Neck sensor-supported hyoid bone movement tracking during swallowing.

Area of Science:

• Biomedical Engineering
• Physiology
• Machine Learning

Background:

• Hyoid bone movement is crucial for normal swallowing function.
• Current gold-standard methods like videofluoroscopic swallowing studies have limitations including radiation exposure and high costs.
• There is a need for non-invasive, accessible methods to assess hyoid bone movement.

Purpose of the Study:

• To develop and validate a non-invasive system for tracking hyoid bone movement using accelerometry.
• To utilize deep neural networks to model the relationship between hyoid bone motion and sensor signals.
• To provide an alternative to traditional swallowing assessment methods.

Main Methods:

• A non-invasive accelerometry sensor was attached to the neck surface to capture hyoid bone movement.
• Deep neural networks were employed to analyze sensor signals and predict hyoid bone movement.
• The system was trained and validated on a dataset of 400 swallows from 114 patients.

Main Results:

• The developed system demonstrated promising performance in predicting hyoid bone movement compared to expert judgments.
• The study confirmed that universal patterns of hyoid bone movement can be captured by nonlinear algorithms.
• The computer-aided prediction showed high accuracy in assessing swallowing function.

Conclusions:

• A sensor-supported strategy using accelerometry and deep learning offers a viable, radiation-free alternative for online hyoid bone movement tracking.
• This approach provides a flexible method for identifying dysphagia and other swallowing disorders.
• The technology has the potential to improve diagnostic accessibility and patient outcomes in swallowing assessments.