Deep learning-based framework for cardiac function assessment in embryonic zebrafish from heart beating videos

Naderi, Amir Mohammad; Bu, Haisong; Su, Jingcheng; Huang, Mao Hsiang; Vo, Khuong; Torres, Ramses Seferino Trigo; Chiao, Jung-Chih; Lee, Juhyun; Lau, Michael; Xu, Xiaolei; Cao, Hung

doi:10.1016/j.compbiomed.2021.104565

Cited by 15 publications

(15 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Additionally, the Intersection of Union (IoU) score of deep-learning network (ZACAF) reached 0.80 ( Supplementary Figure 4B ). The model can be qualitatively compared to show that the model prediction is similar to the hand segmented images ( Naderi et al, 2021 ).…”

Section: Resultsmentioning

confidence: 99%

“…Since the frames are 2D images, we have to estimate the ventricle volume to its area. For calculating FS, measurements of the short axis in ES and ED frames are taken ( Naderi et al, 2021 ).…”

Section: Methodsmentioning

confidence: 99%

“…The heart was visualized and viewed on a microscope at 4× magnification. Videos were taken in a bright field of the heart for analysis of ejection fraction (EF), fractional shortening (FS), end-systolic (ES) and end-diastolic (ED) volume by the deep-learning framework Zebrafish Automatic Cardiovascular Assessment Framework (ZACAF), which was recently developed and validated (Naderi et al, 2021). For segmentation, briefly, ZACAF utilizes a deep learning model with a U-net architecture, which was designed for sematic segmentation in biomedical imaging.…”

Section: Reverse Transcriptase Quantitative Pcrmentioning

confidence: 99%

See 2 more Smart Citations

In vivo Evaluation of Non-viral NICD Plasmid-Loaded PLGA Nanoparticles in Developing Zebrafish to Improve Cardiac Functions

et al. 2022

Self Cite

View full text Add to dashboard Cite

In the era of the advanced nanomaterials, use of nanoparticles has been highlighted in biomedical research. However, the demonstration of DNA plasmid delivery with nanoparticles for in vivo gene delivery experiments must be carefully tested due to many possible issues, including toxicity. The purpose of the current study was to deliver a Notch Intracellular Domain (NICD)-encoded plasmid via poly(lactic-co-glycolic acid) (PLGA) nanoparticles and to investigate the toxic environmental side effects for an in vivo experiment. In addition, we demonstrated the target delivery to the endothelium, including the endocardial layer, which is challenging to manipulate gene expression for cardiac functions due to the beating heart and rapid blood pumping. For this study, we used a zebrafish animal model and exposed it to nanoparticles at varying concentrations to observe for specific malformations over time for toxic effects of PLGA nanoparticles as a delivery vehicle. Our nanoparticles caused significantly less malformations than the positive control, ZnO nanoparticles. Additionally, the NICD plasmid was successfully delivered by PLGA nanoparticles and significantly increased Notch signaling related genes. Furthermore, our image based deep-learning analysis approach evaluated that the antibody conjugated nanoparticles were successfully bound to the endocardium to overexpress Notch related genes and improve cardiac function such as ejection fraction, fractional shortening, and cardiac output. This research demonstrates that PLGA nanoparticle-mediated target delivery to upregulate Notch related genes which can be a potential therapeutic approach with minimum toxic effects.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Reverse Transcriptase Quantitative Pcrmentioning

confidence: 99%

See 1 more Smart Citation

In vivo Evaluation of Non-viral NICD Plasmid-Loaded PLGA Nanoparticles in Developing Zebrafish to Improve Cardiac Functions

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…To obtain the cardiac parameters from the TSA and KYM methods, it required several steps, including manual counting [ 5 ]. In addition, most of the tools in the ImageJ software is a freehand ruler, which might cause inaccuracy, especially if the heart chambers are small [ 15 ]. Since the regions of interest (ROI) in the TSA method are selected manually by using a circle tool with a limited size selection [ 31 ], the results might be affected depending on the users.…”

