Real-Time GPU-Accelerated Digital Heart Twin: Integrating Bidirectional Interactions Between Living Optogenetic Monolayers and Computational Simulations

Reentrant arrhythmias are life-threatening cardiac events that are difficult to study due to limited experimental control over the complex circuit dynamics. We aimed to develop a real-time digital heart twin that allows in real-time dynamic manipulation of reentrant pathways in vitro using physiologically relevant simulations. We designed a closed-feedback loop system that couples a cultured cardiac monolayer with a two-dimensional computational simulation of cardiac tissue. The simulation, based on GPU-accelerated models (e.g., cellular automata), predicts wave propagation in real-time using the Abubu.js library. Optical mapping captures monolayer activation patterns, and simulation outputs are converted into light-based stimulation via optogenetics, using LEDs and microcontrollers to depolarize cardiac tissue. Our platform is capable of accurately detecting and responding to electrical waves in real-time, enabling interactive control of reentrant circuits. The system replaces traditional fixed-delay stimulation protocols with computationally guided interventions, better mimicking physiological conduction dynamics. This digital twin provides a novel and responsive method to study reentrant arrhythmias. Its integration of optical stimulation, real-time modeling, and tissue feedback enables the construction of user-defined reentry pathways under dynamic control. By merging computational and biological systems, this work introduces a versatile experimental framework for investigating arrhythmias. The platform may inform future control and anti-arrhythmic strategies and pave the way for personalized cardiac electrophysiology studies.