Modeling Cardiac Cell Biophysics Using Long-Short-Term Memory Networks

Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) are being used to investigate mechanisms of various life-threatening cardiac disorders and drug effects. Long Short-Term Memory (LSTM) is a type of artificial neural network that has become popular for finding patterns from previous time steps to help predict the future response. In this paper, we developed LSTM-based predictive models to represent the electrophysiological response of a typical hiPSC-CM. Our model is able to reproduce experimentally observed effects on the action potential (AP) morphology for alterations in 5 main ionic currents in hiPSC-CMs. The proposed model achieved an impressive accuracy of 98.5% when compared to a detailed mathematical model of hiPSC-CM based on AP duration parameters (APD50, APD75, APD90). The proposed approach has the potential to reduce the computational time of multiscale cardiac electrophysiology simulations and can be used in mechanistic studies to investigate inherited arrhythmia syndromes and patient-specific drug therapies.

• Publication year

  2022

• Venue

  Annual Modeling and Simulation Conference

• Publication date

  2022-07-18

• Fields of study

  Medicine, Computer Science, Engineering

• Identifiers
  DOI 10.23919/ANNSIM55834.2022.9859442
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• An in silico hiPSC-Derived Cardiomyocyte Model Built With Genetic Algorithm

  2021cited by this paper
• A review of irregular time series data handling with gated recurrent neural networks

  2021cited by this paper
• All-Optical Electrophysiology Refines Populations of In Silico Human iPSC-CMs for Drug Evaluation

  2020cited by this paper
• Time series forecasting of petroleum production using deep LSTM recurrent networks

  2019cited by this paper
• Hyperparameters Tuning for Machine Learning Models for Time Series Forecasting

  2019cited by this paper
• A Review of Deep Learning Models for Time Series Prediction

  2019cited by this paper
• IK1-enhanced human-induced pluripotent stem cell-derived cardiomyocytes: an improved cardiomyocyte model to investigate inherited arrhythmia syndromes.

  2016cited by this paper
• The Comprehensive in Vitro Proarrhythmia Assay (CiPA) initiative - Update on progress.

  2016cited by this paper
• A review of unsupervised feature learning and deep learning for time-series modeling

  2014cited by this paper
• How to Construct Deep Recurrent Neural Networks

  2013cited by this paper
• Training and Analysing Deep Recurrent Neural Networks

  2013cited by this paper
• DEAP: evolutionary algorithms made easy

  2012cited by this paper
• On the difficulty of training recurrent neural networks

  2012cited by this paper
• High purity human-induced pluripotent stem cell-derived cardiomyocytes: electrophysiological properties of action potentials and ionic currents.

  2011cited by this paper
• Patient-specific induced pluripotent stem-cell models for long-QT syndrome.

  2010cited by this paper
• Many-Layered Learning

  2002cited by this paper
• Long Short-Term Memory

  1997cited by this paper

• No citing papers are available for this paper.