The main objective of the study was to assess effects of long-term lowering of glucosylated hemoglobin (HbA1%) on neurosensory function in insulin-dependent diabetes. Individual (OP-1, OP-2, OP-3) and summed (OP-sum) amplitudes of oscillatory potentials (OPs) of electroretinography were recorded at study start and 7-years later in 45 patients (the Oslo study). As an overall 7-year change, amplitudes of OP-2, OP-3 and OP-sum were reduced (p < 0.0001-0.01), retinopathy worsened (p = 0.005), intraocular pressure decreased (p < 0.001), systolic blood pressure increased (p < 0.0002), and glycemic control improved from HbA1 of 11.2 +/- 2.2% at study start to a 7-year cumulative mean of 9.5 +/- 1.5% (p < 0.0001). Multiple regression analysis did not identify any independent relations between change in OP-1, OP-2, OP-3, OP-sum and change in glycemic control or background variables, including change in age and duration of diabetes. However, cross-sectional observations at 7 years showed negative correlations between all OPs and age (p < 0.0001-0.003), and between OP-3 and duration (p = 0.003) and counts of microaneurysms (p = 0.02). The data suggest that various clinical background variables may influence individual and summed amplitudes of OPs differently. Reduced neurosensory retinal function (OPs) seemed to appear after 7-years, independently of vascular defects of retinopathy and long-term improvement in glucose control.