In the context of VR, most systems have used one of three different kinds of EEG signals [85,176,187,188,274]: 1) motor imagery, 2) SSVEP, and 3) P300. The most common is motor imagery, which is a mental process that a person performs before executing an action. During this time, the person rehearses or simulates the motion in the brain, which leads to measurable activations in the primary motor cortex. Users imagine rotating in place or making footsteps to achieve locomotion in the virtual world. Unfortunately, most successful systems are limited to a couple of simple commands, such as starting and stopping walking. Nevertheless, users have been able to explore maze-like environments by simply imagining the motions.

One advantage of motor imagery is that it does require any interference or special stimulus from the system, thereby allowing the user to proceed without disruption or particular timing. The other two kinds of signals unfortunately require a stimulus to be generated, and then the response is measured by EEG. One of them is SSVEP (steady state visually evoked potential), which occurs when a flashing visual stimulus is presented in the range of 3.5 to 75 Hz. The signal-to-noise ratio is very strong for SSVEP, and the user can affect its outcome based on attention paid to the flashing. The decision of whether to pay attention is used as the basis of the command. The other signal is P300, which appears about 300ms after a rare and relevant stimulus is presented. Once again, the response is measured based on how much attention the user pays to the stimulus.

Steven M LaValle 2020-01-06