Method for Estimating Amount of Saliva Secreted Using a Throat Microphone.

Area of Science:

• Biomedical Engineering
• Oral Health
• Wearable Technology

Background:

• Insufficient saliva secretion can lead to oral health issues like glossitis and stomatitis.
• Daily fluctuations in saliva amount necessitate continuous monitoring for timely intervention.
• Current methods lack the capability for constant saliva measurement.

Purpose of the Study:

• To develop a non-invasive method for continuously estimating saliva secretion.
• To utilize sound data from a wearable throat microphone for saliva monitoring.
• To address the limitations of existing methods for real-time saliva assessment.

Main Methods:

• A deep learning model was employed to classify swallowing sounds captured by a throat microphone.
• The classification of swallowing events was used as a basis for estimating saliva secretion.
• The system was evaluated for its accuracy in swallowing detection and saliva amount estimation.

Main Results:

• The deep learning model achieved a high accuracy of 96.96% in classifying swallowing sounds.
• Saliva secretion estimation demonstrated a correlation coefficient (R) of 0.600.
• The Mean Absolute Error (MAE) for saliva amount estimation was 0.0487.

Conclusions:

• The proposed method offers a promising approach for continuous, non-invasive saliva secretion monitoring.
• Detecting swallowing sounds via wearable technology can effectively estimate saliva levels.
• This technology has the potential to improve the management of conditions related to salivary dysfunction.