Section: Discussionmentioning

confidence: 99%

Using DeepLabCut as a Real-Time and Markerless Tool for Cardiac Physiology Assessment in Zebrafish

Suryanto

Saputra

Kurnia

et al. 2022

Biology

View full text Add to dashboard Cite

DeepLabCut (DLC) is a deep learning-based tool initially invented for markerless pose estimation in mammals. In this study, we explored the possibility of adopting this tool for conducting markerless cardiac physiology assessment in an important aquatic toxicology model of zebrafish (Danio rerio). Initially, high-definition videography was applied to capture heartbeat information at a frame rate of 30 frames per second (fps). Next, 20 videos from different individuals were used to perform convolutional neural network training by labeling the heart chamber (ventricle) with eight landmarks. Using Residual Network (ResNet) 152, a neural network with 152 convolutional neural network layers with 500,000 iterations, we successfully obtained a trained model that can track the heart chamber in a real-time manner. Later, we validated DLC performance with the previously published ImageJ Time Series Analysis (TSA) and Kymograph (KYM) methods. We also evaluated DLC performance by challenging experimental animals with ethanol and ponatinib to induce cardiac abnormality and heartbeat irregularity. The results showed that DLC is more accurate than the TSA method in several parameters tested. The DLC-trained model also detected the ventricle of zebrafish embryos even in the occurrence of heart abnormalities, such as pericardial edema. We believe that this tool is beneficial for research studies, especially for cardiac physiology assessment in zebrafish embryos.

show abstract

“…Since the ROI selection in each video was different and should be manually selected, the current version does not support batch conduction. The future study by machine learning to conduct automated ROI selection might overcome this limitation [81]. Furthermore, in the future, the addition of Graphical User Interface (GUI) could benefit a lot of people who use this tool, as the addition of GUI will improve the usability of the tool, especially for researchers without a computer science background.…”

Section: Potential Limitations and Future Workmentioning

confidence: 99%

An OpenCV-Based Approach for Automated Cardiac Rhythm Measurement in Zebrafish from Video Datasets

et al. 2021

View full text Add to dashboard Cite

Cardiac arrhythmia has been defined as one of the abnormal heart rhythm symptoms, which is a common problem dealt with by cardiologists. Zebrafish were established as a powerful animal model with a transparent body that enables optical observation to analyze cardiac morphology and cardiac rhythm regularity. Currently, research has observed heart-related parameters in zebrafish, which used different approaches, such as starting from the use of fluorescent transgenic zebrafish, different software, and different observation methods. In this study, we developed an innovative approach by using the OpenCV library to measure zebrafish larvae heart rate and rhythm. The program is designed in Python, with the feature of multiprocessing for simultaneous region-of-interest (ROI) detection, covering both the atrium and ventricle regions in the video, and was designed to be simple and user-friendly, having utility even for users who are unfamiliar with Python. Results were validated with our previously published method using ImageJ, which observes pixel changes. In summary, the results showed good consistency in heart rate-related parameters. In addition, the established method in this study also can be widely applied to other invertebrates (like Daphnia) for cardiac rhythm measurement.

show abstract

Deep learning-based framework for cardiac function assessment in embryonic zebrafish from heart beating videos

Cited by 15 publications

References 23 publications

In vivo Evaluation of Non-viral NICD Plasmid-Loaded PLGA Nanoparticles in Developing Zebrafish to Improve Cardiac Functions

In vivo Evaluation of Non-viral NICD Plasmid-Loaded PLGA Nanoparticles in Developing Zebrafish to Improve Cardiac Functions

Using DeepLabCut as a Real-Time and Markerless Tool for Cardiac Physiology Assessment in Zebrafish

An OpenCV-Based Approach for Automated Cardiac Rhythm Measurement in Zebrafish from Video Datasets

Contact Info

Product

Resources

About