Semi-supervised adaptation in ssvep-based brain-computer interface using tri-training

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Science

Background:

• Brain-computer interfaces (BCIs) enable communication and control through brain signals.
• Steady-state visual evoked potentials (SSVEPs) are a common BCI modality.
• Semi-supervised learning offers a way to improve BCI performance with limited labeled data.

Purpose of the Study:

• To present a novel, computationally simple, semi-supervised SSVEP-based BCI.
• To evaluate the effectiveness of tri-training in enhancing BCI accuracy.

Main Methods:

• Implemented a tri-training based semi-supervised learning framework.
• Utilized autocorrelation-based features and a Naïve-Bayes classifier (NBC).
• Employed a system with nine characters, a 100 Hz monitor, three scalp electrodes, and two PCs.

Main Results:

• Preliminary tests on nine healthy subjects demonstrated improved accuracy with tri-training.
• The tri-training approach enhanced the performance of the SSVEP-BCI.

Conclusions:

• Tri-training is an effective method for improving SSVEP-BCI accuracy.
• The proposed system offers a computationally simple and effective solution for SSVEP-BCI.