Mind over virtual matter: using virtual environments for neurofeedback training

This paper describes on-going research at Lancaster University to develop a brain-computer interface (BCI) with which to conduct neurofeedback training. We have built a system that translates EEG signals detected from the scalp of a subject into movement and interaction within a VRML world. The training protocol parameters can be set prior to a session commencing. These correspond to signal thresholds within which a subject will be rewarded for maintaining his or her EEG component signal amplitude for a predetermined period. The training environments are constructed from a set of VRML components. Interactivity parameters, in terms of VRML object appearance and behaviour corresponding to changes in the EEG signal, can be chosen to suit the requirements of the session.