Description:
The University of Texas at Dallas presents a real-time, adaptive technique that provides robust signal enhancement against motion artifact for wearable photoplethysmographic (PPG) systems, enabling the reliable extraction of heart rate and oxygen saturation.
Despite the benefits of using PPG biosensors, most of the useful information they capture can be distorted or concealed by motion artifact. PPG biosensors capture changes in optical density that are recorded as voltage/current signals, and these PPG signals are known to be sensitive to pulsatile blood flow and capture the peripheral pulses. By carefully establishing an adaptive, mathematical approach to formulate optical density, the presented approach is able to accurately remove motion artifacts, such as tissue effect and venous effect, without any additional hardware. Additional sources of motion artifact include tissue effect, muscular activity and the respiratory pump, wherein body movements introduce a new rhythmic pattern and variation in the PPG signal and can lead to an inaccurate and unreliable fundamental period.
The key contribution of the presented adaptive algorithm is the separation of an extremely clean signal corresponding to arterial blood flow – significantly enhancing the accuracy of further extracted features, including the fundamental period (heart rate), oxygen saturation, and separation of the venous signal. This has been demonstrated in active motion experiments, as oxygen saturation values extracted from the original signal were unreliable, but the algorithm-enhanced signals provided robust computation of oxygen saturation.
Technical Summary:
In order to appropriately address motion artifact issues, the proposed technology utilizes a novel two-stage adaptive algorithm that efficiently removes the effect of tissue, venous blood noise, and motion artifact. This approach implements mathematical models based on optical density to define the optimum reference signals for removal of both tissue effect during motion and venous blood movement. Additional signal processing techniques provide additional noise reduction/removal and signal enhancement for accurate extraction of heart rate and oxygen saturation.
The algorithm provides an extremely clean signal corresponding to arterial blood flow - this allows the use of a conventional technique, such as ratio of ratios, to accurately extract the oxygen saturation value. Comparative analysis with the Discrete Saturation Transform (DST) algorithm reveals that the described algorithm provides reliable pulse oximeter oxygen saturation (SpO2) readings with higher correlation - our algorithm outperforms DST, with significant improvement for high levels of motion artifact. Experimental results further validated reliable extraction of heart rate and oxygen saturation with correlation of more than 0.98 and 0.7, respectively, compared to reference stationary sensors in the presence of the motion artifact.
Value Proposition:
The presented approach is demonstrated to effectively enhance PPG biosensor signal quality in the presence of motion artifact, achieving reliable separation of both arterial and venous signals with a very high accuracy and correlation.
Applications:
- Non-invasive monitoring devices & wearables
- Diagnostic applications: Hypoxemia, blood pressure, sleep apnea, heart abnormalities, lung function test, hemodynamic monitoring, venous capacity, fluid responsiveness, seizure alerts, wound healing
- Real-time monitoring: of arterial signal and central venous pressure in anesthesia, preoperative and postoperative care
- Continuous tissue perfusion monitoring: for implanted organs after surgery
Key Benefits:
- Accurate – Optimized, adaptive removal of motion artifact prevents the deterioration of signal integrity for PPG biosensors, enabling their use wherever motion is present
- Robust – Provides adaptive removal of tissue and venous effect (noise) to produce reliable computation of SpO2 and heart rate; demonstrates more reliable SpO2 with higher accuracy and correlation than commercial solution (DST method)
- Tunable – May be adjusted, based on features meant to be extracted from enhanced output signals, to further reduce effect of rhythm irregularities on reference noise
Publication:
Yousefi, Rasoul, et al. “A Motion-Tolerant Adaptive Algorithm for Wearable Photoplethysmographic Biosensors.” IEEE Journal of Biomedical and Health Informatics, vol. 18, no. 2, 20 May 2013, pp. 670–681., doi:10.1109/jbhi.2013.2264358.
IP Status: US patent 9,918,666 issued.
Licensing Opportunity: This technology is available for exclusive or non-exclusive licensing.
ID Number: MP-13013
Contact: otc@utdallas.edu