Fast and accessible morphology-free functional fluorescence imaging analysis

Optical calcium imaging is a powerful tool for recording neural activity across a wide range of spatial scales, from dendrites and spines to whole-brain imaging through two-photon and widefield microscopy. Traditional methods for analyzing functional calcium imaging data rely heavily on spatial features, such as the compact shapes of somas, to extract regions of interest and their associated temporal traces. This spatial dependency can introduce biases in time trace estimation and limit the applicability of these methods across different neuronal morphologies and imaging scales. To address these limitations, the Graph Filtered Temporal Dictionary Learning (GraFT) uses a graph-based approach to identify neural components based on shared temporal activity rather than spatial proximity, enhancing generalizability across diverse datasets. Here we present significant advancements to the GraFT algorithm, including the integration of a more efficient solver for the L1 least absolute shrinkage and selection operator (LASSO) problem and the application of compressive sensing techniques to reduce computational complexity. By employing random projections to reduce data dimensionality, we achieve substantial speedups while maintaining analytical accuracy. These advancements significantly accelerate the GraFT algorithm, making it more scalable for larger and more complex datasets. Moreover, to increase accessibility, we developed a graphical user interface to facilitate running and analyzing the outputs of GraFT. Finally, we demonstrate the utility of GraFT to imaging data beyond meso-scale imaging, including vascular and axonal imaging.Author summary: Calcium imaging enables the recording of large populations of cells in real-time within living organisms. However, current methods for analyzing these recordings often rely on the shapes of cells in the images, which can bias results and make it difficult to apply the same techniques across different types of data and scales. In our work, we introduce a significantly improved version of our algorithm, Graph Filtered Temporal Dictionary Learning (GraFT), that instead focuses on patterns of activity over time to identify groups of neurons. We made several key advances: we sped up the computations with new optimization tools, applied data compression techniques to improve efficiency, and built a user-friendly interface so that others can easily apply GraFT to their own work. We further showcase GraFT’s strength in analyzing different types of imaging data and scales, from small blood vessels to long-range axons. By making analysis faster, more flexible, and more accessible, our contributions position GraFT as a robust and versatile approach to analyzing increasingly large and heterogeneous imaging datasets that support new directions in systems neuroscience.