Controlling Virtual Environments by Thoughts

A brain-computer interface (BCI) is a new communication channel between the human brain and a computer. BCIs have been developed during the last years for people with severe disabilities to improve their quality of life. Applications of BCI systems comprise the restoration of movements, communication and environmental contro. However, recently BCI applications have been also used in different research areas e.g. in the field of virtual reality. For the BCI experiments 2 bipolar EEG derivations where mounted on the subject's head (electrode positions C3 and Cz). The electrodes were connected to a portable amplifier and digitization unit. A Pocket PC was used to control the experimental paradigm for the BCI training of the subject. The subject had to imagine a foot movement and a right hand movement 80 times each. Then the EEG data was analyzed in order to distinguish the 2 different imaginations. After the initial training three subjects all with classification accuracy above 80% were participating in an experiment in a highly-immersive CAVE like, virtual reality (VR) system. We have used a virtual street populated by 16 avatars and shops on both sides of the street. The BCI output signal was transmitted to the VR system in order to navigate in the VE. The goal was to reach the end of the street (see Figure 1). The subject was instructed by an acoustic cue to imagine a foot movement (double beep) or a right hand movement (single beep). If the foot movement was classified correctly the subject was moving forward, otherwise the subject was remaining on the same position. If a right hand movement was correctly detected the subject was also remaining on the same position otherwise as a punishment the subject was moving backwards. Therefore only with a 100% BCI classification accuracy the subject was able to reach the end of the street. The accuracy was determined as achieved cumulative mileage and measured how far the subject could move. S1 had a performance of 63.6%, S2 of 78.9% and S3 of 85.4%. The work showed that motor imagery can be used as input signal for a BCI system to control a VE in a highly immersive CAVE system. Subjects reported about an exciting experience of moving forward and backward just by the imagination of different types of movements.