Objectives: New or increased cough is an important sign of pulmonary exacerbation. There is unmet need to objectively quantify and analyse cough trends in early life. Advances in flexible electronics and materials allow development of soft, skin-like, accelerometer-based wearable devices that seamlessly interface with the human body in unique locations, enabling the simultaneous capture of low-frequency body, chest, and throat motion along with high-frequency vocal, throat, and lung sound signals associated with various body processes including coughs and vital signs. Methods: A small, flexible, fully wireless, accelerometer-based mechano-acoustic sensor (MAS) was applied using gentle adhesives to the suprasternal notch of 10 paediatric CF subjects during regular clinic visits. A 15-minute protocol consisting of eliciting cough, throat clearing, speech, and laughs in various head/body orientations, ambient environments, and physical activity intensities. The MAS recorded at a high bandwidth 1.6kHz rate and automatically uploaded to a cloud server upon replacement on a wireless charging platform. Results: The captured sensor data was analysed using an existing machine learning algorithm developed for COVID-19 symptom tracking. Preliminary results show the device differentiates cough from other vocal, respiratory noises and motion artifacts, both at rest and during activity. A majority of children found the device to be acceptable, though device size and adhesive removal caused minor discomfort for some. Conclusion: The extraction of well-classified cough events is feasible, facilitating further analysis of cough episode duration, force, and clinically relevant features associated with early decompensation, response to treatment, and daily symptom tracking. Objective measurement of cough may be useful as a clinical study outcome measure and for clinical monitoring. Future work will include evaluation for longer time periods during stability and pulmonary exacerbations.
Read full abstract