A survey on EMD sensitivity to SNR for EEG feature extraction in BCI application

One of the most complicated biological signals is EEG which its processing has vast applications in Brain-Computer Interface (BCI) Speller is very important. For P300 detection in EEG, miscellaneous Single-Trial and Multi-Trial methods have been introduced, with their own advantages and disadvantages. Our purpose in this paper was to investigate Empirical Mode Decomposition-Dependent (EMD-Dependent) methods of feature extraction sensitivity to Signal to Noise Ratio (SNR) of original signal.