Brain-computer interaction research at the Computer Vision and Multimedia Laboratory, University of Geneva

Area of Science:

• Neuroscience
• Computer Science
• Human-Computer Interaction

Background:

• Brain-computer interaction (BCI) research often focuses on rehabilitation.
• Augmenting classical human-computer interaction with physiological measurements is an emerging area.
• Current techniques for electroencephalogram (EEG) analysis have known limitations.

Purpose of the Study:

• To advance brain-computer interaction (BCI) for multimodal interaction, augmenting classical interfaces with physiological data.
• To develop robust methods for brain source activity recognition from EEG.
• To investigate BCI protocols and performance using information theory and assess user emotional status via physiological signals.

Main Methods:

• Iterative robust stochastic reconstruction procedures were employed for brain source activity recognition, modeling source and noise statistics.
• Procedures for optimal electroencephalogram (EEG) sensor system design (electrode placement and number) were developed.
• Information-theoretic measures were used to compare various BCI protocols and assess performance.

Main Results:

• Novel methods for EEG-based brain source activity recognition were established, overcoming limitations of deterministic approaches.
• Optimal EEG sensor configurations were designed for improved signal acquisition.
• Information-theoretic analysis provided insights into BCI protocol efficiency and capabilities for emotional state assessment.

Conclusions:

• The study demonstrates the potential of BCI and physiological signals to enhance multimodal interaction beyond rehabilitation.
• Advanced signal processing and information-theoretic approaches can significantly improve BCI performance and applications.
• EEG and other physiological signals offer a viable pathway for real-time assessment of user emotional status.