Abstract In previous studies with Sprague-Dawley rats, we demonstrated that NMDA channel blockade during waking massively stimulates the delta (1–4 Hz) EEG of non-rapid eye movement (NREM) sleep. However, non-competitive channel blockers also produce neurotoxicity that is manifested by posterior cingulate vacuolization and heat shock protein production. These neurotoxic effects can be blocked by coadministering gabaergic drugs, including barbiturates and benzodiazepines, with the MK-801. To determine whether delta stimulation by MK-801 would be similarly blocked, we administered an anesthetic dose (40 mg kg ) of pentobarbital followed immediately by 0.3 mg kg of MK-801. Neither the MK-801 motor syndrome nor the NREM delta stimulation was blocked. When the rats recovered from nearly two hours of barbiturate anesthesia, they behaved as though they had just received the MK-801 injection, exhibiting the typical motor syndrome, spikes in the waking EEG and strong stimulation of NREM delta EEG. These findings support our previous evidence that NREM delta stimulation by NMDA channel blockade does not depend on toxic brain changes. They also raise interesting questions regarding the fate of MK-801 during pentobarbital anesthesia. We propose that the drug is not metabolized during the period of anesthesia because it is sequestered within the NMDA cation channel. However, neurons do not respond to the channel block because they have been rendered inert by the anesthesia. When the neurons emerge from anesthesia, the cascade of MK-801 events unfolds. This and other possible explanations can be tested experimentally. Establishing the fate of MK-801 during barbiturate anesthesia could shed new light on the cellular processing of non-competitive NMDA channel blockers.