Speech Behavior and Science Lab

The Speech Behavior and Science Lab (SBS Lab), based at the Graduate Institute of Linguistics at National Taiwan University, investigates the physiological foundations of speech production, focusing on how speech emerges from coordinated bodily movements and how these processes interface with language and technology.

Our research is organized around three major themes. The first is articulatory dynamics and biomechanics, where we examine how physical constraints shape speech production. Using techniques such as lingual ultrasound imaging, we analyze how articulators—especially the tongue—move and coordinate over time, and how these dynamics give rise to structured speech patterns. The second area is speech motor control, which explores how the nervous system plans and executes speech movements. In particular, we investigate mechanisms of motor preparation and rapid triggering, aiming to explain how complex articulatory gestures can be produced efficiently and reliably. The third area is human–AI speech systems, where we integrate insights from speech physiology into computational modeling and interactive technologies, contributing to the development of biologically grounded speech systems.

Methodologically, our lab adopts a multi-modal and integrative approach. We employ ultrasound imaging to capture real-time tongue motion and extract articulatory trajectories with high temporal resolution. We also use physiological tracking techniques to study coordination across articulators and muscle systems. In addition, we develop physics-based biomechanical simulations using the ArtiSynth platform, which models muscles, soft tissues, and the vocal tract to investigate how biomechanical constraints shape speech behavior. Finally, we explore biofeedback methods, particularly ultrasound-based visual feedback, with applications in speech training, pronunciation learning, and clinical rehabilitation.

Overall, SBS Lab aims to build an integrated framework for understanding speech, bridging articulatory biomechanics, motor control, computational modeling, and applied technologies. Through this interdisciplinary approach, we seek to advance both theoretical insights into speech behavior and practical innovations in language technology and clinical practice.