Evaluating brain-computer interface performance using color in the P300 checkerboard speller

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Current Brain-Computer Interface (BCI) systems predominantly use grey-to-white (GW) visual stimuli.
• Optimizing BCI performance and user experience is crucial for broader adoption.

Purpose of the Study:

• To evaluate BCI performance using uniquely colored stimuli.
• To compare grey-to-color (GC) and color intensification (CI) stimuli against standard GW stimuli.
• To test hypotheses regarding increased BCI performance, enhanced event-related potentials (ERPs), and improved offline performance with unique color stimuli.

Main Methods:

• A checkerboard paradigm was employed to present three types of visual stimuli: grey-to-white (GW), grey-to-color (GC), and color intensification (CI).
• Online BCI performance metrics, including accuracy and information transfer rate, were recorded for 36 participants.
• Offline performance was analyzed using an electrode selection algorithm (Jumpwise).
• Event-related potentials (ERPs) were analyzed to assess neural responses.

Main Results:

• The grey-to-color (GC) stimuli condition yielded higher accuracy and information transfer rates compared to CI and GW conditions.
• GC stimuli elicited significantly higher amplitude ERPs than both CI and GW stimuli.
• The Jumpwise-selected electrode locations improved offline performance across all stimulus conditions.

Conclusions:

• Unique color stimuli, specifically GC, demonstrably enhance BCI performance and neural responses (ERPs).
• The use of color stimuli is preferred by participants, indicating potential for increased end-user acceptance.
• Optimized electrode selection algorithms further improve offline BCI performance, complementing advancements in stimulus design.