Automated strength-interval curve generation using actors.

Area of Science:

• Computational electrophysiology
• Biophysics
• Computer science

Background:

• Strength-interval (SI) curves quantify excitable tissue response to electrical stimuli.
• Traditionally, SI curves are experimentally derived, a process that is labor-intensive and time-consuming.
• Computational simulation offers a more convenient alternative but can be computationally demanding.

Purpose of the Study:

• To automate the generation of SI curves using the Actor Model of concurrent computation.
• To reduce the computational burden and manual researcher intervention in SI curve generation.
• To maximize the utilization of available computational resources.

Main Methods:

• Utilized the C++ Actor Framework to automate the openCARP simulation platform.
• Employed sophisticated parallelization techniques, including dynamic information passing between parallel simulations via actors.
• Implemented dynamic monitoring, assessment, and reallocation of computational resources based on actor states.

Main Results:

• Generated a bidomain SI curve with 31 data points in 15.4 hours, a 40% improvement over conventional methods (27.5 hours with 80 CPU cores).
• Achieved full automation, eliminating the need for manual researcher intervention.
• Demonstrated faster computation compared to traditional parallel programming techniques like MPI.

Conclusions:

• The Actor Model offers novel parallelization techniques that significantly enhance the efficiency of computational SI curve generation.
• This improved efficiency impacts future studies on cardiac and other excitable tissues.
• Enables rapid generation of general and patient-specific SI curves for designing and testing therapeutic tools, such as personalized pacemakers.