Working memory (WM) deficit in schizophrenic patients has been well established. Still, underlying biological substrate of the impairment is not clear. Among neurotransmitter hypotheses of schizophrenia, N-methyl-D-aspartate (NMDA) receptor model is mostly supported, considering that NMDA receptor antagonist can elicit both psychosis and cognitive impairment observed in schizophrenic patients. In current study, to test the neuropsychological and the electrophysiological effects of NMDA receptor in WM, event-related potentials (ERPs) of Sternberg's short-term memory scanning task (SMST) were analyzed in 10 healthy subjects under intravenous administration of a subanesthetic dose of ketamine (0.65 mg/kg/h) or placebo (normal saline). Late positive component (LPC) of ERP was hypothesized to reflect later stage of WM. Brief Psychiatric Rating Scale score was significantly increased (t=-5.75, df=9, P<.001) and correct response rate was significantly decreased (t=2.21, df=9, P=.054) after ketamine administration. Neither reaction time nor LPC latency, which reflect memory scanning time, was changed. Amplitude of LPC was significantly reduced after ketamine administration (z=-2.31, number of observations=120, P=.021). In conclusion, NMDA receptor antagonist administration elicited WM deficit both in behavioral and electrophysiological level. Electrophysiological component reflecting later stage of WM was impaired by NMDA antagonist.