Timothy Sanders, Helen Carey, Dalit Shalom

The Speech Therapist’s App for /R/ Treatment (START) is a mobile application-based research tool and treatment platform for biofeedback therapy in /r/ misarticulation.

https://files.nyu.edu/tkm214/public/biofeedback.html

Description

Our team is developing a mobile app with the goal of making evidence-based biofeedback treatment freely available to speech-language pathologists worldwide. Recent research strongly supports the hypothesis that treatment for speech sound disorders (SSD) can be enhanced through the incorporation of biofeedback technologies, which create a real-time visual representation of the user’s speech. By viewing real-time visualizations of their speech, individuals learn to manipulate their vocal output to match a personalized visual model representing correct pronunciation. Through practice and accurate, immediate feedback, individuals develop and retain the motor skills necessary for correct pronunciation.

The core functionality of our app focuses on residual misarticulation of the American /r/ sound. The team is developing an LPC-based (linear predictive coding) algorithm to detect and process the resonant frequencies in the speech signal. The American /r/ has the distinctive acoustic property of an average F3 (third formant) value falling at or below the 80% of the mean F3 of the four cardinal vowels as elicited by the same speaker. Once the formant coefficients of a speaker’s vowel and /r/ elicitations are saved into our program, a personalized visual representation of correct /r/ pronunciations can calculated or drawn to the screen. Our app intends to use this process to create an open-source therapeutic tool for biofeedback therapy in speech. Through wide distribution of the app, we also hope to establish a large-scale database of aggregate user inputs with which researchers and clinicians can test hypotheses about speech development and therapeutic effectiveness.

Classes

Ability Lab