Optimizing the resolution-fidelity trade-off in SRRF nanoscopy for live-cell clathrin imaging

Total internal reflection fluorescence microscopy (TIRFM) enables live-cell imaging of clathrin-coated pits (CCPs) but is diffraction-limited (∼250 nm), hindering visualization of their dense nanostructures (∼150 nm). Super-Resolution Radial Fluctuations (SRRF) provides computational super-resolution; however, unoptimized parameters can disrupt the critical balance between resolution and structural fidelity. Here, we establish a comprehensive parameter optimization framework for SRRF-based nanoscopy of live CCPs. Using multi-modal metrics in CLTA-GFP HeLa cells, we identify a ring radius of 1.0 as optimal, balancing resolution (full width at half maximum = 180 ± 29 nm, p < 0.05 vs. TIRFM) with fidelity (resolution-scaled Pearson = 0.935 ± 0.018). The temporal radiality pairwise product mean (TRPPM) mode achieved superior resolution (154 ± 30 nm) while maintaining fidelity metrics comparable to Temporal radiality average (TRA) mode. In contrast, temporal radiality auto-correlations (TRAC) introduced artifactual structures and reduced fidelity. Parameters such as "remove positivity constraint" and gradient weighting induced severe artifacts and should be avoided. This optimized framework resolves the resolution-fidelity trade-off, enabling robust nanoscale imaging of clathrin-mediated endocytosis in live cells.