Diabetic retinopathy has long been considered to be a retinal manifestation of systemic diabetic angiopathy. Indeed, it is therapeutically true. However, the prolongation of OP peak latency in diabetic eyes without any angiographic evidence of angiopathy leads us to presume that certain neuronal disorders occur early in diabetic eyes. Even though we cannot neglect the possibility that the prolongation of the OP peak latency may derive from undetectable retinal hypoperfusion, it is still far from conventional diabetic angiopathy. Rather, the status should be properly termed "intraretinal diabetic neuropathy" in that the neurones are the disturbed cells to cause visual dysfunction. Thereafter, the OP amplitude diminishes as retinopathy advances, probably depending on the degree of retinal circulatory disturbance. Marked diminution of the OP amplitude predicts rapid progression and poor prognosis of retinopathy. Diabetic retinal pigment epitheliopathy as manifested by one of our non-photic EOG responses is another kind of early ocular involvement of diabetes. Because its mechanisms are not yet known, so far we have not succeeded in correlating it to any kind of subjective visual index. Routine fundus inspection or fluorescent fundus angiography is incapable of detecting the compromised neural retina and/or retinal pigment epithelial integrity and thus the electrophysiology of vision has the edge in ophthalmology.