Spatio-temporal analysis of error-related brain activity in active and passive brain-computer interfaces

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalography (EEG)-based brain-computer interface (BCI) systems are crucial for inferring brain signals non-invasively.
• User brain responses to BCI errors introduce non-stationarity in EEG signals, challenging active BCI control.
• Detecting error-related brain activity is a primary goal for passive BCIs and essential for reliable active BCIs.

Purpose of the Study:

• To propose and evaluate CREST, a novel covariance-based method for detecting error-related brain activity in BCI systems.
• To combine spatial and temporal aspects of feedback-related brain activity using Riemannian and Euclidean geometry.
• To improve the reliability of both active and passive BCI control through enhanced error detection.

Main Methods:

• Developed CREST, a covariance-based method utilizing Riemannian and Euclidean geometry.
• Integrated spatial and temporal features of brain activity related to BCI error feedback.
• Evaluated CREST on two datasets: active BCI (1-D cursor control, motor imagery) and passive BCI (2-D cursor control).

Main Results:

• CREST demonstrated significant improvements in detecting error-related brain activity across participants in both active and passive BCI datasets.
• The proposed method showed superior performance compared to existing techniques for EEG signal analysis in BCI applications.
• Enhanced detection of error signals contributed to more reliable BCI control.

Conclusions:

• CREST offers a robust and effective approach for detecting error-related brain activity in EEG-based BCI systems.
• The method's ability to combine spatial and temporal information improves the reliability of BCI control.
• This advancement has implications for both active BCI development and passive BCI applications requiring precise error signal identification.