Automatic Removal of Ocular Artifacts from Electroencephalogram Using Hilbert-Huang Transform
Yan Long Wang, Jin Hua Liu, Yuan Chun Liu · 2008
Hilbert-Huang transform (HHT) is a time-frequency analysis method, which extract intrinsic mode functions (IMFs) that admit well-behaved Hilbert transforms from the analyzed signals using empirical mode decomposition. With the Hilbert transform, the IMF yields meaningful instantaneous frequencies as functions of time. This paper presents a signal processing technique for automatic removal of ocular artifacts from Electroencephalogram (EEG) based on HHT. First, EEG contaminated by ocular artifacts was decomposed IMFs. Then the instantaneous frequencies of each IMF were computed respectively. If all instantaneous frequencies of an IMF are less than 3 Hz, then the value of the IMF is set to zero. If all instantaneous frequencies of an IMF are less than 16 Hz, then the value of IMF, which is greater than a threshold, is set to zero. Computing the sum of all IMFs having been processed, we can obtain corrected EEG. The experimental results show that this method, which remove most ocular artifacts in EEG and distort EEG slightly, is effective.