Depression is a common and disabling condition and is especially disabling for patients who also have epilepsy. Antidepressants, particularly the tricyclic antidepressants are well known to be associated with seizure activity, but this is a very neglected area of research. Most of the data on the proconvulsive effects of antidepressants come from either work in animal models or from research into the effects of antidepressants in overdose. Both of these situations may tell us little about the behaviour of antidepressants in patients with epilepsy. The selective serotonin [5-hydroxytryptamine; 5-HT] reuptake inhibitors have a low seizure propensity, are well tolerated in overdose and have a favourable adverse effect profile, making them suitable as first line treatments for depression in patients with epilepsy. Other antidepressants, e.g. trazodone, moclobemide, mirtazepine, are also likely to have minimal proconvulsive effects, but adverse effects, interactions with other drugs, especially anticonvulsants, or the lack of clinical data may make their use less attractive. Although this review has focused on these clinically important issues it is clear that considerably more research needs to be undertaken on the seizure propensity and clinical efficacy of antidepressants in patients with epilepsy.