Voice onset time contrasts in the speech of profoundly hearing impaired children

Children with profound hearing impairments typically err in the production of the voiced voiceless contrast in stops. In order to examine how speech training and experience with various types of prosthetic devices influences these contrasts, four groups of children were recorded producing words beginning with voiced and voiceless stops. Each group was comprised of a conventional hearing aid subject, a tactile aid subject, and a cochlear implant subject. Recordings were made before the children were fitted with either the tactile aid or cochlear implant and 2 years later. Measurements of voice onset time (VOT) were made from the waveforms. Preliminary results from one group of children revealed the voiced and voiceless distinction was correctly produced for bilabial and alveolar stops at both dates. For the velar stops, no child produced a VOT contrast at the first recording date. The child fitted with a cochlear implant produced longer VOTs for voiced velar stops than for voiceless ones. After 2 years experience with the devices, all children demonstrated normal VOT contrasts for velar stops. Data from the remaining three groups will be presented and discussed relative to prosthetic device and experience with the devices. [Work supported by the NIH.]