Application of Hilbert-Huang transform for the study of motor imagery tasks.

A motor based Brain-Computer Interface (BCI) translates the subject's motor intention into a control signal by means of the method which extracts characteristic feature from EEG recorded from the scalp. In this paper, the EEG signal recorded during three motor imagery tasks, which were imagination of left hand, right hand and foot movements, was investigated. A novel method named Hilbert-Huang transform (HHT) is introduced to extract the feature from signal. Firstly, raw signal is decomposed using Empirical Mode Decomposition (EMD). And then, several Intrinsic Mode Functions (IMF) are gained. For further study, the IMFs whose main frequency is higher than 5 Hz are selected. Secondly, based on the IMFs selected above, Hilbert spectrum is calculated. In each motor imagery task, local instantaneous energies, within specific frequency band of electrode C3 and C4, are selected as the features. A three-layer BP Neural Network classifier is structured for pattern classification. The classification results show that HHT can be used in EEG-based BCI research as a method to analysis non-linear and non-stationary signal.