Deep learning-assisted mapping of dendritic spines using sequential 2D two-photon calcium imaging

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cell Biology

Background:

• Neurons process information through complex synaptic inputs.
• Synapse location critically influences dendritic events and neuronal computation.
• Mapping synaptic organization is essential for understanding neuronal function.

Purpose of the Study:

• To present a novel software pipeline for targeted dendritic imaging and analysis.
• To facilitate the mapping of synaptic organization through spine calcium imaging.
• To enable detailed study of neuronal input-output transformations.

Main Methods:

• Developed a software pipeline using sequential 2D scanning on two-photon microscopes.
• Integrated ROIpy for dendrite-aligned region-of-interest generation for depth-specific imaging.
• Utilized Spyne with deep learning for spine detection and calcium activity classification.

Main Results:

• The pipeline enables targeted dendritic imaging across intricate 3D dendrites.
• Deep learning modules accurately detect spines and classify calcium activity.
• The method supports diverse experimental setups, including patch-clamp and all-optical configurations.

Conclusions:

• The presented software pipeline offers a robust method for analyzing synaptic organization.
• This approach enhances the understanding of how synapse location shapes neuronal computation.
• The tool supports various experimental designs for comprehensive neuronal analysis.