Statistical analysis and optimization of frequency-domain fluorescence lifetime imaging microscopy using homodyne lock-in detection.

Fluorescence lifetime imaging microscopy is used widely in biological research, but the accuracy and precision of lifetime measurements are limited. Photon noise is an inherent error source that cannot be eliminated. In this paper, we present a general approach to compute the probability density of the estimated lifetime for frequency-domain fluorescence lifetime imaging microscopy using homodyne lock-in detection. The analysis for commonly used excitation methods, including sinusoidal modulation, square-wave modulation, and a periodically pulsed light source, are given and compared to the results of Monte Carlo simulations. The optimum parameters of the excitation waveforms to minimize the variance of the estimated lifetimes are also derived and compared to previously published results.