Exploring inter-trial coherence for inner speech classification in EEG-based brain-computer interface.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalogram (EEG)-based brain-computer interfaces (BCIs) offer communication potential for individuals with speech impairments.
• Current EEG-based inner speech classification methods lack sufficient accuracy for practical application.

Purpose of the Study:

• To investigate the efficacy of inter-trial coherence (ITC) as a feature extraction technique for enhancing inner speech classification accuracy in EEG-based BCIs.
• To develop and evaluate a novel methodology employing ITC for improved BCI performance.

Main Methods:

• A novel methodology utilizing ITC within a complex Morlet time-frequency representation for feature extraction was developed.
• EEG recordings of four words from ten subjects across three sessions were analyzed.
• Extracted features were classified using k-nearest neighbors (kNN) and support vector machine (SVM) algorithms.

Main Results:

• The proposed ITC-based methodology achieved average classification accuracies of 56.08% with kNN and 59.55% with SVM.
• These results indicate performance comparable or superior to existing methods in inner speech classification.
• Inter-trial phase coherence demonstrated potential for improving accuracy in EEG-based BCI systems.

Conclusions:

• The study introduces a promising feature extraction method using ITC for EEG-based inner speech classification.
• The findings suggest that ITC can significantly enhance the accuracy of BCI systems for communication.
• This research provides a foundation for future advancements in inner speech decoding technologies.