Related Experiment Video
Updated: May 26, 2025
Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
Published on: August 9, 2024
Screening Voice Disorders: Acoustic Voice Quality Index, Cepstral Peak Prominence, and Machine Learning
Ahmed M Yousef, Adrián Castillo-Allendes1,2,3, Mark L Berardi3,4
1Center for Laryngeal Surgery and Voice Rehabilitation, Massachusetts General Hospital, Boston, Massachusetts, USA.
Introduction:
The Acoustic Voice Quality Index (AVQI) and Smoothed Cepstral Peak Prominence (CPPs) have been reported to effectively support the assessment of voice quality in persons seeking voice care across many languages. This study aimed to evaluate the diagnostic accuracy of these two measures in detecting voice disorders in American English speakers, comparing their performance to machine learning (ML) models.
Methods:
This retrospective study included a cohort of 187 participants: 138 patients with clinically diagnosed voice disorders and 49 vocally healthy individuals. Each participant completed two voicing tasks: sustaining [a:] vowel and producing a running speech sample, which were then concatenated. These samples were analyzed using VOXplot software for AVQI-3 (version 03.01) and CPPs. Additionally, four ML models (random forest, k-nearest neighbors, support vector machine, and decision tree) were trained for comparison. The diagnostic accuracy of the two measures and models was assessed using various evaluation metrics, including receiver operating characteristic curve and Youden Index.
Results:
A cutoff score of 1.54 for the AVQI-3 (with 55% sensitivity and 80% specificity) and 14.35 dB for CPPs (with 65% sensitivity and 78% specificity) were identified for detecting voice disorders. Compared to an average ML sensitivity of 89% and specificity of 55%, CPPs offered a better balance between sensitivity and specificity, outperforming AVQI-3 and nearly matching the average ML performance.
Conclusions:
ML shows great potential for supporting voice disorder diagnostics, especially as models become more generalizable and easier to interpret. However, current tools like AVQI-3 and CPPs remain more practical and accessible for clinical use in evaluating voice quality than commonly implemented models. CPPs, in particular, offers distinct advantages for identifying voice disorders, making it a recommended and feasible choice for clinics with limited resources.
More Related Videos
09:09Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody
Published on: September 27, 2024
06:04Systematic Hearing Performance Evaluation Process for Adolescents with Cochlear Implantation at Early Ages
Published on: March 24, 2023
Related Concept Videos
01:28Physical Assessment of the Respiratory Tract IV: Auscultation
Breath Sounds
Breath sounds are categorized into vesicular, bronchovesicular, and bronchial.
01:21Perceiving Loudness, Pitch, and Location
Place theory, or place coding, suggests that different pitches are heard because various sound waves activate specific locations along the cochlea's basilar membrane. The brain determines the pitch of a sound by...
01:31Respiratory System Abnormal Finding II: Palpation and Auscultation
Palpation Findings
During a respiratory assessment, palpation can reveal several vital abnormalities:
01:30Assessment of Airway, Skin Color, and Use of Accessory Muscles
Introduction
The initial evaluation of a patient's respiratory system...
01:30Classification of Signals
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...