Cough monitoring for pulmonary tuberculosis using combined microphone/accelerometer measurements

Abstract

A laboratory-free test for assessing recovery from pulmonary tuberculosis (TB) would be very helpful in regions of the world where laboratory facilities are lacking. Our hypothesis is that analysis of cough sound recordings may provide such a test, as recovering patients should cough less frequently. We have carried out several studies on cough data from a cohort of TB patients in Lima, Peru [Larson et al., PLOS One]. Our previous work provides a foundation to support larger-scale studies of coughing rates over time for TB patients undergoing treatment, but it only used recordings from lapel microphones. For the current study, we use an additional channel recorded by a throat-mounted accelerometer. The accelerometer only responds to patient-generated noise events and thus provides robustness to background noise in the environment. We describe algorithm development for cough data analysis using combined microphone/accelerometer measurements, and compare several event detection and classification strategies. We show that adding the accelerometer improves performance on detection and classification.