Alertness-based subject-dependent and subject-independent filter optimization for improving classification efficiency...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Brain-computer interface (BCI) systems are crucial for neural prostheses and medical rehabilitation.
• High mental workload in BCI tasks hinders user manipulation.
• Electroencephalography (EEG) signals offer insights into both BCI control and user mental states.

Purpose of the Study:

• To propose and evaluate novel methods for identifying non-effective trials in Steady State Visual Evoked Potential (SSVEP)-based BCI systems.
• To enhance the performance and reliability of SSVEP-BCI by filtering out non-effective trials.

Main Methods:

• Utilized subject-dependent and subject-independent alertness models to identify non-effective trials.
• Applied these models to SSVEP-BCI systems for performance evaluation.

Main Results:

• The subject-dependent alertness model demonstrated significant improvements in classification accuracy.
• The subject-independent alertness model enhanced the predictive capabilities of the SSVEP-BCI system.
• Both models showed effectiveness in filtering non-effective trials.

Conclusions:

• The proposed alertness models, particularly the subject-dependent approach, offer valuable improvements over conventional methods like canonical correlation analysis (CCA).
• These methods enhance precision, contributing to the technical advancement of BCI technologies.
• Demonstrated the practical effectiveness of subject-dependent and subject-independent alertness models in SSVEP-BCI systems.