1. Intravenous lactate prevents cerebral dysfunction during hypoglycaemia in healthy volunteers. This study examines whether this also occurs in insulin-dependent diabetes. Changes in four-choice reaction time, auditory brain stem response, and P300 latency were used as measures of cerebral function. 2. Ten subjects were studied twice at least 4 weeks apart. Blood glucose was maintained between 5 and 8 mmol/l for 1 h before starting a 60 m-unit min-1 m-2 stepped hyperinsulinaemic clamp, achieving blood glucose concentrations of 4.5, 3.3 and 2.5 mmol/l. At one visit, 40 mumol min-1 kg-1 sodium lactate was infused, and at the other, normal saline. Cerebral function was measured at each blood glucose concentration. 3. Blood lactate rose to 3.32 +/- 0.06 mmol/l during lactate infusion compared with 0.9 +/- 0.03 mmol/l during saline infusion. Compared with the results at 4.5 mmol/l there were no significant changes at 3.3 mmol/l in any measure of cerebral function at either visit. At 2.5 mmol/l a significant increase in reaction time and P300 latency occurred with saline [mean change 33.1 +/- 8.6 ms (P < 0.01) and 30.1 +/- 9.2 ms (P < 0.01) respectively] but not lactate [mean change -5.9 +/- 3.7 ms (P > 0.05) and -6 +/- 7.6 ms (P > 0.05) respectively]. No significant changes occurred in auditory brain stem response. The catecholamine response to hypoglycaemia was attenuated by lactate (P < 0.05 for adrenaline and noradrenaline). 4. Thus intravenous lactate prevents cerebral dysfunction during hypoglycaemia in insulin-dependent diabetes.