Smartwatch algorithm to detect viral infections

Smartwatch algorithm to detect viral infections

Technology News |
Engineers at Purdue University and digital medicine specialist physIQ have announced the co-development of a viral detection algorithm for smartwatches.
By Rich Pell

Share:

The algorithm, which is designed to detect early signs of viral infections, including COVID-19, will be commercialized by physIQ, which develops solutions designed to improve health care outcomes by applying artificial intelligence (AI) to real-time physiological data from wearable sensors.

“Smartwatches are well-suited for the detection of early viral infection, including COVID-19,” says Craig Goergen, Purdue’s Leslie A. Geddes Associate Professor of Biomedical Engineering. “Infections can happen at any time, making the continuously tracked data available through an individual’s smartwatches uniquely suited to identify the earliest signs of illness. In particular, knowledge of a person’s usual heart rate and respiratory during sleep and activity over long periods of time is especially valuable for detecting subtle changes from normal.”

The research involved a study of 100 participants, including Purdue students, staff and faculty, to determine whether wearing a smartwatch to collect data was practical, unobtrusive and user-friendly. Each participant received a Samsung Galaxy smartwatch with a pre-loaded physIQ app to collect data.

Along with the smartwatch, the research participants also wore FDA-cleared adhesive chest-based biosensors to capture a single-lead electrocardiogram signal and multiple other parameters for five days of continuous monitoring. Goergen’s lab analyzed data from the app remotely using physIQ’s cloud-based accelerateIQ platform.

Data from the chest patches were processed by physIQ’s FDA-cleared AI-based algorithms in deriving heart rate, respiration rate and heart rate variability. These data served as “gold standard” references to compare with data from the smartwatches.

“The algorithms for enabling early detection are built off physiological features derived from the biosensor data collected by the smartwatches,” says Stephan Wegerich, physIQ’s chief science officer. “Generating accurate and robust physiological features forms the input to subsequent viral detection algorithms. This requires the development of sophisticated signal processing and machine learning algorithms. Combined, these make the most out of smartwatch biosensor data, which is a big part of our collaboration with Purdue.”

The viral infection detection algorithm complements its other health care applications, says the company. The goal across of all physIQ’s applications is the ability to characterize dynamic human physiology over time, whether it is for assessing the efficacy of a new therapy, safety monitoring during treatment or general wellness.

Dr. Steve Steinhubl, physIQ’s chief medical officer and Purdue alumnus says, “The collaborative nature of our relationship and work with Purdue University has the potential to greatly expand physIQ’s physiological monitoring applications that can be targeted to a wide range of clinical needs using the pinpointIQ and accelerateIQ platforms.”

physIQ

Related articles:
Wearable health AI system aims to detect COVID-19 before symptoms
COVID-19 study to use continuous biosensor data, AI
Biosensor study to measure early signs of infectious disease
‘Revolutionary’ lab-on-chip detects viruses, contaminants
Machine learning speeds search for unknown viruses

 

Linked Articles
Smart2.0
10s