A machine learning-powered smartphone application could help patients with a chronic cough document their symptoms over an extended period of time, according to new findings published in Digital Biomarkers. Such a solution would help improve care for these patients in a meaningful way.
“Monitoring and measurement of cough in clinical trials and routine care typically relies upon patients self-reporting the frequency, severity, and quality of their own coughs,” wrote lead author Lucia Kvapilova, a HealthMode product lead, and colleagues. “However, individuals’ self-perception is unavoidably influenced by perception bias (such as over- and under-perception of respiratory symptoms). This presents an opportunity for the development of objective cough monitoring systems.”
The authors developed an application (HealthMode Cough) that can continuously record audio data with a person’s own smartphone. Recordings are then sent to a server on the cloud where only authorized users can access them.
Convolutional neural networks (CNNs) were used to train the application to automatically assess audio recordings and detect coughing sounds. Cough sounds and “background noise” sounds used for training purposes were all taken from publicly available sources such as YouTube.
“Both cough sounds and background noises were split into non-overlapping training and test sample datasets (each cough-originator from the recordings was assigned to either the training or test set) and mixing was performed in each set separately,” the authors wrote. “A small portion of the training set was used as a validation set for hyperparameter tuning.”
Overall, the team found that their application could successfully collect continuous audio data, detecting coughs in said data with significant specificity. The team did note that more research is needed, but these findings “yielded encouraging results.” More detailed clinical studies are expected to take place in the near future.
Note: Kvapilova works as a representative from HealthMode. Researchers from the Albert Einstein College of Medicine and Tufts University School of Medicine also participated in this work.
Michael has more than 16 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.