Related Experiment Video
Updated: Aug 29, 2025
Monitoring Neuronal Survival via Longitudinal Fluorescence Microscopy
Published on: January 19, 2019
Fluorescently labeled nuclear morphology is highly informative of neurotoxicity
Shijie Wang1, Jeremy W Linsley1, Drew A Linsley2,3
1Center for Systems and Therapeutics, Gladstone Institutes, San Francisco, CA, United States.
Abstract:
Neurotoxicity can be detected in live microscopy by morphological changes such as retraction of neurites, fragmentation, blebbing of the neuronal soma and ultimately the disappearance of fluorescently labeled neurons. However, quantification of these features is often difficult, low-throughput, and imprecise due to the overreliance on human curation. Recently, we showed that convolutional neural network (CNN) models can outperform human curators in the assessment of neuronal death from images of fluorescently labeled neurons, suggesting that there is information within the images that indicates toxicity but that is not apparent to the human eye. In particular, the CNN's decision strategy indicated that information within the nuclear region was essential for its superhuman performance. Here, we systematically tested this prediction by comparing images of fluorescent neuronal morphology from nuclear-localized fluorescent protein to those from freely diffused fluorescent protein for classifying neuronal death. We found that biomarker-optimized (BO-) CNNs could learn to classify neuronal death from fluorescent protein-localized nuclear morphology (mApple-NLS-CNN) alone, with super-human accuracy. Furthermore, leveraging methods from explainable artificial intelligence, we identified novel features within the nuclear-localized fluorescent protein signal that were indicative of neuronal death. Our findings suggest that the use of a nuclear morphology marker in live imaging combined with computational models such mApple-NLS-CNN can provide an optimal readout of neuronal death, a common result of neurotoxicity.
Insights
Detecting neurotoxicity in live microscopy is challenging. New convolutional neural network (CNN) models accurately identify neuronal death using nuclear morphology, outperforming human analysis for improved neurotoxicity assessment.
Area of Science:
- Neuroscience
- Computational Biology
- Toxicology
Background:
- Neurotoxicity detection via live microscopy relies on morphological changes, often difficult to quantify accurately.
- Human curation for assessing neuronal death is imprecise and low-throughput.
- Convolutional neural networks (CNNs) show potential for superior neurotoxicity assessment.
Purpose of the Study:
- To investigate the utility of nuclear morphology in detecting neurotoxicity using CNNs.
- To compare the efficacy of nuclear-localized fluorescent protein versus freely diffused fluorescent protein for classifying neuronal death.
- To identify novel features indicative of neuronal death through explainable AI.
Main Methods:
- Systematic comparison of fluorescent neuronal morphology from nuclear-localized and freely diffused fluorescent proteins.
- Development and application of biomarker-optimized (BO-) CNNs, specifically mApple-NLS-CNN.
- Utilized explainable artificial intelligence (XAI) methods to analyze CNN decision strategies.
Main Results:
- Biomarker-optimized CNNs achieved superhuman accuracy in classifying neuronal death using only nuclear-localized fluorescent protein.
- Nuclear morphology alone was sufficient for accurate classification of neuronal death.
- Novel features within the nuclear signal predictive of neuronal death were identified.
Conclusions:
- Nuclear morphology markers combined with computational models like mApple-NLS-CNN offer an optimal readout for live imaging of neuronal death.
- This approach enhances the accuracy and efficiency of neurotoxicity assessment.
- CNNs can identify subtle indicators of neurotoxicity not apparent to human observers.
More Related Videos
07:41A Neurite Outgrowth Assay and Neurotoxicity Assessment with Human Neural Progenitor Cell-Derived Neurons
Published on: August 6, 2020
08:53Assessment of Neuronal Viability Using Fluorescein Diacetate-Propidium Iodide Double Staining in Cerebellar Granule Neuron Culture
Published on: May 10, 2017