Related Experiment Video
Updated: Jul 10, 2026
14:13
Coordinate Mapping of Hyolaryngeal Mechanics in Swallowing
Published on: May 6, 2014
Analysis of swallowing sounds using hidden Markov models
Mohammad Aboofazeli1, Zahra Moussavi
1Department of Electrical and Computer Engineering, University of Manitoba, Winnipeg MB R3T 5V6, Canada.
Medical & Biological Engineering & Computing
|November 15, 2007
Summary
This study introduces a hidden Markov model (HMM) method for analyzing swallowing sounds, improving segmentation and classification accuracy for healthy and dysphagic individuals. The HMM approach effectively distinguishes swallowing phases and identifies patients with swallowing difficulties.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Acoustics
Background:
- Acoustical analysis of swallowing shows diagnostic potential.
- Accurate segmentation and classification of swallowing sounds are crucial for diagnosing dysphagia.
Purpose of the Study:
- To develop and evaluate a hidden Markov model (HMM) based method for swallowing sound segmentation and classification.
- To compare the performance of different HMM configurations and features for analyzing swallowing sounds.
Main Methods:
- Swallowing sound signals from healthy and dysphagic subjects were analyzed.
- Signals were segmented into 25 ms frames and represented by seven features.
- Hidden Markov Models (HMMs) were trained for segmentation and classification tasks.
- Multi-scale product of wavelet coefficients and root mean square (RMS) were among the features investigated.
Main Results:
- HMMs accurately segmented swallowing sounds into initial quiet period, initial discrete sounds (IDS), and bolus transit sounds (BTS).
- Segmentation accuracy was highest using HMMs with multi-scale product of wavelet coefficients.
- Classification accuracy improved with an increased number of HMM states (up to 8).
- RMS and waveform fractal dimension (WFD) were the best performing features for classification, achieving up to 85.5% accuracy for IDS segments.
Conclusions:
- HMM-based acoustical analysis is a promising method for swallowing sound segmentation and classification.
- The proposed method can aid in the diagnosis of dysphagia by differentiating between healthy and impaired swallowing mechanisms.
- Further optimization of HMM configurations and feature selection can enhance diagnostic capabilities.
Related Concept Videos
01:25Deglutition
Swallowing, otherwise known as deglutition, facilitates the transport of food from the mouth to the stomach. It is a multifaceted process that involves both the tongue and the muscles of the throat and esophagus. Saliva and mucus aid in this process, which takes approximately 4 to 8 seconds for semi-solid or solid food and around 1 second for liquids or very soft food.
Swallowing can be divided into three stages: the voluntary phase, the pharyngeal phase, and the esophageal phase. Although the...
Swallowing can be divided into three stages: the voluntary phase, the pharyngeal phase, and the esophageal phase. Although the...
01:02¹H NMR: Interpreting Distorted and Overlapping Signals
Spin systems where the difference in chemical shifts of the coupled nuclei is greater than ten times J are called first-order spin systems. These nuclei are weakly coupled, and their chemical shifts and coupling constant can generally be estimated from the well-separated signals in the spectrum.
As Δν decreases and the signals move closer, the doublets appear increasingly distorted. The intensities of the inner lines increase at the cost of those of the outer lines as the signals are slanted or...
As Δν decreases and the signals move closer, the doublets appear increasingly distorted. The intensities of the inner lines increase at the cost of those of the outer lines as the signals are slanted or...