Experiments on prosody perception with cochlear implants

Background and objective. Prosody has a myriad of linguistic functions and involves specific aspects of speech, such as stress, intonation and pauses. The underlying acoustic quantities (amplitude envelope, pitch frequency, and temporal structure) can be processed and transmitted by cochlear implants (CI) only to a limited extent. At present, no adequate tests are available in the German-speaking world for evaluation of the perception of prosodic elements. Different experiments have been conducted to address several prosodic cues, and the results are to be used as a basis for appropriate tests. Methods. Various prosodic materials were used for the experiments. Discrimination was measured for minimal pairs differing in frequency and/or duration, accents in words and phrases, questions versus statements and phrasing. Measurements were performed in ten normal-hearing subjects and five with cochlear implants. Results and conclusions. In all test modules, the subjects with normal hearing proved to have high discrimination rates of 96-100%. The test of word stresses was problematic because the results were influenced by different confounders. The other measurements did prove to be basically suitable for use in the subjects with implants. Early results revealed that the subjects with CI had few problems with prosodic cues based on the temporal structure, the outcome being similar to that of the subjects with normal hearing in these tests. In contrast, the performance of subjects with CI in perceiving prosodic cues based on amplitude variations and, especially, on alterations in pitch frequency was worse, even though some of them achieved very good results in these tests too. These preliminary tests can form the basis for development of a German-language prosody test battery with a limited number of subtests addressing different prosodic cues .