Classification of BCI Users Based on Cognition

Area of Science:

• Neuroscience and Human-Computer Interaction
• Physiological Computing and Cognitive State Monitoring

Background:

• Brain-Computer Interfaces (BCI) were initially designed to aid individuals with paralysis by translating brain activity into computer commands.
• Assessing cognitive states objectively, easily, and interpretably remains a challenge in human-computer interaction research.

Purpose of the Study:

• To develop and validate an objective, user-friendly method for cognitive state assessment using Brain-Computer Interface (BCI) systems.
• To classify cognitive states as low-risk or high-risk based on physiological and performance data.

Main Methods:

• Seventy healthy participants performed six tasks while their pupil dilation, blink rate, and Galvanic Skin Response (GSR) were recorded.
• Cognitive state clusters were generated using the K-means algorithm, integrating physiological data and task performance metrics.
• Machine learning models (Logistic Regression, Decision Tree, Neural Networks) were employed to validate the classification consistency.

Main Results:

• A novel method utilizing BCI systems successfully classified participants' cognitive states into low-risk or high-risk categories.
• The developed classification method demonstrated high consistency, ranging from 87.1% to 100%, when compared with established machine learning techniques.
• Physiological signals (pupil dilation, blink rate, GSR) and task performance provided a robust basis for cognitive state assessment.

Conclusions:

• The proposed BCI-based method offers an objective and interpretable approach to cognitive state assessment.
• This technique has potential applications in various fields requiring real-time monitoring of cognitive load and performance.
• Further research can explore the application of this method in diverse populations and more complex task environments.