Dual stream neural networks for brain signal classification

Area of Science:

• Neuroscience
• Machine Learning
• Biomedical Engineering

Background:

• Brain-computer interfaces (BCIs) require robust classification of neuroimaging signals.
• Existing methods may lack adaptability or subject independence.

Purpose of the Study:

• To develop a novel neural network classifier for functional neuroimaging signals in BCIs.
• To create a subject-independent classifier applicable to various signal types.

Main Methods:

• Proposed a dual stream neural network (DSNN) architecture.
• The first stream processes raw time-dependent signals for feature identification.
• The second stream integrates dynamic functional connectivity for enhanced multi-channel information.

Main Results:

• DSNN performance was benchmarked against three public datasets.
• Achieved performance comparable to or exceeding state-of-the-art results.
• Information-theoretic analysis provided insights into network signal parsing.

Conclusions:

• The DSNN offers a versatile and high-performing solution for BCI signal classification.
• The architecture is subject-independent and adaptable to different neuroimaging data.
• Enables deeper understanding of neural network processing of biological signals.