Generative Adversarial Network Model to Classify Human Induced Pluripotent Stem Cell-Cardiomyocytes based on Maturation Level.

Objective: Our study develops a generative adversarial network (GAN)-based method that generates faithful synthetic image data of human cardiomyocytes at varying stages in their maturation process, as a tool to significantly enhance the classification accuracy of cells and ultimately assist the throughput of computational analysis of cellular structure and functions. Methods: Human induced pluripotent stem cell derived cardiomyocytes (hiPSC-CMs) were cultured on micropatterned collagen coated hydrogels of physiological stiffnesses to facilitate maturation and optical measurements were performed for their structural and functional analyses. Control groups were cultured on collagen coated glass well plates. These image recordings were used as the real data to train the GAN model. Results: The results show the GAN approach is able to replicate true features from the real data, and inclusion of such synthetic data significantly improves the classification accuracy compared to usage of only real experimental data that is often limited in scale and diversity. Conclusion: The proposed model outperformed four conventional machine learning algorithms with respect to improved data generalization ability and data classification accuracy by incorporating synthetic data. Significance: This work demonstrates the importance of integrating synthetic data in situations where there are limited sample sizes and thus, effectively addresses the challenges imposed by data availability.

Heart-on-Chip for Combined Cellular Dynamics Measurements and Computational Modeling Towards Clinical Applications.

Organ-on-chip or micro-engineered three-dimensional cellular or tissue models are increasingly implemented in the study of cardiovascular pathophysiology as alternatives to traditional in vitro cell culture. Drug induced cardiotoxicity is a key issue in drug development pipelines, but the current in vitro and in vivo studies suffer from inter-species differences, high costs, and lack of reliability and accuracy in predicting cardiotoxicity. Microfluidic heart-on-chip devices can impose a paradigm shift to the current tools. They can not only recapitulate cardiac tissue level functionality and the communication between cells and extracellular matrices but also allow higher throughput studies conducive to drug screening especially with their added functionalities or sensors that extract disease-specific phenotypic, genotypic, and electrophysiological information in real-time. Such electrical and mechanical components can tailor the electrophysiology and mechanobiology of the experiment to better mimic the in vivo condition as well. Recent advancements and challenges are reviewed in the fabrication, functionalization and sensor assisted mechanical and electrophysiological measurements, numerical and computational modeling of cardiomyocytes' behavior, and the clinical applications in drug screening and disease modeling. This review concludes with the current challenges and perspectives on the future of such organ-on-chip platforms.

Noninvasive detection of the hemodynamic stress of exercise using the photoplethysmogram.

Exercise induced hemodynamic stress has been studied extensively using a wide range of physiological sensors. While athletes can modulate their training intensity using EKG-based heart rate monitors, there are currently no noninvasive monitors that can be used to ascertain with a high degree of certainty the hemodynamic stress an individual is experiencing because of fatigue or an underlying pathology. We propose that cardiac stress will result in detectable changes in skin blood flow. In a clinical trial with eleven healthy subjects performing the Bruce Protocol treadmill test low frequency waves were observed in the blood flow to both the forehead and ear, but not the finger, using photople- thysmographs (PPG) measured by a pulse oximeter. As volitional fatigue approached, the low frequency (f = 0.05-0.2 Hz) amplitude modulation observed in the PPG became more pronounced; then, within several seconds of the cessa- tion of the protocol, they disappeared. Using a software-based detector, these distinct waves are reliably detected, with a low incidence of false positives, in all subjects before the onset of volitional fatigue. We hypothesize that the low frequency waves observed in the PPG of individuals exercising to volitional fatigue provide a mechanism for noninvasively detecting hemodynamic stress to the human vascular system.