Inner speech classification using inter-trial coherence framework for feature extraction

EEG signal analysis for inner speech recognition has garnered interest in recent years. Participants in these studies are prompted to perform a mental task involving inner speech within a specific time frame. This study utilizes a dataset of inner speech recordings for four words. While the dataset includes recordings from multiple subjects and sessions, this work focuses on the first session of one subject. A novel feature extraction method based on the inter-trial coherence framework is proposed for inner speech classification. Two classifiers, k-nearest neighbors and support vector machines, are trained and evaluated using the extracted features. Results indicate that the proposed method achieves comparable or superior performance compared to previous works. Future work aims to extend the methodology to include the remaining subjects and sessions in the dataset. By doing so, the generalizability and robustness of the method can be further assessed.