Control of brain seizures after exposure to nerve agents is imperative for the prevention of brain damage and death. Animal models of nerve agent exposure make use of pretreatments, or medication administered within 1 minute after exposure, in order to prevent rapid death from peripheral toxic effects and respiratory failure, which then allows the testing of anticonvulsant compounds. However, in a real-case scenario of an unexpected attack with nerve agents, pretreatment would not be possible, and medical assistance may not be available immediately. To determine if control of seizures and survival are still possible without pretreatment or immediate pharmacologic intervention, we studied the anticonvulsant efficacy of the GluK1 (GluR5)/α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor antagonist (3S,4aR,6R,8aR)-6-[2-(1(2)H-tetrazole-5-yl)ethyl]decahydroisoquinoline-3-carboxylic acid (LY293558) in rats that did not receive any treatment until 20 minutes after exposure to the nerve agent soman. We injected LY293558 intramuscularly, as this would be the most likely route of administration to humans. LY293558 (15 mg/kg), injected along with atropine and the oxime HI-6 at 20 minutes after soman exposure, stopped seizures and increased survival rate from 64% to 100%. LY293558 also prevented neuronal loss in the amygdala and hippocampus, and reduced neurodegeneration in a number of brain regions studied 7 days after soman exposure. Analysis of the LY293558 pharmacokinetics after intramuscular administration showed that this compound readily crosses the blood–brain barrier. There was good correspondence between the time course of seizure suppression by LY293558 and the brain levels of the compound.