engleski

Processing of vocal pitch information in children with cochlear implants and neurocognitive speech studies in newborn and adults

The first part of this work examined gender identification of adult human voices in a population of forty-one children with cochlear implants (CI), and a control population of 15 age-matched hearing peers. An identification experiment tested participants’ abilities to identify the gender of speakers from 2-sec speech excerpts (natural speech segments with fundamental frequencies between 80 Hz and 220 Hz), and a discrimination experiment quantified the participants’ discrimination abilities between genders as a function of the average fundamental frequency difference between pairs of natural speech segments. Control participants all had perfect or near-perfect performance in both tasks. CI users fell into three performance groups. One group (n=18) performed above chance in both experiments, another group (n=18) performed at chance in both experiments, and a third group (n=5) could discriminate the gender of voices even at small fundamental frequency differences in direct comparisons, but were unable to identify the voice gender of single stimuli correctly. The spectral and temporal information provided by the CI device of each participant in response to the specific stimuli used in the experiments were measured empirically, and information was gathered on the hearing history of each subject (age of first deafness diagnosis, age of implantation, duration of CI use). Temporal and place cues were available in all CI devices. Successful performance in the gender identification and discrimination tasks relied on temporal cues, with no evidence for participant utilization of place cues. CI subjects who had a shorter duration of auditory deprivation were better able to perform the long-term recall of gender features required by the identification task, and the age of cochlear implantation was the most important factor related to successful short-term comparison of gender-relevant acoustic features in the discrimination task. The duration of implant use was not a performance factor in either of these tasks in this study population. The second part of this work contains two published papers on gender congruency in speech production in adults, and one published paper on language processing in newborns as revealed by optical topography that the author contributed to while a graduate student at SISSA.

cochlear implant; voice gender; pitch; temporal cues; spectral cues

nije evidentirano