Development and Validation of a Wearable Sensor-Based Algorithm for the Detection of Swallowing Events in Parkinson's Disease

Oropharyngeal Dysphagia (OD) is a swallowing disorder characterized by difficulties in bolus transport and airway protection. It represents one of the axial symptoms of Parkinson's Disease (PD) and affects up to 80% of patients. OD is often underestimated, but significantly increases the risk of aspiration pneumonia, which is the main cause of hospitalization in PD. Despite its high prevalence, OD often remains undiagnosed until the later stages of the disease and its gold-standard diagnostic methods are invasive, expensive, and unsuitable for continuous monitoring. For these reasons, there is a growing need for non-invasive and accessible tools for the early detection of dysphagia. This thesis proposes an innovative approach for the identification and characterization of swallowing events in PD. In particular, the aim is to develop and validate an algorithm for the automatic detection of swallowing events on the basis of accelerometric signals. The study involved twenty-six PD patients and seven healthy controls who performed four swallowing tasks. A triaxial accelerometer was placed on the thyroid cartilage as part of a multimodal non-invasive setup to record swallowing activity. Accelerometric signals were processed in MATLAB, using manual frame-by-frame segmentation as the reference to design and optimize the automatic algorithm and to evaluate its performance. The algorithm prediction showed good agreement with manual segmentation, with an overall F1-score of 0.76 and global mean errors in predicting the onset and offset of 0.27 ± 1.32s and 0.16 ± 1.03s, respectively. The results demonstrated the promising accuracy of the proposed algorithm in localizing swallowing events within the signal. These findings represent the possibility of using accelerometry to assess swallowing activity in an automatic, non-invasive and reliable way, potentially enabling the early detection of OD and the prevention of its complications.