A BCI using VEP for continuous control of a mobile robot

Area of Science:

• Neuroscience
• Robotics
• Human-Computer Interaction

Background:

• Brain-computer interfaces (BCI) translate neural activity into device commands.
• Visual evoked potentials (VEP) are commonly used for BCI, with steady-state (SSVEP) and code-modulated (c-VEP) approaches offering high information transfer rates (ITR).

Purpose of the Study:

• To evaluate the applicability of SSVEP and c-VEP for continuous robotic control.
• To determine which VEP method provides superior performance in a robotic steering task.

Main Methods:

• Eleven healthy subjects controlled a robot along a track using a computer-based BCI system.
• Two VEP methods, SSVEP and c-VEP, were employed for control, with subjects receiving visual feedback.
• Performance was assessed by task completion time and online classification accuracy.

Main Results:

• The c-VEP method resulted in significantly faster task completion times (222.57 s) compared to SSVEP (573.43 s).
• Maximum online classification accuracy was higher for c-VEP (98.18 %) than for SSVEP (91.36 %).

Conclusions:

• Code-modulated VEP is better suited for continuous robotic control applications than SSVEP.
• The findings suggest c-VEP enhances the usability of brain-computer interfaces for real-time device manipulation.