A novel 3D region-based speller paradigm for BCI systems
Scientific Reports, cilt.16, sa.1, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 16 Sayı: 1
- Basım Tarihi: 2026
- Doi Numarası: 10.1038/s41598-026-49989-9
- Dergi Adı: Scientific Reports
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, BIOSIS, Chemical Abstracts Core, EMBASE, MEDLINE, Directory of Open Access Journals, Zoological Record, Academic Search Ultimate (EBSCO), Natural Science Collection (ProQuest), Biological Science Database (ProQuest), Biomedical Reference Collection: Corporate Edition (EBSCO), Health Research Premium Collection (ProQuest)
- Anahtar Kelimeler: 3D, Brain-computer interface, EEG, Event-related potentials, Paradigm, Region based
- Atatürk Üniversitesi Adresli: Evet
Özet
This study proposes and evaluates a novel three-dimensional region-based (3D-RB) speller paradigm designed to enhance classification performance. EEG data were recorded from 15 participants using 32 channels. Classification accuracy was examined across both single electrodes and predefined electrode groups. Subject-dependent analyses revealed that electrodes located in the parietal and occipital regions (e.g., Pz, P7, P8, O1, O2, Oz) achieved the highest single-channel accuracies (approximately 80–85%), whereas central electrodes (e.g., Cz, C3, C4) yielded lower accuracies (around 70–73%). Electrode grouping provided a distinct advantage; for most participants, Group 4 (Parietal + Occipital) and Group 5 (Parietal + Occipital + Central) achieved the highest performance, reaching nearly 99% accuracy. Notably, despite including fewer electrodes, Group 4 performed nearly as well as Group 5, underscoring the practical benefit of optimized electrode selection. Subject-independent (LOSO) analyses showed similar trends. Among single electrodes, P7, P8, O1, and O2 achieved the highest accuracies (approximately 78–79%), while central electrodes (e.g., Cz, Cp1, Cp2, C3, C4) remained lower (70–73%). Electrode groups again outperformed single channels, with Group 4 and Group 5 reaching approximately 89–91% accuracy. The comparable performance of Group 4, despite fewer electrodes, highlights its practical advantage for real-world applications. Grand Average ERP analyses indicated that differences between target and non-target stimuli primarily emerged within early and mid-latency time windows, with these effects being more pronounced over parietal and occipital regions. Taken together, these findings demonstrate that incorporating three-dimensional visual effects within a region-based paradigm significantly enhances classification performance by leveraging parietal–occipital activity. The proposed 3D-RB paradigm therefore offers an efficient and user-friendly approach for future BCI speller designs.