Toward asynchronous EEG-based BCI: Detecting imagined words segments in continuous EEG signals

An asynchronous Brain--Computer Interface (BCI) based on imagined speech is a tool that allows to control an external device or emit message at the moment user desires by decoding EEG signals of speech. In order correctly implement these types BCI, we must be able detect from continuous signal, when subject starts imagine words. this work, five methods feature extraction wavelet decomposition, empirical mode frequency energies, fractal dimension and chaos theory features are presented solve task detecting words segments as preliminary study for latter implementation BCI These tested in three datasets using four different classifiers higher F1 scores obtained 0.73, 0.79, 0.68 each dataset, respectively. This results promising build system automatizes segmentation classification.