Antiepileptic drugs (AEDs) protect against seizures by modulation of voltage-gated sodium and calcium channels, enhancement of GABAA (γ-aminobutyric acid, type A) receptor-mediated synaptic inhibition, and inhibition of ionotropic glutamate receptor-mediated synaptic excitation. Sodium channel-blocking AEDs, including phenytoin, carbamazepine and lamotrigine, inhibit high-frequency repetitive spike firing during the spread of seizure activity without affecting ordinary ongoing neural activity. These drugs bind preferentially to depolarized sodium channels and induce a non-conducting state that is similar to channel inactivation, but from which recovery is much slower, allowing the block to accumulate during repetitive activation of the channel, as occurs with epileptic bursting. In addition, the drug binds slowly so that there is preferential block of firing during sustained epileptic depolarizations. Persistent or non-inactivating sodium current represents only a fraction of the sodium current, but might contribute to the initiation and maintenance of epileptiform activity. Preferential inhibition of persistent sodium current probably contributes to the protective activity of phenytoin and possibly other sodium channel blocking AEDs. Gabapentin binds with high affinity to the auxiliary calcium channel subunits α2δ-1 and α2δ-2. The functional consequences of this interaction are not fully defined, but recent studies indicate that gabapentin might inhibit calcium currents, resulting in reduced excitatory neurotransmission. In the thalamus, T-type calcium channels are essential for the abnormal oscillatory behaviour that underlies generalized absence seizures. Ethosuximide inhibits these channels, accounting for its anti-absence activity. Inhibitory GABAA receptor-mediated synaptic interactions are important in restraining the natural tendency of brain circuits in regions that are susceptible to epileptic activity (including the neocortex, hippocampus and amygdala) to undergo the transition into synchronized epileptiform activity. Many AEDs enhance GABAA receptor inhibition either through positive modulatory interactions with GABAA receptors (benzodiazepines, barbiturates, felbamate and topiramate) or by modifying the dynamics of GABA-mediated inhibitory synaptic function, as is the case for vigabatrin, an irreversible suicide inhibitor of the GABA degradative enzyme GABA transaminase, and tiagabine, an inhibitor of the high-affinity GABA transporter GAT1. Several marketed AEDs might act partly by inhibition of ionotropic glutamate receptors, including felbamate, which inhibits NMDA (N-methyl-D-aspartate) receptors, and topiramate, which inhibits kainate receptors. There is remarkable overlap between the ion channel targets of AEDs and human epilepsy genes, illustrating the pivotal importance of ion channels in epilepsy. In many cases, the AEDs and mutations induce functionally opposite effects on channel behaviour. So, whereas the mutations lead to increased seizure susceptibility through gain-of-function effects on voltage-gated sodium and calcium channel gating or reduced efficacy of GABAA receptors or potassium channels, AEDs inhibit sodium or calcium channels, or enhance the activity of GABAA receptors or potassium channels.