Automatic detection of swallowing events based on surface electromyography

Area of Science:

• Biomedical Engineering
• Neuroscience
• Clinical Diagnostics

Background:

• Clinical evaluation of dysphagia is often invasive and subjective, relying heavily on Videofluoroscopic Swallowing Study (VFSS).
• Accurate detection of bolus transit through the mandible line (ManL) and upper esophageal sphincter (UES) is crucial for diagnosing oropharyngeal dysphagia.
• There is a need for non-invasive, objective methods to complement existing diagnostic tools.

Purpose of the Study:

• To evaluate the efficacy of surface electromyography (sEMG) signals for detecting bolus passage in patients with oropharyngeal dysphagia.
• To introduce and validate a novel method utilizing GRU networks and sEMG spectrograms for dysphagia assessment.
• To compare the performance of the proposed sEMG-based method against the gold standard VFSS.

Main Methods:

• sEMG signals were recorded from patients with oropharyngeal dysphagia.
• Spectrograms were generated from processed sEMG signals.
• Gated Recurrent Unit (GRU) networks were employed to analyze spectrograms for detecting bolus passage at ManL and UES.
• VFSS was used as the gold standard for comparison.

Main Results:

• The sEMG-based method achieved high accuracy, with F1 scores up to 96% for binary and 80% for tri-class classifications.
• Performance was superior to existing state-of-the-art methods.
• Detection accuracy improved with larger bolus volumes and thinner consistencies.

Conclusions:

• sEMG is a promising, non-invasive biosignal for dysphagia diagnosis and monitoring.
• The proposed GRU network-based analysis of sEMG spectrograms offers an objective and accurate method for detecting bolus passage.
• sEMG can serve as an effective complement to VFSS, enhancing dysphagia assessment.