Efficient Sensor Configuration for Accelerometer-Based Silent Speech Interfaces: Regionally Distributed Sensor Placement Strategy Considering Articulatory Dynamics

Silent speech interfaces (SSIs) decode articulatory activity without audible speech. Although accelerometer-based SSIs can capture facial and neck movements in compact wearable setups, accelerometer-specific sensor placement has not been systematically examined. This study evaluated all combinations of ten accelerometers across the lower face and neck in a 100 English word silent speech classification task with 20 participants. Subject-dependent classification was performed using a convolutional neural network followed by a bidirectional long short-term memory network. Accuracy improved as the number of sensors increased, reaching a plateau with five sensors. At this point, Sensors 1 and 7, positioned over the upper and lower labial regions, Sensors 4 and 8, located bilaterally along the buccal-mandibular region, and Sensor 5, placed over the left submental region beneath the chin, achieved the highest accuracy of 92.30 ± 4.64% and were selected as the efficient configuration. These five sensors consistently ranked among the most informative locations in standalone performance, perturbation-based explainability, and sensor ablation analyses, indicating that they contributed unique and influential articulatory information. Regional coverage comparisons showed that configurations spanning broader functional articulatory regions outperformed those with more restricted coverage. This placement strategy was further validated using an independent 100-word vocabulary, demonstrating that broad regional coverage remained advantageous despite slight changes in the optimal sensor identities. A practical three-sensor configuration based on this strategy achieved approximately 90% recognition accuracy while substantially reducing hardware complexity. Together, these results support regionally distributed sensor placement as a design strategy for capturing complementary articulatory information in compact accelerometer-based SSI systems.