Adaptation and Validation of the Acoustic Voice Quality Index for Voice Disorder Assessment for the Mandarin...

This study aimed to determine the standardized syllable numbers (SSNs) for analyzing the third version of Acoustic Voice Quality Index in Mandarin (AVQIv3-M), to adapt and validate AVQIv3-M using the determined SSNs, and to identify the most suitable continuous speech (CS) material for clinical practice.

A total of 141 native Mandarin speakers (101 with dysphonia and 40 vocally healthy) were instructed to read a passage, count numbers, and sustain the vowel /a/. First, the SSNs corresponding to approximately three seconds of CS were determined for the reading passage and counting numbers. Second, auditory-perceptual judgments were performed by five experienced raters using the Grade (G) parameter of the GRBAS scale. Third, rater reliability, concurrent validity, and diagnostic accuracy were evaluated.

The optimal SSNs were 22 syllables for the passage and 19 syllables for the numbers. AVQIv3-M scores showed strong correlations with G scores (rₛ = 0.808 for the passage; rₛ = 0.816 for the numbers). Receiver operating characteristic analysis demonstrated good diagnostic accuracy, with the area under curve reaching 0.814 for the passage and 0.842 for the numbers. For the passage and numbers, the optimal cutoff values were 2.595 and 1.79, respectively. The passage showed higher specificity (100%) but lower sensitivity (54.7%), whereas the numbers demonstrated higher sensitivity (70.2%) with slightly reduced specificity (88.9%).

AVQIv3-M is a valid and reliable acoustic measure for assessing voice quality in Mandarin. Speech material selection influences diagnostic performance, with the reading passage offering higher specificity and the counting numbers greater sensitivity for identifying dysphonia.