix. Then, we substitute mutual information matrix for covariance matrix in PCA and conduct the relevance analysis of segmented EEG. The experimental results show that the contribution rate of first principal component(FPC) of segmented EEG is more larger than others, which can effectively reflect the difference of epileptic EEG and normal EEG with the change of segment number. In addition, the evolution of FPC conduce to identify the time-segment locations of abnormal dynamic processes of brain activities,these conclusions are helpful for the clinical analysis of EEG.
基金Natural Science Foundatoin of Fujian Province of Chinagrant number:2012J01280
文摘A new wavelet variance analysis method based on window function is proposed to investigate the dynamical features of electroencephalogram(EEG).The exprienmental results show that the wavelet energy of epileptic EEGs are more discrete than normal EEGs, and the variation of wavelet variance is different between epileptic and normal EEGs with the increase of time-window width. Furthermore, it is found that the wavelet subband entropy (WSE) of the epileptic EEGs are lower than the normal EEGs.
