Related Experiment Video
Updated: Oct 1, 2025
08:26
Simultaneous Imaging of Microglial Dynamics and Neuronal Activity in Awake Mice
Published on: August 23, 2022
2.5K
Light-Field Microscopy for Optical Imaging of Neuronal Activity: When Model-Based Methods Meet Data-Driven
Pingfan Song1, Herman Verinaz Jadan2, Carmel L Howe3
1Department of Engineering, University of Cambridge, Cambridge, CB2 1PZ, UK.
IEEE Signal Processing Magazine
|March 9, 2022
Summary
This paper surveys computational methods for light-field microscopy (LFM) to image neuronal activity. It highlights integrating physics-based and data-driven machine learning for advanced neuroscience research.
Area of Science:
- Neuroscience
- Microscopy
- Computational Biology
Background:
- Understanding neural network information processing requires observing large-scale neuronal dynamics.
- Light-field microscopy (LFM) enables high-speed 3D imaging of neuronal populations at video frame-rates.
Purpose of the Study:
- To provide a comprehensive survey of state-of-the-art computational methods for LFM.
- To focus on integrating model-based and data-driven approaches for analyzing LFM data.
Main Methods:
- Reviewing existing computational techniques for LFM data analysis.
- Exploring the application of signal processing and wave-optics theories.
- Investigating the synergy between model-driven and data-driven machine learning.
Main Results:
- Identified the need for novel machine learning approaches tailored to LFM.
- Highlighted the potential of physics and optics models to enhance LFM data analysis.
- Emphasized the benefits of integrating diverse theoretical frameworks.
Conclusions:
- Computational methods are crucial for advancing LFM applications in neuroscience.
- Hybrid approaches combining model-based and data-driven techniques offer enhanced interpretability and generalization.
- Further research in this area will accelerate discoveries in neural dynamics.
