Robustness and fidelity of voltage imaging analysis pipelines

Tracking neuronal voltage changes using fluorescent voltage indicators is rapidly reshaping neuroscientific research. Voltage imaging enables direct visualization of electrical signals from subcellular compartments to large-scale networks, yet requires sophisticated image-analysis procedures. Here, we present a comprehensive study of current voltage imaging analysis pipelines and discuss the experimental conditions for which they are most suitable. We compare strengths and limitations of these pipelines in motion correction, denoising and segmentation routines, and discuss how different signal processing strategies can influence data integrity and interpretation. Our results show that most real-time analyses require GPU-accelerated algorithms and that denoising prior to signal processing is needed for analysis of subcellular dynamics. We conclude that voltage imaging analyses needs to be tailored to different experimental settings: we propose a decision-tree to determine analysis strategies for diverse experimental conditions. Together, these insights pave the way for reproducible, high-fidelity voltage imaging studies.