Summary of Federal Project Grant Award

The National Institute on Deafness and Other Communication Disorders (NIDCD) awarded a $132,695 Project Grant to the University of Maryland, College Park, effective July 1, 2026, through June 30, 2028, under the Research Related to Deafness and Communication Disorders program (CFDA 93.173). The award supports research into automated vocal tract gesture identification and biofeedback systems to enhance speech sound disorder treatment in children. The project develops artificial intelligence (AI) capabilities to deliver intensive, best-practice speech therapy by inferring child vocal tract gestures from audio recordings using speech inversion technology and machine learning techniques, requiring only a microphone for clinical implementation.

The research delivers three primary technical products: (1) an acoustic-to-articulatory speech inversion system trained on child speech data to accurately infer vocal tract gestures from audio (Aim 1); (2) a machine learning classification system that identifies vocal tract gesture error subtypes to generate automated verbal feedback for AI-driven therapy delivery (Aim 2); and (3) personalized visual biofeedback displays using animated tongue representations derived from speech inversion estimates to guide vocal tract gesture error correction (Aim 3). This work aims to lower the lifelong clinical burden of childhood speech sound disorders by automating intensive biofeedback interventions without increasing demands on speech-language pathology clinicians.