Wrist Photoplethysmography Signal Quality Assessment for Reliable Heart Rate Estimate and Morphological Analysis

Abstract

Photoplethysmographic (PPG) signals are mainly employed for heart rate estimation but are also fascinating candidates in the search for cardiovascular biomarkers. However, their high susceptibility to motion artifacts can lower their morphological quality and, hence, affect the reliability of the extracted information. Low reliability is particularly relevant when signals are recorded in a real-world context, during daily life activities. We aim to develop two classifiers to identify PPG pulses suitable for heart rate estimation (Basic-quality classifier) and morphological analysis (High-quality classifier). We collected wrist PPG data from 31 participants over a 24 h period. We defined four activity ranges based on accelerometer data and randomly selected an equal number of PPG pulses from each range to train and test the classifiers. Independent raters labeled the pulses into three quality levels. Nineteen features, including nine novel features, were extracted from PPG pulses and accelerometer signals. We conducted ten-fold cross-validation on the training set (70%) to optimize hyperparameters of five machine learning algorithms and a neural network, and the remaining 30% was used to test the algorithms. Performances were evaluated using the full features and a reduced set, obtained downstream of feature selection methods. Best performances for both Basic- and High-quality classifiers were achieved using a Support Vector Machine (Acc: 0.96 and 0.97, respectively). Both classifiers outperformed comparable state-of-the-art classifiers. Implementing automatic signal quality assessment methods is essential to improve the reliability of PPG parameters and broaden their applicability in a real-world context.

Photoplethysmography Driven Hypertension Identification: A Pilot Study

“…Traditional machine learning methods have been extensively studied in PPG [ 34 , 35 , 36 , 37 ], and this study also compares SVM and KNN classification models with the LSTM-Attention model. Unlike neural networks, which can directly train on raw signals, feature extraction is required before training these two models.…”

Cardiorespiratory Assessments in Panic Disorder Facilitated by Wearable Devices: A Systematic Review and Brief Comparison of the Wearable Zephyr BioPatch with the Quark-b2 Stationary Testing System

“…Since the HR-measurement algorithms of smart sports-watches are generally proprietary, access to their recorded raw data might be more useful to researchers, allowing them to perform additional analyses and identify more reliable patterns of cardiac activity under different conditions. Finally, due to the generally high susceptibility of PPG-based devices to motion artifacts, several signal-processing techniques, also using simultaneously recorded accelerometer data, may remove motion-artifact effects from PPG signals and improve the detection of PPG pulses, making them suitable for performing HR estimates under various movement intensities [ 40 , 66 ].…”

Upper-Arm Photoplethysmographic Sensor with One-Time Calibration for Long-Term Blood Pressure Monitoring

“…To assess PPG signal quality, 120 PPG pulses measured by both upper-arm and wrist-type BPMs from each subject were randomly selected and assigned to one of the three quality levels by three raters, defined as follows [ 39 ]: Fair—systolic and diastolic peaks cannot be easily distinguished from noise. Good—the systolic peak is clearly detectable, but the diastolic peak is not.…”