Signal Typing and Acoustic Analysis of Voice in School-Aged Indian Children

The current study aimed to classify voice signals using narrowband spectrograms and investigate acoustic measures in samples classified as type 1 and rated normal based on auditory-perceptual evaluation.

Voice samples of sustained phonation of /a/ of 234 males and 223 females aged between 4 and 16; 11 years were recorded and classified into 5 age groups. Narrowband spectrograms were generated and signal typing was done by two raters. Acoustic parameters for type-I signals were derived using PRAAT software (version 6.3.01). Inter-rater reliability was measured for both signal typing and auditory-perceptual evaluation.

Based on agreement by two raters, around 79% of samples were classified as type 1, 8.5% as type II, and 7.8% as type-III signals. Inter-rater reliability for signal typing and GRBAS (Grade,Roughness, Breathiness, Asthenia, Strained) rating were found to be good (Intraclass Correlation Coefficient-0.90 and 0.81). Overall comparison using analysis of variance revealed statistically significant difference F0-related measures and NHR values (P < 0.05). Within males and females, F0 measures were statistically significant (P < 0.05). Noise-related measures were significant in males (P < 0.05). Independent t test revealed sex differences for F0 measures, jitter %, and NHR values in children above 13 years of age (P < 0.05).

Normative data obtained in this study can be used as a reference to compare with children with pediatric dysphonia.