Related Experiment Video
Updated: May 17, 2026
09:52
An Implantable System For Chronic In Vivo Electromyography
Published on: April 21, 2020
Artificial neural network classification of pharyngeal high-resolution manometry with impedance data
Matthew R Hoffman1, Jason D Mielens, Taher I Omari
1Department of Surgery, Division of Otolaryngology-Head and Neck Surgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, USA.
The Laryngoscope
|October 17, 2012
Summary
This study shows artificial neural networks can accurately classify swallowing function using high-resolution manometry with impedance. This method helps identify patients at risk for aspiration or penetration.
Area of Science:
- Swallowing physiology and biomechanics
- Medical device technology
- Artificial intelligence in healthcare
Background:
- Dysphagia, or difficulty swallowing, affects many patients and can lead to aspiration.
- Pharyngeal high-resolution manometry (HRM) with impedance provides detailed data on swallow function.
- Objective classification of swallow events is crucial for clinical management.
Purpose of the Study:
- To evaluate the efficacy of artificial neural networks (ANNs) in classifying swallows.
- To differentiate between safe swallows, penetration, and aspiration using HRM-impedance data.
- To assess the diagnostic accuracy of ANN models for swallowing disorders.
Main Methods:
- A case series design was employed to evaluate a novel data analysis method.
- Multilayer perceptron ANNs were trained and tested on data from 25 subjects with swallowing disorders.
- Swallows were classified as safe, penetration, or aspiration based on extracted pharyngeal HRM-impedance parameters.
Main Results:
- The ANN achieved an overall classification accuracy of 89.4 ± 2.4%.
- Models using only manometry parameters achieved 85.0 ± 6.0% accuracy, while impedance-only models yielded 76.0 ± 4.9%.
- Receiver operating characteristic analysis showed strong performance for classifying safe swallows, aspiration, and penetration.
Conclusions:
- ANNs demonstrate high accuracy in classifying swallows in patients with dysphagia.
- HRM-impedance combined with ANNs offers a promising clinical tool for screening patients at risk of aspiration or penetration.
- This technology can aid in early detection and management of swallowing impairments.
