Fast and robust fixed-point algorithms for independent component analysis
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
Addressing the problems existing in traditional phase locking value (PLV) calculating method, an improved method based on Hilbert-Huang transform (HHT) is proposed and applied onto a group of hypoxia EEG in this paper. Improved method preprocesses EEG data by ocular artifact elimination and spatial filtering firstly. Then, EEG sub-components withinα frequency band (8-12Hz) are obtained through empirical mode decomposition (EMD) of HHT and chosen as our research object. Finally, Hilbert transform (HT) is applied onto the target EEG sub-components and PLVs among different channels of EEG records are calculated according to the transform result. According to extracted PLVs used as features, normal and hypoxia EEG recorded from 3 subjects can be distinguished effectively. Primarily analysis shows that improved method has potential to be used widely to analyze EEG.