We will use artificial neural networks to develop an acoustic-based system for automatically identifying stuttering in continuous speech. The system will use these networks, already developed for purposes of speech recognition, to identify differences between perceptually indistinguishable authentic and simulated stutterings.Awardee's statement of the potential commercial applications of the research:An automatic neural network demonstrated to identify stuttering events in chronic stutterers would have, potentially, worldwide application for treatment and research on a disorder that occurs in 1% of all populations.National Institute on Deafness and Other Communication Disorders (NIDCD)
