The Development of Mandarin Tone Auditory Test Battery and Its Application to Cochlear-Implant Users

• Main Authors: Chu-Hsiu Teng, 鄧菊秀
• Other Authors: Hsiao-Chuan Chen
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/36072438348370093207

博士 === 國立高雄師範大學 === 特殊教育學系 === 99 === The purposes of this study were to develop a set of Mandarin Tone Auditory Test Battery (MTATB) as an assessment tool for tonal auditory perception and to explore the performance of cochlear-implant (CI) users as well. Two subsets of batteries in the MTATB were Lexical Pitch Test Battery (LPTB, No. A) and Lexical Tone Test Battery (LTTB, No. B). The former was comprised of three tests, including Frequency Difference Discrimination Test (FDD, A1), F0 Height Recognition Test (F0HR, A2), and F0 Slope Recognition Test (F0SR, A3). The latter was comprised of two tests, Mandarin Lexical Tone Discrimination Test (MLTD, B1) and Mandarin Lexical Tone Recognition Test (MLTR, B2). Besides, this study applied the MTATB to CI users for investigating several aspects of results, such as the tonal auditory perception of CI users, the difference of the performances in all tests between CI users and normal hearing (NH) listeners, and the predictability of test scores in both groups. The subjects had two groups, including 30 NH listeners for the pretest, 39 CI users and 39 age/sex-matched NH listeners for the formal test, and 5 in each group for the retest, respectively. All subjects were tested individually through standard testing procedures held in a typical sound booth. The research tools for this study consisted of tympanometry, audiometer, sound level meter, Praat software, Chinese speech recognition threshold test, vowel and consonant recognition tests, and the self-developed MTATB. The main results were as follows: First, the development of the MTATB showed that：(1) its results had high internal consistency (Cronbach α=.867 to .982), split-half reliability (Spearman-Brown Coefficient=.898 to .986) and test-retest reliability (r=.808 to .999, p＜.01；α=.797 to .998); (2) its results had good content validity (based on theories and multiple cross-validated checking processes), criterion validity (r=.280 to .849, p＜.05), and construct validity (including pattern recognition, tone discrimination, and frequency discrimination factors); (3) the battery itself had multi-dimensional characteristics (including lexical pitch/lexical tone, discrimination/recognition, and non-speech/speech). Second, the performance of the MTATB showed that: (1) the mean scores of the five tests for CI users were ranked as a medium level approximately (i.e. A1 Δf=43.55 Hz, Δf/f=0.162, A2=52.36, A3=41.47, B1=85.66, and B2=65.85), and the correlations between any two of the five scores were significant (r=.342 to .663, p＜.05)；(2) the differences were observed in auditory perception abilities between CI and NH groups, and obviously CI users had achieved lower scores than NH listeners in all tests (about 1.28 to 1.76 SD); (3) CI users’ performance of lexical tone recognition was affected by the interaction between tones and consonants (manners: F=3.66, p＜.001；places: F=1.77, p＜.05), but not by that between tones and vowels (F=1.27, p＞.05)；(4) the response error rates of CI users were reached to a certain degree (between 2.0 % and 40.0 %) in tests A2, A3, and B2, and these errors also showed pattern-specific tendencies. Third, the predictability of lexical pitch and lexical tone performance showed that: (1) the tests, as predictive variables selected from the MTATB, LPTB and LTTB, had a similar order on their importance between both groups; (2) the three tests in LPTB could highly predict lexical pitch performance, and A2 was the most important predictor (R squared change: CI=79.1 %, NH=83.5 %), A3 the less (CI=16.7 %, NH=15.1 %), and A1 the least (CI=4.2 %, NH=1.4 %); (3) the two tests in LTTB could highly predict lexical tone performance and B2 (CI=86.9 %, NH=54.7 %) was more important than B1(CI=13.1 %, NH=45.3 %); (4) the three tests in LPTB predicted the performance of LTTB, B1, and B2 with low to medial level; however, both A2 and A3 for NH listeners (46.2 %), but only A2 for CI users (43.9 %) could predict the performance of Mandarin lexical tone recognition; (5) five tests and two sub-batteries could highly predict the performance of the MTATB, and A2 was the most important predictor (CI=76.2 %, NH=80.5 %), and LPTB (CI=85.0 %, NH=96.8 %) was more important than LTTB (CI=15.0 %, NH=3.2 %); (6) the items, as predictive variables selected from these five tests, LPTB and LTTB, had a different order on their importance between both groups. Based on the above-mentioned findings, the researcher proposed several suggestions for clinical and practical application. Limitations of this study and future researches were also addressed.