DIAGNOSIS OF SWALLOWING DIFFICULTIES USING TIME-FREQUENCY ANALYSIS OF SWALLOWING SOUND SIGNALS BASED ON WAVELET TRANSFORMATION

Since aspiration in the elderly and patients with low resistance can easily develop into aspiration pneumonia, convenient, inexpensive and noninvasive methods that can be performed rapidly at the bedside are needed to identify patients at a higher risk of aspiration. To assess aspiration risk, a stethoscope is placed on the neck of the patient to hear swallowing sounds. However, aspiration may go undetected when testing is conducted by a person unused to identifying abnormal sounds. Furthermore, recent trends have been toward performing thorough screenings without conducting tests involving a risk of radiation exposure, such as X-ray cinefluorography and radiation-emitting tests. We therefore attempted to clarify the characteristic sounds associated with aspiration by capturing swallowing sounds using an acceleration sensor and performing a time-frequency analysis of swallowing sound signals based on wavelet transformation. Subjects in the present study comprised healthy individuals with a mean age of range, 20–60 years. Each subject was instructed to swallow a predetermined amount of water or jelly. In each subject, swallowing sounds were recorded and analyzed. The results identified 3 characteristic sounds associated with eating and swallowing. In addition, the temporal relationships of these sounds were analyzed, clarifying age-related characteristic changes.