Speech understanding in adults and children for sentences with cochlear-implant simulated spectral degradation and shallow insertion depth

Abstract

While previous studies have compared cochlear-implant simulated (i.e., vocoded) speech understanding between adults and children, they have been acute experiments that have not considered long-term adaptation or training effects. Normal-hearing adult listeners demonstrate significant improvement in vocoded speech understanding with training, particularly if there is simulated frequency-to-place mismatch. The purpose of this study was to compare how adults and children adapt to eight-channel sine-vocoded speech with 0, 3, and 6 mm of frequency-to-place mismatch. Twenty adults (>18 yrs) and ten children (8–10 yrs) were trained on vocoded speech understanding over a four-hour period. The stimuli were a closed set of 40 simple words from which five-word nonsense sentences were constructed. Speech understanding was measured in 45-trial blocks where no feedback was provided, followed by 30-trial blocks where visual and auditory feedback was provided. High variability existed within both groups. On average, children performed worse than adults. Over the first five testing/training blocks, children improved at a slower rate than adults. Some children showed minimal improvement over the testing, whereas most of the adults showed noticeable improvement. These results suggest that adults have developed neural mechanisms that can more effectively adapt to and process degraded and frequency-shifted speech.