Detection of Mental Task Related Activity in NIRS-BCI systems Using Dirichlet Energy over Graphs.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Near Infrared Spectroscopy (NIRS)-based Brain Computer Interfaces (NIRS-BCI) typically analyze mean concentration changes and slopes of hemodynamic responses.
• Spatial patterns across NIRS channels are often overlooked but crucial for reliable mental task detection.
• Existing methods may not fully leverage the spatial information inherent in NIRS data.

Purpose of the Study:

• To introduce a novel measure for NIRS-BCI that incorporates spatial features.
• To enhance the robustness and reliability of mental task detection in NIRS-BCI.
• To integrate information from multiple NIRS channels for improved performance.

Main Methods:

• The study proposes using the Dirichlet Energy of NIRS signals defined over a graph.
• This measure quantifies spatial patterns and integrates activity across multiple NIRS channels.
• The method was applied and validated on a real-world NIRS dataset.

Main Results:

• The proposed Dirichlet Energy measure effectively captures spatial NIRS features.
• Integration of multi-channel NIRS activity using this measure leads to robust mental task detection.
• Experimental results on a real dataset confirm the efficiency of the proposed approach.

Conclusions:

• The Dirichlet Energy measure offers a promising approach for NIRS-BCI by incorporating spatial information.
• This method enhances the detection of mental task-related brain activity.
• The findings suggest improved NIRS-BCI performance through the analysis of spatial NIRS features.