Project Abstract

Analyzing cardiomyocyte dynamic networks is crucial for advancing cardiac research, particularly in drug discovery, disease modeling, and tissue engineering, yet manual analysis remains labor-intensive and subjective. This work leverages Cellpose, a deep learning-based segmentation tool, to automate and enhance the study of cardiomyocyte networks, enabling high-throughput, quantitative analysis of cell morphology and interactions. The research addresses a critical gap by combining machine learning with cardiac cell biology to extract previously inaccessible spatial and temporal network dynamics. The study employs time-lapse microscopy of cardiomyocyte cultures, processed using Cellpose for robust cell segmentation, followed by feature extraction and machine learning to analyze structural and functional network properties. Key findings include the identification of distinct cardiomyocyte subpopulations based on network connectivity, quantification of synchronization patterns, and detection of abnormal propagation dynamics under pharmacological perturbations. The analysis pipeline successfully produced automated segmentation of cardiomyocyte networks with high accuracy, along with quantitative metrics for network organization and functional behavior. This work contributes a scalable, data-driven framework for studying cardiomyocyte networks, providing researchers with new tools to assess tissue-level cardiac function, predict arrhythmic risk, and optimize engineered cardiac tissues. The approach demonstrates how machine learning can uncover hidden patterns in complex biological systems, advancing both computational biology and cardiac research.

Keywords: Machine Learning, deep learning, cell segmentation

 Conference Details

Session: A

Location: Sir Stanley Clarke Auditorium at 11:00 13:00

Markers: Fabio Caraffini, Casey Hopkins

Course: BSc Computer Science 3yr FT

Future Plans: I’m continuing studies