A multistage deep neural network model for blood pressure estimation using photoplethysmogram signals

OBJECTIVE Easy access bio-signals are useful to alleviate the shortcomings and difficulties of cuff-based and invasive blood pressure (BP) measuring techniques. This study proposes a multistage model based on deep neural networks to estimate systolic and diastolic blood pressures using the photoplethysmogram (PPG) signal. METHODS The proposed model consists of two key ingredients, using two successive stages. The first stage includes two convolutional neural networks (CNN) to extract morphological features from each PPG segment and then to estimate systolic and diastolic BPs separately. The second stage relies on long short-term memory (LSTM) to capture temporal dependencies. Further, the method incorporates the dynamic relationship between systolic and diastolic BPs to improve accuracy. RESULTS The proposed multistage model was evaluated on 200 subjects using the standards of the British Hypertension Society (BHS) and the Association for the Advancement of Medical Instrumentation (AAMI). The results revealed that our model performance met the requirements of the AAMI standard. Also, according to the BHS standard, it achieved grade A in estimating both systolic and diastolic BPs. The mean and standard deviation of error for systolic and diastolic blood pressure estimations were +1.91±5.55mmHg and +0.67±2.84mmHg, respectively. CONCLUSION Our results highlight the benefits of the proposed model in terms of appropriate feature extraction as well as estimation consistency.

[1]  Dursun Delen,et al.  Predicting hospital readmission for lupus patients: An RNN-LSTM-based deep-learning methodology , 2018, Comput. Biol. Medicine.

[2]  E. O’Brien,et al.  The British Hypertension Society protocol for the evaluation of automated and semi-automated blood pressure measuring devices with special reference to ambulatory systems. , 1990, Journal of hypertension.

[3]  John Allen.  Photoplethysmography and its application in clinical physiological measurement , 2007, Physiological measurement.

[4]  Giuseppe Mancia,et al.  Short- and long-term blood pressure variability: present and future. , 2012, Hypertension.

[5]  D. Jamison,et al.  Disease Control Priorities in Developing Countries , 1993 .

[6]  Mahmood A. Rashid,et al.  Protein secondary structure prediction using neural networks and deep learning: A review , 2019, Comput. Biol. Chem..

[7]  N. Selvaraj,et al.  Assessment of heart rate variability derived from finger-tip photoplethysmography as compared to electrocardiography , 2008, Journal of medical engineering & technology.

[8]  T. H. Kyaw,et al.  Multiparameter Intelligent Monitoring in Intensive Care II: A public-access intensive care unit database* , 2011, Critical care medicine.

[9]  Mingshan Sun,et al.  Optical blood pressure estimation with photoplethysmography and FFT-based neural networks. , 2016, Biomedical optics express.

[10]  Mateusz Buda,et al.  Association of genomic subtypes of lower-grade gliomas with shape features automatically extracted by a deep learning algorithm , 2019, Comput. Biol. Medicine.

[11]  Md. Kamrul Hasan,et al.  Cuffless blood pressure estimation from electrocardiogram and photoplethysmogram using waveform based ANN-LSTM network , 2018, Biomed. Signal Process. Control..

[12]  Guanglin Li,et al.  New photoplethysmogram indicators for improving cuffless and continuous blood pressure estimation accuracy , 2018, Physiological measurement.

[13]  Mitja Lustrek,et al.  Continuous Blood Pressure Estimation from PPG Signal , 2018, Informatica.

[14]  Feng Liu,et al.  Deep Learning and Its Applications in Biomedicine , 2018, Genom. Proteom. Bioinform..

[15]  A. G. Ramakrishnan,et al.  Threshold-Independent QRS Detection Using the Dynamic Plosion Index , 2014, IEEE Signal Processing Letters.

[16]  U. Rajendra Acharya,et al.  Deep learning for healthcare applications based on physiological signals: A review , 2018, Comput. Methods Programs Biomed..

[17]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[18]  Matjaz Gams,et al.  Non-Invasive Blood Pressure Estimation from ECG Using Machine Learning Techniques , 2018, Sensors.

[19]  M. Landowne,et al.  A Method Using Induced Waves to Study Pressure Propagation in Human Arteries , 1957, Circulation research.

[20]  Vaidotas Marozas,et al.  Modeling of the photoplethysmogram during atrial fibrillation , 2017, Comput. Biol. Medicine.

[21]  Mohammad Kachuee,et al.  Non-invasive blood pressure estimation using phonocardiogram , 2017, 2017 IEEE International Symposium on Circuits and Systems (ISCAS).

[22]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[23]  Mahdi Shabany,et al.  Cuffless Blood Pressure Estimation Algorithms for Continuous Health-Care Monitoring , 2017, IEEE Transactions on Biomedical Engineering.

[24]  Wei Zhang,et al.  Assessing cognitive mental workload via EEG signals and an ensemble deep learning classifier based on denoising autoencoders , 2019, Comput. Biol. Medicine.

[25]  Giacomo Pucci,et al.  The dynamic relationship between systolic and diastolic blood pressure: yet another marker of vascular aging? , 2010, Hypertension Research.

[26]  B. Alpert,et al.  Validation of the Welch Allyn Spot Vital Signs blood pressure device according to the ANSI/AAMI SP10: 2002. Accuracy and cost-efficiency successfully combined , 2007, Blood pressure monitoring.

[27]  Hiroshi Motoda,et al.  Computational Methods of Feature Selection , 2022 .

[28]  T. Dawber,et al.  Characteristics of the Dicrotic Notch of the Arterial Pulse Wave in Coronary Heart Disease , 1973, Angiology.

[29]  Qiao Zhang,et al.  Cuff-free blood pressure estimation using signal processing techniques , 2010 .

[30]  Kwang Suk Park,et al.  Second Derivative of Photoplethysmography for Estimating Vascular Aging , 2007, 2007 6th International Special Topic Conference on Information Technology Applications in Biomedicine.

[31]  Michael Bursztyn,et al.  Linear relationship between systolic and diastolic blood pressure monitored over 24 h: assessment and correlates , 2008, Journal of hypertension.

[32]  Christina Orphanidou,et al.  Derivation of respiration rate from ambulatory ECG and PPG using Ensemble Empirical Mode Decomposition: Comparison and fusion , 2017, Comput. Biol. Medicine.

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

[34]  Yuan-Ting Zhang,et al.  Continuous Cuffless Blood Pressure Estimation Using Pulse Transit Time and Photoplethysmogram Intensity Ratio , 2016, IEEE Transactions on Biomedical Engineering.

[35]  T. L. Rusch,et al.  Signal processing methods for pulse oximetry , 1996, Comput. Biol. Medicine.

[36]  Qiang Fang,et al.  Continuous non-invasive blood pressure monitoring using photoplethysmography: A review , 2015, 2015 International Symposium on Bioelectronics and Bioinformatics (ISBB).

[37]  U. Rajendra Acharya,et al.  Deep convolutional neural network for the automated detection and diagnosis of seizure using EEG signals , 2017, Comput. Biol. Medicine.

[38]  L A Geddes,et al.  Pulse transit time as an indicator of arterial blood pressure. , 1981, Psychophysiology.

[39]  E. Y. K. Ng,et al.  SVR ensemble-based continuous blood pressure prediction using multi-channel photoplethysmogram , 2019, Comput. Biol. Medicine.

[40]  Rong Zhang,et al.  Arterial blood pressure feature estimation using photoplethysmography , 2018, Comput. Biol. Medicine.

[41]  Gan Kok Beng,et al.  Applicability of adaptive noise cancellation to fetal heart rate detection using photoplethysmography , 2008, Comput. Biol. Medicine.