A Collaborative Brain-Computer Interface Framework for Enhancing Group Detection Performance of Dynamic Visual...

Area of Science:

• Neuroscience
• Computer Science
• Human-Computer Interaction

Background:

• Collaborative brain-computer interfaces (cBCIs) offer performance advantages for dynamic visual target detection.
• Existing cBCIs lack dynamic information interaction and learning guidance among multiple agents.
• Current multi-mind fusion modes are static and unidirectional, limiting collaborative potential.

Purpose of the Study:

• To propose a novel cBCI framework enhancing group detection performance for dynamic visual targets.
• To develop a mutual learning domain adaptation network (MLDANet) for improved information interaction and learning.
• To overcome the limitations of static, unidirectional multi-mind fusion in existing cBCIs.

Main Methods:

• Developed a Mutual Learning Domain Adaptation Network (MLDANet) as the core of the cBCI framework.
• Integrated information interaction, dynamic learning, and individual transferring abilities within MLDANet.
• Established a dynamic interactive learning mechanism between individual networks and collaborative decision-making at the neural decision level, using P3-sSDA as the base network.

Main Results:

• The MLDANet-cBCI framework achieved superior group detection performance for dynamic visual targets.
• The mutual learning strategy within MLDANet enhanced the detection abilities of individual networks.
• With three collaborators, MLDANet-cBCI showed F1 score improvements of 0.12 for collaborative detection and 0.19 for individual networks compared to multi-classifier cBCIs.

Conclusions:

• The proposed MLDANet-cBCI framework surpasses traditional multi-mind collaborative modes.
• This novel framework demonstrates superior group detection performance for dynamic visual targets.
• The findings hold significant implications for the practical applications of multi-mind collaboration in BCI technology.