AbstractPurpose To assess visual impairment in diabetic patients with and without established blood‐retinal barrier leakage and relation with the status of the blood‐brain barrier.Methods We assessed 37 diabetic patients (29 type2; mean Hb1Ac 7.93[1.44]) using four different approaches: a two‐alternative forced‐choice multi‐psychophysical channel test for speed discrimination and achromatic contrast as well as chromatic L, M and S cone contrast discrimination; multifocal electrorretinography(mfERG) and brain barrier integrity assessment using a permeability measure Ktrans for dynamic contrast enhanced MRI (Cramer et. al, 2013) in 4 regions of interest(ROIs). We selected two diabetic groups with and without measured retinal barrier leakage. Spearman’s rank were used for correlations and Mann‐Whitney U test for comparison between groups for a set p value threshold of 0.05.Results We found that presence of retinal leakage was not associated with changes in blood brain barrier permeability at any of the studied ROIs. Interestingly we found, in particular in diabetic type 2 patients, significant negative correlations between the amplitude and/or implicit time of N1 and P1 components of the multifocal electrorretinogram and brain leakage, for the most external stimulation rings. Psychophysical performance was impaired in particular for S but was not correlated with barrier status.Conclusion Our findings suggest that retina and brain blood barriers are patho physiologically unrelated in diabetes. We found a direct correlation between neurophysiological measures and brain barrier damage, suggesting that both may have a direct common cause.