RESPAN is an open-source deep-learning pipeline for restoration, segmentation, and quantitative analysis of neuronal morphology. Developed in collaboration with the Polleux Lab at Columbia University, RESPAN enables automated analysis of dendritic spines, dendritic branches, and soma across challenging volumetric microscopy datasets.
The pipeline combines content-aware image restoration, automated segmentation, model training, analysis, and validation within a single user-friendly graphical interface. By improving signal, contrast, and isotropic resolution before segmentation, RESPAN supports robust quantification across fixed tissue, live imaging, and in vivo two-photon datasets.
The published study validates RESPAN against expert annotations and benchmarks it against existing software, demonstrating accurate and reproducible analysis across diverse imaging modalities. RESPAN is freely available on GitHub and can run on standard workstations or laptops equipped with an NVIDIA GPU.
A deep learning pipeline for accurate and automated restoration, segmentation, and quantification of dendritic spines. Bernal-Garcia S, Schlotter AP, Pereira DB, Recupero AJ, Polleux F, Hammond LA*. Cell Reports Methods. 2025;5(10):101179. *corresponding author.