Related Experiment Video
Updated: Apr 30, 2026
10:16
Synthetic, Multi-Layer, Self-Oscillating Vocal Fold Model Fabrication
Published on: December 2, 2011
14.0K
Subject-Specific Modeling by Domain Adaptation for the Estimation of Subglottal Pressure from Neck-Surface
Emiro J Ibarra1, Julián D Arias-Londoño2, Juan I Godino-Llorente2
1Department of Electronic Engineering and Advanced Center for Electrical and Electronic Engineering, Universidad Técnica Federico Santa Maria, Valparaiso, 2390123, Chile.
Biomedical Signal Processing and Control
|March 26, 2025
Summary
This study estimates subglottal air pressure using neck-surface acceleration signals and machine learning. The novel approach significantly improves accuracy for voice disorder assessment.
Area of Science:
- Biomedical Engineering
- Speech Science
- Computational Linguistics
Background:
- Subglottal air pressure is crucial for understanding voice disorders.
- Current measurement methods are invasive and impractical for ambulatory settings.
- Estimating subglottal pressure from non-invasive signals is a significant challenge.
Purpose of the Study:
- To develop and validate a machine learning model for estimating subglottal air pressure from neck-surface acceleration signals.
- To improve the accuracy and practicality of subglottal pressure assessment in both laboratory and ambulatory settings.
- To create subject-specific models for enhanced estimation accuracy.
Main Methods:
- Utilized a physiologically relevant voice production model and neural network architecture.
- Employed a domain adaptation strategy to refine the model from synthetic to in vivo data.
- Collected synchronous oral airflow, intraoral pressure, microphone, and neck-surface accelerometer data from 135 participants with normal and disordered voices.
Main Results:
- The proposed subject-specific models achieved over a 21% improvement in subglottal pressure estimation compared to previous methods (measured by RMSE).
- Demonstrated the effectiveness of a non-linear, subject-specific regression approach.
- Successfully applied the method to various voice conditions, including unilateral vocal fold paralysis and vocal hyperfunction.
Conclusions:
- The developed method offers a non-invasive, accurate, and practical approach for estimating subglottal air pressure.
- Subject-specific refinements significantly enhance the estimation of subglottal pressure from neck-surface vibration signals.
- This technique holds promise for improved diagnosis and monitoring of voice disorders.
