Code-Modulated Visual Potentials for Fast and Flexible BCI

We explore a new BCI design for the control of robotic devices. Specifically, we show the first use of a code-modulating, Visually-Evoked Potential (cVEP)-based BCI for a navigation and control task. The few studies, which have been done so far using cVEP, show high information-transfer rates and low latencies in spelling systems compared to other EEG-based BCIs. Still, there is no previous work on transferring the cVEP to applications beyond spelling, such as control of robotic devices. We initially study three exploratory experiments where we investigate the stability of the cVEP potential in different conditions. Afterwards, the findings inform the development of a complete BCI using cVEP for the control of a devices. For the evaluation subjects have to control a virtual agent in a kitchen task using the cVEP-BCI system. We are able to achieve an average accuracy above 80% and an average latency of about 2 s per command. All but one subjects were able to control the agent and complete the task.

Related People

Riechmann, HannesContact

Finke, AndreaContact

Ritter, HelgeContact