Impulse Data Model For Solving The Inverse Problem of Electrocardiography.

OBJECTIVE To develop, train and test neural networks for predicting heart surface potentials (HSPs) from body surface potentials (BSPs). The method re-frames traditional inverse problems of electrocardiograpy into regression problems, constraining the solution space by decomposing signals with multidimensional Gaussian impulse basis functions. METHODS Impulse HSPs were generated with single Gaussian basis functions at discrete heart surface locations and projected to corresponding BSPs using a volume conductor torso model. Both BSP (inputs) and HSP (outputs) were mapped to regular 2D surface meshes and used to train a neural network. Predictive capabilities of the network were tested with unseen synthetic and experimental data. RESULTS A dense full connected single hidden layer neural network was trained to map body surface impulses to heart surface Gaussian basis functions for reconstructing HSP. Synthetic pulses moving across the heart surface were predicted from the neural network with root mean squared error of 9.1 +/ 1.4%. Predicted signals were robust to noise up to 20 dB and errors due to displacement and rotation of the heart within the torso were bounded and predictable. A shift of the heart 40 mm toward the spine resulted in a 4% increase in signal feature localization error. The set of training impulse function data could be reduced and prediction error remained bounded. Recorded HSPs from in-vitro pig hearts were reliably decomposed using space-time Gaussian basis functions. Activation times calculated from predicted HSPs for left-ventricular pacing had a mean absolute error of 10.4 +/ 11.4 ms. Other pacing scenarios were analyzed with similar success. CONCLUSION Impulses from Gaussian basis functions are potentially an effective and robust way to train simple neural network data models for reconstructing HSPs from decomposed BSPs. SIGNIFICANCE The HSPs predicted by the neural network can be used to generate activation maps that non-invasively identify features of cardiac electrical dysfunction and can guide subsequent treatment options.

[1]  L. Bear,et al.  Evaluation of Fifteen Algorithms for the Resolution of the Electrocardiography Imaging Inverse Problem Using ex-vivo and in-silico Data , 2018, Front. Physiol..

[2]  Patrick E. McSharry,et al.  A dynamical model for generating synthetic electrocardiogram signals , 2003, IEEE Transactions on Biomedical Engineering.

[3]  Nenad Sarapa,et al.  Automatic analysis of cardiac repolarization morphology using Gaussian mesa function modeling. , 2008, Journal of electrocardiology.

[4]  Ken-ichi Funahashi,et al.  On the approximate realization of continuous mappings by neural networks , 1989, Neural Networks.

[5]  Leo K. Cheng,et al.  The inverse problem of electrocardiography , 2011 .

[6]  J. Horvath,et al.  Deep learning neural nets for detecting heart activity , 2019, 1901.09831.

[7]  A. van Oosterom,et al.  Non-Invasive Imaging of Cardiac Activation and Recovery , 2009, Annals of Biomedical Engineering.

[8]  Rob S. MacLeod,et al.  Inverse electrocardiography by simultaneous imposition of multiple constraints , 1999, IEEE Transactions on Biomedical Engineering.

[9]  Y. Rudy,et al.  Noninvasive Reconstruction of Epicardial Measures , 2001 .

[10]  Yesim Serinagaoglu Dogrusoz,et al.  The Impact of Torso Signal Processing on Noninvasive Electrocardiographic Imaging Reconstructions , 2020, IEEE Transactions on Biomedical Engineering.

[11]  Mostafa Bendahmane,et al.  A Spatial Adaptation of the Time Delay Neural Network for Solving ECGI Inverse Problem , 2019, FIMH.

[12]  Maxime Sermesant,et al.  Deep learning formulation of electrocardiographic imaging integrating image and signal information with data-driven regularization. , 2021, Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European Society of Cardiology.

[13]  P. Savard,et al.  Forward and inverse problems of electrocardiography: modeling and recovery of epicardial potentials in humans , 1994, IEEE Transactions on Biomedical Engineering.

[14]  Y. Rudy,et al.  Noninvasive Electrocardiographic Imaging , 1999 .

[15]  T Peng,et al.  Predictive modeling of drug effects on electrocardiograms , 2019, Comput. Biol. Medicine.

[16]  Ebadollah Kheirati Roonizi,et al.  Morphological modeling of cardiac signals based on signal decomposition , 2013, Comput. Biol. Medicine.

[17]  Avinash Malik,et al.  Parametric Modeling of Electrocardiograms using Particle Swarm optimization , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[18]  Rajkumar Thirumalainambi,et al.  Training data requirement for a neural network to predict aerodynamic coefficients , 2003, SPIE Defense + Commercial Sensing.

[19]  Jian Wu,et al.  Multiscale-Wavelet Regularization Method for the Inverse Problem of Electrocardiography , 2019, 2019 International Conference on Medical Imaging Physics and Engineering (ICMIPE).

[20]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1989, Math. Control. Signals Syst..

[21]  Y. Rudy,et al.  Regularization of the inverse problem in electrocardiography: A model study , 1988 .

[22]  Rémi Dubois,et al.  Cardiac electrical dyssynchrony is accurately detected by noninvasive electrocardiographic imaging. , 2018, Heart rhythm.

[23]  B. Taccardi,et al.  Finite element approximation of regularized solutions of the inverse potential problem of electrocardiography and applications to experimental data , 1985 .

[24]  B. Taccardi,et al.  A mathematical procedure for solving the inverse potential problem of electrocardiography. analysis of the time-space accuracy from in vitro experimental data , 1985 .

[25]  Leo K. Cheng,et al.  Forward Problem of Electrocardiography: Is It Solved? , 2015, Circulation. Arrhythmia and electrophysiology.

[26]  Bruce H Smaill,et al.  Laminar Arrangement of Ventricular Myocytes Influences Electrical Behavior of the Heart , 2007, Circulation research.

[27]  Y. Rudy,et al.  Application of L1-Norm Regularization to Epicardial Potential Solution of the Inverse Electrocardiography Problem , 2009, Annals of Biomedical Engineering.

[28]  Feng Liu,et al.  A Hybrid Model of Maximum Margin Clustering Method and Support Vector Regression for Noninvasive Electrocardiographic Imaging , 2012, Comput. Math. Methods Medicine.

[29]  Hervé Delingette,et al.  Transfer Learning From Simulations on a Reference Anatomy for ECGI in Personalized Cardiac Resynchronization Therapy , 2019, IEEE Transactions on Biomedical Engineering.

[30]  Avinash Malik,et al.  A machine learning approach to reconstruction of heart surface potentials from body surface potentials , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[31]  Joël M. H. Karel,et al.  CT-Scan Free Neural Network-Based Reconstruction of Heart Surface Potentials From ECG Recordings , 2020, 2020 Computing in Cardiology.

[32]  Nejib Zemzemi,et al.  A machine learning regularization of the inverse problem in electrocardiography imaging , 2013, Computing in Cardiology 2013.

[33]  Linwei Wang,et al.  Improving Generalization of Deep Networks for Inverse Reconstruction of Image Sequences , 2019, IPMI.

[34]  B. V. Van Veen,et al.  A spatial feature extraction and regularization model for the head-related transfer function. , 1995, The Journal of the Acoustical Society of America.

[35]  Joël M. H. Karel,et al.  Wavelet-promoted sparsity for non-invasive reconstruction of electrical activity of the heart , 2018, Medical & Biological Engineering & Computing.

[36]  Matija Milanič,et al.  Assessment of regularization techniques for electrocardiographic imaging. , 2014, Journal of electrocardiology.