Validation of the Acoustic Voice Quality Index in the Korean Language

Abstract

The purpose of this study was to explore the concurrent validity of the Acoustic Voice Quality Index (AVQI) in a Korean population. We investigated the feasibility of its cutoff values and diagnostic accuracy in discriminating between normal and dysphonic voices. A total of 1,524 native Korean subjects with normal voices (n = 113) and with voice disorders (n = 1,411) were asked to speak and sustain the vowel "a" and to read the Korean text "Walk" aloud. A 2-second mid-speech portion of the sustained vowel sound and two sentences (with 26 syllables) were edited and concatenated, and the AVQI was calculated. Additionally, two auditory-perceptual assessments, namely the Grade, Roughness, Breathiness, Asthenia, Strain scale, and the Consensus Auditory-Perceptual Evaluation of Voice were used by five speech language pathologists to rate the severity of dysphonia. Both auditory-perceptual assessments showed high confidence levels among the five speech language pathologists. The AVQI correlated with grade (G) and overall severity (OS). There were statistically significant differences in AVQI, G, and OS between normal and pathological speech groups. In the receiver operating characteristic curve analysis, AVQI, G, and OS cutoff values were <3.33, <0.00, and <22.00, respectively. The receiver operating characteristic curve analysis indicated that AVQI had excellent diagnostic accuracy for discriminating between normal and dysphonic voices (area under the curve: 0.970-0.977). We demonstrated the concurrent validity of AVQI as a promising tool for assessing overall voice quality and quantifying dysphonia in a Korean-speaking population.