Swallowing Activity Classification Model for Older Adults
Author Information
Author(s): Li Dan, Zhang Yaqing, Wu Junhui, Luo Wei, Liu Tao, Lyu Benshuai, Shang Shaomei
Primary Institution: Peking University
Hypothesis
This study aims to develop a swallowing activity classifier based on acoustic recording and analysis using machine learning algorithms.
Conclusion
The study found that distinct acoustic patterns can be effectively classified and identified using machine learning methods, which may help in continuous monitoring of swallowing activities in older adults.
Supporting Evidence
- Dysphagia poses a significant health threat for old people worldwide.
- Acoustic diagnosis has shown promising effectiveness in dysphagia screening and aspiration detecting.
- The MLP, CNN, and CRNN models achieved accuracy values of 74%, 68%, and 54%, respectively.
Takeaway
Researchers created a tool that listens to how older people swallow to help doctors know if they're having trouble swallowing.
Methodology
The study recorded and annotated swallowing audio data from 61 old individuals and analyzed signal features using machine learning models.
Limitations
Current assessment tools for swallowing function have pronounced limitations.
Participant Demographics
Old individuals.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website