Deep learning approaches for plethysmography signal quality assessment in the presence of atrial fibrillation

OBJECTIVE Photoplethysmography (PPG) monitoring has been implemented in many portable and wearable devices we use daily for health and fitness tracking. Its simplicity and cost-effectiveness enabled a variety of biomedical applications, such as continuous long-term monitoring of heart arrhythmias, fitness, and sleep tracking, and hydration monitoring. One major issue that can hinder PPG-based applications is movement artifacts, which can lead to false interpretations. In many implementations, noisy PPG signals are discarded. Misinterpreted or discarded PPG signals pose a problem in applications where the goal is to increase the yield of detecting physiological events, such as in the case of paroxysmal atrial fibrillation (AF) - a common episodic heart arrhythmia and a leading risk factor for stroke. In this work, we compared a traditional machine learning and deep learning approaches for PPG quality assessment in the presence of AF, in order to find the most robust method for PPG quality assessment. APPROACH The training data set was composed of 78278 30-second long PPG recordings from 3764 patients using bedside patient monitors. Two different representations of PPG signals were employed - a time-series based (1D) and an image-based (2D). Trained models were tested on an independent set of 2683 30-second PPG signals from 13 stroke patients. MAIN RESULTS ResNet18 showed a higher performance (0.985 accuracy, 0.979 specificity, and 0.988 sensitivity) than SVM and other deep learning approaches. 2D-based models were generally more accurate than 1D-based models. SIGNIFICANCE 2D representation of PPG signal enhances the accuracy of PPG signal quality assessment.

[1]  M. Elgendi,et al.  Photoplethysmography and Deep Learning: Enhancing Hypertension Risk Stratification , 2018, Biosensors.

[2]  Giuseppe Baselli,et al.  A review of methods for the signal quality assessment to improve reliability of heart rate and blood pressures derived parameters , 2016, Medical & Biological Engineering & Computing.

[3]  H. Kamel,et al.  Atrial Fibrillation and Mechanisms of Stroke: Time for a New Model , 2016, Stroke.

[4]  Dragica Radosav,et al.  Deep Learning and Medical Diagnosis: A Review of Literature , 2018, Multimodal Technol. Interact..

[5]  W Karlen,et al.  Photoplethysmogram signal quality estimation using repeated Gaussian filters and cross-correlation , 2012, Physiological measurement.

[6]  Houshang Darabi,et al.  LSTM Fully Convolutional Networks for Time Series Classification , 2017, IEEE Access.

[7]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[8]  David J Gladstone,et al.  Atrial fibrillation in patients with cryptogenic stroke. , 2014, The New England journal of medicine.

[9]  Joshua C. Denny,et al.  Combining billing codes, clinical notes, and medications from electronic health records provides superior phenotyping performance , 2016, J. Am. Medical Informatics Assoc..

[10]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[11]  P. Kirchhof,et al.  2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. , 2016, European heart journal.

[12]  Jo Woon Chong,et al.  Photoplethysmograph Signal Reconstruction Based on a Novel Hybrid Motion Artifact Detection–Reduction Approach. Part I: Motion and Noise Artifact Detection , 2014, Annals of Biomedical Engineering.

[13]  Xiao Hu,et al.  A Supervised Approach to Robust Photoplethysmography Quality Assessment , 2020, IEEE Journal of Biomedical and Health Informatics.

[14]  Tim Oates,et al.  Time series classification from scratch with deep neural networks: A strong baseline , 2016, 2017 International Joint Conference on Neural Networks (IJCNN).

[15]  N H Lovell,et al.  Signal quality measures for pulse oximetry through waveform morphology analysis , 2011, Physiological measurement.

[16]  Pasi Liljeberg,et al.  A Real-time PPG Quality Assessment Approach for Healthcare Internet-of-Things , 2019, ANT/EDI40.

[17]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[18]  Sasank Chilamkurthy,et al.  Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study , 2018, The Lancet.

[19]  B. Drew,et al.  Insights into the Problem of Alarm Fatigue with Physiologic Monitor Devices: A Comprehensive Observational Study of Consecutive Intensive Care Unit Patients , 2014, PloS one.

[20]  P. Kirchhof,et al.  2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. , 2016, European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery.

[21]  Xiao Hu,et al.  Monitoring significant ST changes through deep learning. , 2018, Journal of electrocardiology.

[22]  Xiao Hu,et al.  A Robust Approach Toward Recognizing Valid Arterial-Blood-Pressure Pulses , 2010, IEEE Transactions on Information Technology in Biomedicine.

[23]  Arvid Lundervold,et al.  An overview of deep learning in medical imaging focusing on MRI , 2018, Zeitschrift fur medizinische Physik.

[24]  William Speier,et al.  Evaluating utility and compliance in a patient-based eHealth study using continuous-time heart rate and activity trackers , 2018, J. Am. Medical Informatics Assoc..

[25]  Jaques Reifman,et al.  Application of Information Technology: A Method for Automatic Identification of Reliable Heart Rates Calculated from ECG and PPG Waveforms , 2006, J. Am. Medical Informatics Assoc..

[26]  Q. Mcnemar.  Note on the sampling error of the difference between correlated proportions or percentages , 1947, Psychometrika.

[27]  Nima Tajbakhsh,et al.  Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning? , 2016, IEEE Transactions on Medical Imaging.

[28]  H. Ko,et al.  Atrial Fibrillation on Intensive Care Unit Admission Independently Increases the Risk of Weaning Failure in Nonheart Failure Mechanically Ventilated Patients in a Medical Intensive Care Unit , 2016, Medicine.

[29]  Xiao Hu,et al.  Robust Assessment of Photoplethysmogram Signal Quality in the Presence of Atrial Fibrillation , 2018, 2018 Computing in Cardiology Conference (CinC).

[30]  Q Li,et al.  Dynamic time warping and machine learning for signal quality assessment of pulsatile signals , 2012, Physiological measurement.

[31]  Qiao Li,et al.  A deep learning approach to monitoring and detecting atrial fibrillation using wearable technology , 2017, 2017 IEEE EMBS International Conference on Biomedical & Health Informatics (BHI).

[32]  Qiao Li,et al.  Evaluation of the accuracy and noise response of an open-source pulse onset detection algorithm on pulsatile waveform databases , 2016, 2016 Computing in Cardiology Conference (CinC).

[33]  P. Wolf,et al.  Epidemiologic assessment of chronic atrial fibrillation and risk of stroke , 1978, Neurology.

[34]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[35]  David A. Clifton,et al.  Signal-Quality Indices for the Electrocardiogram and Photoplethysmogram: Derivation and Applications to Wireless Monitoring , 2015, IEEE Journal of Biomedical and Health Informatics.