Multi-Modal Integration of EEG-fNIRS for Brain-Computer Interfaces - Current Limitations and Future Directions

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Multi-modal integration of neurophysiological signals enhances reliability by combining complementary features from different modalities.
• Electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) integration presents a cost-effective and portable solution for brain-computer interface (BCI) development.
• Current EEG/fNIRS integration in BCI shows limited performance gains due to challenges in feature fusion and spatial registration.

Purpose of the Study:

• To conduct a comprehensive literature review on EEG/fNIRS integration for BCI applications.
• To identify and discuss the current limitations hindering the performance of integrated EEG/fNIRS BCI systems.
• To propose future research directions for optimizing multi-modal integration in BCI.

Main Methods:

• Systematic review of existing literature on EEG/fNIRS integration in BCI.
• Analysis of studies focusing on feature extraction and fusion techniques for multi-modal neurophysiological data.
• Evaluation of challenges related to sensor placement and signal synchronization in combined EEG/fNIRS systems.

Main Results:

• The integration of EEG and fNIRS shows potential but has not yet achieved significant BCI performance enhancement.
• Key limitations include inadequate methods for integrating disparate EEG and fNIRS features and potential mismatches in recording locations.
• Existing studies highlight the need for advanced signal processing and co-registration techniques.

Conclusions:

• Further research is required to develop sophisticated approaches for integrating EEG and fNIRS data to overcome current performance bottlenecks.
• Addressing feature-level fusion and spatial alignment is crucial for unlocking the full potential of EEG/fNIRS in BCI.
• Future work should focus on novel integration strategies and standardized methodologies for multi-modal BCI development.