Young deaf children with hearing aids and/or cochlear implants (CIs) have reduced access to speech information needed to acquire language and subsequently to develop normal reading skills. Cued Speech (CS) is a manual system designed to disambiguate lipread speech information. Cued Speech could also supplement the speech information available through a hearing aid or Cl. The long-term goal of the proposed research is to realize a computer-based automatic and comprehensive CS training system. Experiments will test the fundamental hypothesis that coarticulation of cues is a basis for the effectiveness of CS. Three additional research questions will be tested: (1) Whether the proposed system is effective in training; (2) Whether participants perceptually integrate synthetic CS with their lipreading (and residual hearing), and whether cue learning transfers to untrained words; and (3) Whether the proposed system has potential for Cl users to disambiguate auditory input. Ultimately, if successful, the use of this system would contribute to development of children's reading proficiency, which depends on phonological representations obtained via speech perception. Even without CS later in life, the phonological representations and language skills enhanced via CS will continue to be used. Another clinical implication is that parents who opt not to obtain a Cl for their children can have an effective, efficient, and affordable alternative to assist their children during the critical language development period. The proposed research synergistically integrates engineering and human perception research and is designed to be carried forward if successful. Project 1 has the goal of engineering an automatic American English CS synthesis system. To synthesize CS manual coarticulation, a concatenative synthesis method will be adopted, for which the transitions between sequential hand gestures are video-captured, extracted, and stored in short segments. Project 2 is to evaluate the accuracy and effectiveness of the proposed system for conveying CS information to augment speech with four potential user groups: adults who learned CS as deaf children, normal-hearing adults, Cl adults, and Cl children. This research is relevant to public health: It will contribute new knowledge about CS production and perception; it will help facilitate Cl recipients' auditory habilitation; and it will improve language acquisition, reading ability, psychosocial development, and quality of life among deaf and hard-of-hearing children. ? ? ?
