A Novel EEG Feature Extraction Method Using Hjorth Parameter 论文

摘要

When processing electroencephalography (EEG) signals in motor imagery case, it is essential to analyze them in both time and frequency domains. An EEG signal has a non-stationary property and its frequency feature also differs from individual to individual. Thus, we can infer that each subject has one’s own dominant timing and frequency band for extracting distinguishable features. Based on this inference, after analyzing EEG signals with the Hjorth parameter, we select the principal frequency band and the timing using the Fisher ratio of the Hjorth parameter. By doing these, the performance of the feature extraction in EEG-based BCI systems was improved in terms of the classification accuracy by 4.4% on average. 1