Real-Time Optical Mapping of Contracting Cardiac Tissues With GPU-Accelerated Numerical Motion Tracking

Area of Science:

• Cardiovascular Physiology
• Biomedical Imaging
• Computational Biology

Background:

• Optical mapping of cardiac action potentials and calcium transients is hindered by motion artifacts.
• Accurate tracking of fluorescence changes in moving cardiac tissue is crucial for reliable electrophysiological and mechanical measurements.
• Current motion compensation techniques are computationally intensive and performed offline, limiting widespread adoption.

Purpose of the Study:

• To evaluate and compare the performance of open-source, GPU-accelerated numerical motion-tracking algorithms for optical mapping data.
• To assess the effectiveness of these algorithms in tracking, stabilizing, and compensating for motion in cardiac optical mapping videos.
• To determine the potential for real-time processing and widespread use of motion tracking in routine optical mapping studies.

Main Methods:

• Evaluation of 5 open-source numerical motion-tracking algorithms implemented on Graphics Processing Units (GPUs).
• Comparison of algorithm performance on synthetic and real optical mapping videos of contracting cardiac tissues (human cell cultures, mouse, rabbit, pig hearts).
• Assessment of tracking accuracy, motion stabilization, sensitivity to fluorescence signals and noise, and processing speed.

Main Results:

• GPU-accelerated motion tracking substantially reduces video processing times for optical mapping.
• Real-time motion tracking and stabilization are achievable for both low and high-resolution videos.
• The Farnebäck algorithm demonstrated efficacy in motion compensation across various species and imaging conditions.

Conclusions:

• GPU-accelerated numerical motion tracking significantly enhances processing speed for optical mapping.
• This advancement facilitates real-time motion compensation, enabling more detailed simultaneous measurements of electrophysiological phenomena and tissue mechanics.
• The findings pave the way for broader implementation of motion tracking and stabilization in routine cardiac optical mapping research.