Anqi Xu

This study investigates the neural mechanisms underlying sensorimotor serial learning in speech production, demonstrating that motor compensation plays a more critical role than sensory error correction.

Using multimodal neurophysiological methods, the research reveals how the brain regulates sequential speech motor planning through adaptive motor strategies rather than relying primarily on auditory feedback.

The findings suggest that English diphthongs are more effectively learned and produced when treated as single dynamic targets rather than sequences of two static vowel positions.

Through computational modeling and behavioral experiments, the study demonstrates advantages of dynamic representation in both perception and production of gliding vowels.

This paper demonstrates that artificial vocal learning systems guided by speech recognition can successfully simulate key aspects of early vocal development in children.

Using computational models that integrate articulatory synthesis with deep learning, the study provides insights into how feedback mechanisms might support natural language acquisition.

This research shows that vocal learning extends beyond simple imitation, with computational models revealing how learners develop internal representations that go beyond surface-level copying.

Through simulations using advanced articulatory synthesis, the study demonstrates how vocal learning involves complex mapping between acoustic targets and articulatory gestures.

This paper presents a multimodal approach to artificial vocal learning that combines phoneme recognition with visual information to guide articulatory development.

The integration of auditory and visual feedback streams enables more robust vocal learning, mimicking how human infants use both hearing and visual cues during speech acquisition.

This study identifies that consonantal effects on fundamental frequency (F0) in American English are produced through multiple interacting mechanisms rather than a single phonetic process.

Using empirical analysis of speech production data, the research reveals how different consonant types influence pitch patterns through distinct physiological pathways.

The research reveals specific prosodic strategies used in Hijazi Arabic to encode focus, including distinctive patterns of pitch accentuation and duration modification.

Through cross-linguistic comparison, the study contributes to understanding how focus marking varies across languages while maintaining universal principles of prosodic organization.