Fuzzy Entropy based Complexity Analysis for Target Classification during Hybrid BCI Paradigm

Electroencephalography (EEG) is one of the most widely used noninvasive system in the field of brain-computer interfacing (BCI). Visual evoked potentials (VEPs) are the efficient BCI techniques designed to detect target/non-target events through brain responses. Fuzzy based entropy measures have received increased attention in analyzing the complex multichannel EEG signals. Although, fuzzy entropy performs robustly compared to non-fuzzy methods, it does not examine the time series signals over multiple temporal scales, which is crucial for multivariate signals. This study proposed an empirical mode decomposition (EMD) featured fuzzy entropy by coarse-graining the time-series signal at a multi-scale level (EMFuzzyEn) to increase the performance of the BCI during hybrid steady state visual evoked potential (SSVEP) and rapid serial visual presentation (RSVP) BCI paradigm. The results showed that the EMFuzzyEn features achieved significantly higher classification performance of 89 ± 1% for 9 channel combination and 87 ± 2% for 2 channel combination. Further, the EMFuzzyEn also showed superior performance when compared to our published event related potential (ERP) based BCI technique and popular non-fuzzy entropy algorithms. Overall, the results demonstrated that EMFuzzyEn algorithm enhances the discrimination between target and non-target events efficiently by evaluating their complexity differences thereby improving the classification performance and can be a potential indicator to measure the BCI performance.