| Nowadays, individuals with motor disabilities have a greater recognition among the society. Progress is on track in regards to brain cortical mechanisms, EEG-based communications to acquire neuronal signals, and hardware and software development. Technology either did not exist or was extremely expensive before. Brain-Computer Interface technology has the possibility of helping motor disabled human beings: a BCI system collects the brain electrical activity of a user and processes it to finally transmit it to an external device which will display a feedback. When the user’s retina is excited, a Steady-State Visually Evoked Potential is generated in the brain. This particular electrical signal is largely used in the research area because it can be recorded via EEG, without any prior training of users, and it has an excellent signal-to-noise ratio.The work presented here involves an SSVEP-based BCI system where a user gazes at a flickering dot among others on a computer screen, without any feedback for him or her. Firstly, the LCD-based stimulator used is presented, followed by the experiment scheme and protocol, which involves EEG recording. Afterwards, the processing part is explained: the system pre-processes the SSVEP with a bandpass filter to eliminate noise and then extracts the spectral feature using the Empirical mode decomposition(HHT) or the wavelet transform; feature classification is conducted by Fast Fourier Transform(FFT) or Canonical Correlation Analysis(CCA).Results show that spatial filtering techniques are better at eliminating nuisance signals for multi-channel signals. Still, the influence of inter-subject variation appears in terms of detectability. FFT-based spectra combined with HHT shows that targeted frequencies and their harmonics are detected. Finally, the CCA algorithm combined with the wavelet transform shows high detectability, with best results for high frequencies when targets are bigger, and low frequencies when targets are smaller. |
PDF Full Text Request