The development and evaluation of a new test of pitch perfection for cochlear implant users

Abstract

Pitch perception, which is important for perceiving music, understanding tonal language and intonation cues and analysing the auditory scene, is limited by suboptimal temporal and spectral cues in cochlear implant (CI) users. The measurement of pitch perception has been quantified by several tests, however they demonstrate flaws in methodology and design and so the motivation behind this thesis was to improve on these tests.

Seven existing tests of pitch perception were evaluated using both normal hearing listeners (NHL) and CI users. The best performing test was the Melodic Contour Identification (MCI) test: it used a non-adaptive method, showed minimal floor and ceiling effects and had good reliability on retest. It had limitations too: the 9 melodic contours differed in their complexity, making some contours easier than others and the contours spanned up to 5 notes, meaning that it was not possible to assess single intervals.

A new test of pitch perception, the Pitch Contour Test (PCT) was designed to improve on existing tests. It used a non-adaptive method, which allowed the psychometric function relating pitch interval size to performance, to be estimated and visualised. The stimuli consisted of 4 contours, equal in difficulty and representative of single intervals, which allowed pitch discrimination and pitch ranking ability to be assessed simultaneously. It provided sufficient numbers of trials to ensure statistical confidence in the result and specified levels required for success.

The PCT was evaluated by comparing it to three existing tests of pitch perception: the University of Washington Clinical Assessment of Music Perception (UW CAMP), the Melodic Contour Identification (MCI) test and the South of England Cochlear Implant Centre Music Test Battery (SOECIC MTB), using NHL and CI users.

The PCT was superior to these tests with regard to numbers of trials, its ability to assess pitch discrimination and ranking simultaneously and its ability to estimate the psychometric function and be suitable for participants who demonstrate a non-monotonic function. The PCT performed similarly to the UW CAMP in terms of reliability, and similar to the MCI and the SOECIC MTB in terms of being sensitive to musicianship in the NHL group, however it did demonstrate floor and ceiling effects.

The development of the PCT impacts clinicians’ ability to assess pitch perception in a more holistic way, and allows psychometric functions to be more fully explored. It has been used clinically to assess and determine the benefits to switching certain electrodes off in order to improve the listening experience for CI users. Use of the PCT throughout this thesis has demonstrated that both CI users and NHL can demonstrate non-monotonic psychometric functions. Using traditional adaptive methodologies when a psychometric function is non-monotonic can result in erroneous final results and without an estimation of the psychometric function, these results would appear to be accurate. Using a non-adaptive method when testing the pitch perception of CI users is therefore considered to be essential.

More information

Identifiers

Catalogue record

Export record