PURPOSE. Diabetes appears to induce a visual cycle defect because rod dysfunction is correctable with systemic treatment of the visual cycle chromophore 11-cis-retinaldehyde. However, later studies have found no evidence for visual cycle impairment. Here, we further examined whether photoreceptor dysfunction is corrected with 11-cis-retinaldehyde. Because antioxidants correct photoreceptor dysfunction in diabetes, the hypothesis that exogenous visual chromophores have antioxidant activity in the retina of diabetic mice in vivo was tested.METHODS. Rod function in 2-month-old diabetic mice was evaluated using transretinal electrophysiology in excised retinas and apparent diffusion coefficient (ADC) MRI to measure light-evoked expansion of subretinal space (SRS) in vivo. Optokinetic tracking was used to evaluate cone-based visual performance. Retinal production of superoxide free radicals, generated mostly in rod cells, was biochemically measured with lucigenin. Diabetic mice were systemically treated with a single injection of either 11-cis-retinaldehyde, 9-cisretinaldehyde (a chromophore surrogate), or all-trans-retinaldehyde (the photoisomerization product of 11-cis-retinaldehyde).RESULTS. Consistent with previous reports, diabetes significantly reduced (1) dark-adapted rod photo responses (transretinal recording) by~18%, (2) rod-dominated light-stimulated SRS expansion (ADC MRI) by~21%, and (3) cone-dominated contrast sensitivity (using optokinetic tracking [OKT]) by~30%. Both 11-cis-retinaldehyde and 9-cis-retinaldehyde largely corrected these metrics of photoreceptor dysfunction. Higher-than-normal retinal superoxide production in diabetes by~55% was also significantly corrected following treatment with 11-cis-retinaldehyde, 9-cis-retinaldehyde, or all-trans-retinaldehyde.CONCLUSIONS. Collectively, data suggest that retinaldehydes improve photoreceptor dysfunction in diabetic mice, independent of the visual cycle, via an antioxidant mechanism.