Deep and shallow data science for multi-scale optical neuroscience

TL;DR

The paper addresses the challenge of analyzing multi-scale optical neuroscience data, where SNR and resolution limitations necessitate both specialized and generalizable computational methods. It evaluates the promise and limits of deep learning in denoising, ROI extraction, and image restoration, and foregrounds the out-of-distribution (OOD) problem that hinders cross-lab generalization and dissemination. A central argument is that no single pipeline fits all contexts, requiring robust validation across diverse data and adaptive retraining capabilities to remain effective in new settings. The authors advocate for community-driven solutions and data infrastructures that support data diversity, model adaptation, and measurement standardization to enable reproducible, scalable analyses across labs.

Abstract

Optical imaging of the brain has expanded dramatically in the past two decades. New optics, indicators, and experimental paradigms are now enabling in-vivo imaging from the synaptic to the cortex-wide scales. To match the resulting flood of data across scales, computational methods are continuously being developed to meet the need of extracting biologically relevant information. In this pursuit, challenges arise in some domains (e.g., SNR and resolution limits in micron-scale data) that require specialized algorithms. These algorithms can, for example, make use of state-of-the-art machine learning to maximally learn the details of a given scale to optimize the processing pipeline. In contrast, other methods, however, such as graph signal processing, seek to abstract away from some of the details that are scale-specific to provide solutions to specific sub-problems common across scales of neuroimaging. Here we discuss limitations and tradeoffs in algorithmic design with the goal of identifying how data quality and variability can hamper algorithm use and dissemination.