Abstract
Fluorescence lifetime imaging microscopy (FLIM) is a sensitive technique that provides insight into molecular interactions, making it a powerful tool for understanding protein dynamics in neurodegenerative diseases. However, longitudinal FLIM studies over a number of days aimed at capturing complex protein behavior in live samples with high spatial granularity are limited by the need for rapid acquisition, implying few signal photons and high noise. Microscopy studies are further challenged by the complex lifetime mixtures present in many samples, leading to a more challenging extraction problem and the prospect of obscuring insight into the underlying molecular interactions. To enable accurate recovery in such situations, we present a noise estimation method that allows precise, pixel-wise determination of fluorescence lifetime parameters in such high-noise environments. In addition, we introduce a multi-exponential constrained fitting approach that enables robust multiparameter extraction from measured data with noisy data typical of neuron studies. The approach is validated using reference dyes and is further illustrated for neuronal imaging of the aggregation of alpha-synuclein, a key protein related to the onset and progression of Parkinson's disease. More broadly, reliable studies of protein dynamics become possible, thereby providing a means to advance our understanding of neurodegenerative disease etiology and offering applications across diverse molecular systems and in various disciplines.