This study validates the use of residence time distribution (RTD) functions in human subjects to assess changes in retinal flow by using the widely recognized model of flow changes due to oxygen breathing. Changes in retinal blood flow may provide important information for clinical decision-making in several populations, including those with diabetic retinopathy, sickle cell disease and retinitis pigmentosa. Normal volunteer subjects were studied before and after oxygen breathing. After i.v. injection, relative fluorescence was obtained using scanning laser ophthalmoscopy/image processing in all vessel branches (average, 17). For each experiment, 64 frames (2/s) were digitized and were normalized using the RTD method. Vessel diameters were measured using densitometry techniques on fundus photos, where the diameter data made it possible to weight each vessel according to relative cross-sectional area to obtain a true mean circulation time (MCT). MCT increased for the group of subjects when breathing oxygen compared to normal air (P = 0.001), representing a decrease in retinal blood flow. Average MCT increased 2.82 +/- 2.51 s for all subjects, with an increase of 2.93 +/- 2.26 s in repeat trials for one subject. The proposed method uses information from all retinal vessels and allows the assessment of overall, as well as selected, regional retinal flow. It is more comprehensive than previous methods analysing single vessel flow. This method will be potentially useful for assessing hemodynamic changes in the retina associated with a wide range of eye disease.