Effect of Channel and Reference Selection on a Non-occipital Steady-State Visual Evoked Potential-Based Brain-Computer Interface

The requirements of occipital electroencephalogram signals restrict the practicality of steady-state visual evoked potential (SSVEP) brain-computer interfaces (BCI). This study evaluated the feasibility and performance of using non-occipital signals to identify SSVEPs. Channel and reference selections were investigated to determine the optimal non-occipital channel combinations and reference electrode. The spectra and signal-to-noise ratio results indicated that signals from temporal and central regions are available for non-occipital SSVEP BCI. The target identification results confirmed that appropriate combinations of non-occipital channels and reference can achieve sufficient accuracy of approximately 80%.