Queen’s University Background
Electrocardiogram (ECG/EKG) is the electrical measurement of cardiac activity. It is a common and non-invasive method used to accurately detect heart problems and monitor cardiac health. ECG/EKG devices are typically used in hospitals and doctor’s offices. There are also a number of portable ECG/EKG devices for short-term ambulatory monitoring. However, there are limited options for long-term or continuous wearable ECG/EKG monitoring and daily use.
Existing wearables that monitor heart rate use a photoplethysmogram (PPG) sensor which measures volumetric changes in blood circulation to calculate heart rate. While used widely in the wearables industry and considered a close alternative to ECGs/EKGs by some, PPGs suffer from inaccurate heart rate estimation and several other limitations in comparison to conventional ECG/EKG monitoring devices due to factors like skin tone, diverse skin types, motion artefacts, and signal crossovers. Moreover, PPGs lack important information about cardiac activity and are typically not used by cardiologists to assess the condition and performance of the heart. Still, the need for two sensors in ECG/EKG monitoring has made PPG the technique of choice for heart monitoring in the wearables industry.
To address the limitations of heart monitoring using PPG, researchers at Queen’s University have developed a technique for generating detailed ECG/EKG signals from corresponding PPG signals using a deep learning approach. The method uses a generative adversarial network (GAN) to learn both time-domain and frequency-domain information in order to map PPG to ECG/EKG while attending to salient features.
In ongoing research by the team, the advantages of using the generated ECG/EKG data in other areas where the use of PPG is limited is being evaluated. These areas include identification of cardiovascular diseases, detection of abnormal heart rhythms, and others. Furthermore, generating multi-lead ECG can also be studied in order to extract more useful cardiac information often missing in single-channel ECG recordings. The research can open a new path towards cross-modality signal-to-signal translation in the biosignal domain, allowing for less available physiological recordings to be generated from more affordable and readily available signals.
Stage of Development
The technology is at the proof-of-concept stage.
Enables the generation of detailed ECG/EKG signals using standard PPG sensors
Eliminates the need for multiple sensors
Can be incorporated into any wearable device that incorporates PPG sensors, without any additional hardware required
Continuous ECG/EKG monitoring for 24/7 use
More accurate heart-rate estimation and the potential to detect cardiac diseases where PPG fails
The technology has application in the wearables market for health monitoring/tracking. It could be incorporated into smartwatches, health monitoring bracelets, or other devices that have a PPG sensor.
Queen’s University are seeking companies that are interested in developing and commercializing the technology.