In 14 patients operated upon for focal cerebral seizures under local anesthesia, cortical electrical activity was compared with the levels of nicotinamide adenine dinucleotide (NADH) observed fluorometrically. NADH levels fell 3 to 15% in response to 5-second intervals of cortical stimulation in 42 of 70 observations. Although a rough correlation was seen between the quantity of current delivered (milliamperes X seconds) and the NADH decrease, this varied from case to case. The presence of cortical afterdischarge often, but not invariably, corresponded to a greater percentage of change in the NADH levels. Averaging the NADH response to sporadic interictal epileptiform discharges failed to demonstrate concomitant NADH reductions. A similar lack of change was seen in four patients in whom low frequency spike foci were induced by topically applied penicillin in cortex destined for excision. Preliminary studies of the topography of spread of NADH change after cortical stimulation indicate that this is usually asymmetrical in human epileptogenic cortex. Under experimental conditions in cats, it seemed possible to differentiate primary from projected epileptiform activity, in that the projected activity had little or no concomitant fall in the NADH level after the electrographic spike. Pathological examination of the excised sites of NADH recording showed, with one exception, fibrous astrocytic transformation of the central cortex layers.