Bayesian In Vivo Tracking of Synapses using Joint Poisson Deconvolution and Diffeomorphic Registration

2026-05-13 · Source: Takara TLDR - Daily AI Papers · Field: Science & Research — Life Sciences & Biology, Mathematics & Computational Sciences, Health & Medical Research · Depth: Expert, medium

Summary

A novel template-based Bayesian framework has been developed for tracking synapses in vivo, addressing challenges in longitudinal 2-photon microscopy data. This method models synapses as varying luminance point sources undergoing nonlinear tissue deformation. It applies a unified Bayesian approach, deriving a posterior that integrates a diffeomorphic mapping for domain warping, a Gaussian point spread function for imaging, and a Poisson observation model for raw photon counts. The framework simultaneously constructs a probabilistic template of synapse locations, denoises and deconvolves image data, infers fluorescence intensities, performs diffeomorphic image registration to correct tissue motion, and provides confidence regions for parameter estimates. The framework's efficacy was demonstrated on both a 2D+t simulated dataset and a 3D+t longitudinal in vivo microscopy dataset of fluorescent synapses imaged in a mouse over two weeks.

Key takeaway

For research scientists studying synaptic dynamics in vivo, this Bayesian framework offers a robust solution to overcome challenges posed by low SNR and tissue motion in 2-photon microscopy. You can achieve more accurate detection and tracking of synapses, even in dense regions, by leveraging its integrated denoising, deconvolution, and registration capabilities. Consider implementing this approach to enhance the reliability of your longitudinal synaptic imaging studies.

Key insights

A Bayesian framework enables robust, simultaneous synapse tracking, denoising, and registration in noisy in vivo microscopy data.

Principles

• Model synapses as varying luminance point sources.
• Integrate diffeomorphic mapping for tissue motion correction.
• Utilize Poisson observation model for photon counts.

Method

The method derives a Bayesian posterior incorporating diffeomorphic mapping, a Gaussian PSF, and a Poisson observation model to simultaneously track synapses, denoise images, infer intensities, and register data.

In practice

• Apply to 2-photon microscopy for synapse tracking.
• Use for longitudinal studies of synaptic dynamics.
• Quantify synaptic changes in neurological disease models.

Topics

• Synapse Tracking
• In Vivo Microscopy
• Bayesian Framework
• Diffeomorphic Registration
• Poisson Deconvolution

Code references

• ssc1230609-spec/MSR-registration

Best for: AI Scientist, Research Scientist

Related on AIssential

• A physics-informed foundation model for quantitative diffusion MRI — PIGMENT uses a physics-informed generative prior to enable reliable quantitative diffusion MRI from sparse, heterogeneous data.
• Co-Diffusion: An Affinity-Aware Two-Stage Latent Diffusion Framework for Generalizable Drug-Target Affinity Prediction — Co-Diffusion enhances drug-target affinity prediction in cold-start scenarios using a two-stage, affinity-aware latent diffusion framework.
• Anatomy-Guided Residual Motion Diffusion for Controllable 4D Cardiac MRI Synthesis — A cascaded diffusion model disentangles anatomy and motion for controllable 4D medical image synthesis.
• Treatment-Conditioned Diffusion for Forecasting Neurodegenerative Disease Progression — A treatment-conditioned diffusion framework accurately forecasts neurodegenerative disease progression by integrating imaging and pharmacological data, preserving anatomical detail.
• Estimation of instrument and noise parameters for inverse problem based on prior diffusion model — A Gibbs-Diffusion Posterior Sampling (G-DPS) based method efficiently estimates observation parameters and images in inverse problems with diffusion priors.
• $${\bf{Micro}}{{\mathbb{S}}}{\bf{plit}}$$ Micro S plit : semantic unmixing of fluorescent microscopy data — MicroSplit enables simultaneous multi-structure imaging in one channel, providing denoised, unmixed outputs with uncertainty quantification.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Takara TLDR - Daily AI Papers.