Experimentalphonetische Analyse der Artikulation und Perzeption von Cochlear-Implantat-Trägern
Final Report Abstract
The immediate aim of this project was to study links between speech perception and speech production in cochlear implant speakers (CI) and matched normal-hearing controls (NH). The overall goal was to provide a foundation for understanding the sensorimotor abilities relevant for successful CI outcomes. Acoustic phonetic analysis of vowel, sibilant and consonant cluster production were carried out. Perceptual discrimination tests were carried out in three modalities: auditory, visual and haptic. The CI speaker group (total n = 48) was divided into 4 groups based on speech status at hearing loss and time from hearing loss to implantation. Speech production results: For vowel articulation the differences between CI and NH were fairly subtle. While there was a general tendency for F1 to be lower in CI, in terms of category contrasts vowel height appeared to be better realized than front-back or rounding oppositions. For sibilants very clear evidence for a shift of the main frication energy to lower frequencies was found. This could partly reflect a strategy to compensate for the highfrequency limitations of implants. However, analysis of total sound pressure levels and supplementary ultrasound imaging of the tongue for some speakers indicate that the efficiency of sibilant noise generation is indeed impaired. The analysis of consonant clusters indicated that the durational properties of complex sound sequences are well preserved in CI speakers. Sibilants occurring in clusters are not more distorted than when occurring in singletons. Perceptual tests: These consisted of two existing tests (acoustic and haptic) of discrimination abilities together with an innovative visual test using a continuum based on morphing between two speech stimuli. All tests used as common framework an adaptive staircase testing procedure. Results from the discrimination tests were used to predict speech production abilities, based on the acoustic distance between the sibilant categories /s/ and /ʃ/. For NH subjects the best regression model was based on visual acuity, haptic acuity and auditory acuity for a vowel continuum. Surprisingly, auditory acuity on a sibilant continuum was not a significant predictor of sibilant production performance. CI had comparable (but not better) performance to NH on visual and haptic acuity. As expected their auditory acuity on the sibilant continuum was worse than NH, but rather surprisingly their auditory acuity on the vowel continuum was even worse. For CI none of the acuity measures proved significant predictors of production performance (even though on a group basis production performance on sibilants was, as expected, worse than that of NH). Future work will focus on further development of the visual and auditory acuity tests to give a direct measure of multimodal integration ability, since this has been argued to be a key skill in deriving maximum benefit from prosthetic hearing.
Publications
-
(2014). Quantitative analysis of multimodal perception in cochlear implant wearers. In Proceedings of the 15th ICPLA Conference (International Clinical Linguistics and Phonetics Association), Stockholm (p. 100)
Hoole, P., Gluth, C.
-
(2014). Speech of Cochlear Implant Patients: An Acoustic Analysis of Vowel Production. In Proceedings of the 15th ICPLA Conference (International Clinical Linguistics and Phonetics Association), Stockholm (p. 97)
Neumeyer, V., Schiel, F. & Hoole, P.
-
(2015) Speech of cochlear implant patients: An acoustic analysis of sibilant production. In The Scottish Consortium for ICPhS 2015 (Ed.), Proc. 18th Int. Cong. Phonetic Sciences. Glasgow
Neumeyer, V., Schiel, F. & Hoole, P.
-
(2015). Akustische Analysen der Sprachproduktion von CI-Trägern. Dissertation IPS München
Neumeyer, V.
-
(2015). How can speech production skills be predicted from visual, auditory, and haptic perception skills? In The Scottish Consortium for ICPhS 2015 (Ed.), Proc. 18th Int. Cong. Phonetic Sciences. Glasgow
Gluth, C. & Hoole, P.
