Review on data analysis methods for mesoscale neural imaging in vivo

Area of Science:

• Neuroscience
• Biomedical Imaging
• Computational Biology

Background:

• Mesoscale neural imaging in vivo is crucial for understanding neuron-behavior relationships, but signal processing is complex and unstandardized.
• Optical imaging faces challenges like noise, artifacts, and demanding computational requirements, hindering accessibility for non-specialists.

Purpose of the Study:

• To develop a general, standardized data analysis pipeline for mesoscale neural imaging.
• To provide a shared framework applicable across different imaging modalities and systems.
• To simplify complex neural data processing for broader neuroscience research.

Main Methods:

• The pipeline comprises two stages: single-cell neural response extraction and data mining.
• Stage 1 includes motion registration, denoising, neuron segmentation, and signal extraction.
• Stage 2 involves neural functional mapping, clustering, and brain-wide network deduction.

Main Results:

• A general pipeline for processing mesoscale neural images is presented.
• Principles, comparisons of approaches, and application scopes are discussed.
• Shortcomings and remaining challenges in mesoscale neural data analysis are highlighted.

Conclusions:

• Large-scale mesoscale data presents challenges in balancing fidelity and efficiency, computational load, and interpretability.
• Future research will likely focus on exploring global circuits at the single-neuron level.
• Standardized pipelines are essential for advancing the field of mesoscale neural imaging analysis.