SUMMARY Twelve patients with severe intractable epilepsy were treated by chronic cerebellar stimulation under double-blind conditions for six months. No reduction in seizure frequency occurred that could be attributed to stimulation, though eleven of the patients considered that the trial had helped them. One patient experienced fewer episodes of incontinence during stimulation. Cerebellar stimulation in its present form cannot be recommended for the treatment of severe intractable epilepsy.Therapeutic effects of cerebellar stimulation in severe epilepsy were first reported by Cooper.' A later report summarised the outcome in a total of 32 patients.2 Seizure frequency fell to half or less of the pre-operative rate in 18 patients and nine were regarded as therapeutic failures. One died at operation and four others died subsequently in their sleep. Two of these had not responded to stimulation, one had been free of grand mal and petit mal attacks for the six week period between implantation and death and the response of the fourth was not described. Treatment was uncontrolled and assessment open.Several other reports3-6 have since appeared which have supported the concept that cerebellar stimulation improves epilepsy, although the effect was less pronounced than in Cooper's series. Only one double-blind study7 has been performed. This involved five patients, of whom three, on the criteria recommended by Cooper et al,8 might not have been expected to respond favourably.Severe epilepsy has an appreciable morbidity. Mortality is also increased, especially in young adult males. Lennox9 quotes an overall death rate of 11-4 in the age group 25-45 compared with a rate of 3-1 in the general population. Therefore it is extremely important to establish whether or not this treatment should be recommended for occasional cases of severe intractable epilepsy. The present uncertainty