Real-time classification of activated brain areas for fMRI-based human-brain-interfaces

Functional MR imaging (fMRI) enables to detect different activated brain areas according to the performed tasks. However, data are usually evaluated after the experiment, which prohibits intra-experiment optimization or more sophisticated applications such as biofeedback experiments. Using a human-brain-interface (HBI), subjects are able to communicate with external programs, e.g. to navigate through virtual scenes, or to experience and modify their own brain activation. These applications require the real-time analysis and classification of activated brain areas. Our paper presents first results of different strategies for real-time pattern analysis and classification realized within a flexible experiment control system that enables the volunteers to move through a 3D virtual scene in real-time using finger tapping tasks, and alternatively only thought-based tasks.