Virtual Reality for Brain Computer Interface
Visualization of the brain activity is one of the key components for BCI-design. Since it is well-known that visual input has a strong impact on motor cortex activity and influences the EEG patterns, it is interesting to find out which type of visual feedback can improve the performance of a BCI system.
Recently, we have combined Virtual Reality (VR) with the BCI system. With VR we can create "realistic" artificial (virtual) environments by means of 3D-computer graphic. The user can immerse themselves into the virtual world, move around and change items inside the virtual scene. It is like playing a video game, but with the difference that every action is doing by thought, without moving any muscle!
Virtual flat
We have created the 3D-model of a flat with several rooms and corridor. We have also predefined the path where one can walk through. This should prevent that the user go though the wall or go out of the flat.
After the start the user is moving to the next cross way. The forward movement is done automatically. At each cross way the user has two choices. Two arrows are shown, indicating the possible directions to go further. Now, the user has to produce control signal (by imagine of body part movements like left hand or right hand motions) to choose the direction.
Currently, virtual flat is used for cue-based BCI training. The user has to imagine body part movements only when the arrows appear. We have inserted some flags into predefined rooms. The user was instructed to get the flags by moving into these rooms.
Virtual park
For the asynchronous training paradigm we have built a virtual park with several hedges. In contrast to the virtual flat, there are no predefined path. The forward movement can be done by imaging foot movements. Left and right rotation are done by focusing left and right hand movements, respectivly. The user can reach any point of the lawn unless he is stopped by a hedge or is going to leave the park.
To scoring the training performance we add some items into the scene. The user should pick up all items where he can choose arbitrary way to the items.
Document Actions
