- 期刊
Development and Verification of the Taiwanese Mandarin Quick Speech-in-Noise Test (TM-QuickSIN)
背景噪音中臺灣華語語句聽辨快速測驗之發展與驗證
摘要
This study aimed to develop a clinically relevant quick speech-in-noise test with a test list containing the majority of the phonetics of the Taiwanese Mandarin language. Eighty-five 20-44-year-old adults were enrolled (normal hearing - NH, 73; hearing loss - HL, 12). The test was developed in the following stages: compilation of Taiwanese Mandarin sentences and four-talker babble noise, screening sentences in babble (in which short sentences containing five keywords each were mixed with four-talker babble noise), constructing test lists, assessing the reliability and homogeneity of lists, and exploring signal-to-noise ratio (SNR) loss in adults with hearing loss. The results showed that the mean SNR at the 50% correct recognition point (SNR_(50)) of adults with normal hearing was -2.22 dB (SD = 0.78), while adults with hearing loss required, on average, an additional 5.31 dB (SD = 3.66) to achieve this same recognition score (i.e. their "SNR loss"). While the age and word recognition scores (WRS) in quiet conditions were similar between the two hearing groups, independent sample t-test results revealed statistically significant differences both in SNR loss and instantaneous slope at the 50% correct point (p < .05). Moreover, the SNR loss in the HL group was not related to hearing levels or WRS in quiet situations. In conclusion, the TM-QuickSIN test we developed consists of 10 test lists with good inter-list equivalence and test-retest reliability for assessing SNR loss. The 108 normalized sentences in babble developed in this study are reliable materials for further auditory and speech perception research in Taiwan. Compared with normal-hearing listeners, adults with hearing loss exhibit significant differences in terms of SNR loss and psychometric functions, thereby emphasizing the value of speech-in-noise tests in comprehensive auditory assessments.
並列摘要
本研究旨在發展與驗證噪音下臺灣華語語句聽辨識快速測驗，且句子涵蓋日常臺灣華語主要語音音素。總計85位年齡20至44歲成人參與本研究，聽力正常者73人，聽力損失者12人。研究程序分成以下階段：彙編短句材料和四人版雜沓式噪音、篩選測驗材料（含五個關鍵詞的短句合併雜沓式噪音）和測驗表單編制、驗證信效度與測驗表單間的同質性考驗，並比較聽力正常與聽力損失成人之訊噪比損失（signal-to-noise ratio loss, SNR loss）。研究結果顯示正常聽力成人噪音下華語短句聆聽辨識達50%正確率之平均訊噪比值（SNR_(50)）為-2.22 dB（SD = 0.78 dB），而聽損成人平均需要額外5.31 dB（SD = 3.66）來達到相同的辨識分數（即他們的訊噪比損失）。本研究聽常與聽損成人雖有相似的安靜情境華語單字詞辨識正確率，但在訊噪比損失與50%辨識正確率的瞬間斜率均有統計顯著差異（p<.05）；此外，結果也指向聽損成人的訊噪比損失與其聽損程度和安靜下單字詞辨識正確率皆沒有關聯。本研究成果如下：（1）發展10張具高度表間同質和再測信度的背景噪音中臺灣華語語句快速測驗表單，供臨床評估訊噪比損失；（2）建構108組臺灣華語短句－雜沓式噪音組合，可作為國內未來在言語聽知覺相關研究之材料；（3）相較於聽常者，聽力損失成人有顯著不同的訊噪比損失和言語聽辨識心理函數關係，強調在完整的聽覺評估中，納入噪音下言語聽辨識測驗有其價值。