Beta (parachlorophenyl)-gamma-aminobutyric acid (baclofen) (Lioresal) has been given in a single blind trial to 17 schizophrenic patients selected among the hospitalized population from the Lillhagen mental hospital, which serves the city of Göteborg. The selection was made on the basis of a relatively young age, a short disease history and normal physical and laboratory findings. At the beginning of the investigation almost all the patients were under treatment with various neuroleptics and were stabilized at the most satisfactory level with a quite acceptable adaptation to daily activities. Previous attempts to reduce the dosages of medicine were a failure.Prior to treatment with baclofen, a complete medical history was obtained from each patient. A physical and laboratory routine examination was done. Routine progress notes, ward and doctors global evaluation were reported periodically. A Brief Psychiatric Rating Scale was performed prior to treatment and after 1 and 6 months of treatment, and thereafter with 6 months intervals. A linear regression analysis of these periods was done and the significance of the variation tested accordingly.Treatment with baclofen began with small starting doses of 15 mg daily for a week, with a progressive increase to the optimal doses, usually 30–50 mg daily. Further increase did not produce any noticeable effect. This dose was maintained during the whole investigation period which ranged from 8 to 30 months with a mean observation period as long as 22 months, each patient being his own control. When the optimal level of baclofen was reached the doses of neuroleptics were carefully decreased, caution being observed to keep the patient at the same behavioral and affective level.The neuroleptics used prior to the treatment with baclofen were grouped as phenothiazines, butyrophenones and depotpreparates. Following the calculation of the mean dose for each group, each individual dose was normalized to the mean and a linear regression analysis performed. The significance of the variations was tested with t-test.Following treatment with baclofen a significant decrease of accompanying neuroleptics, especially phenothiazines and butyrophenones could be done satisfactorily, while previously, any such attempt would have lead to an impairment of the patient condition.On the other hand 14 out of 17 patients responded to baclofen with a marked improvement which persisted throughout the whole treatment period, although it varied in intensity.It is essential to note that, in order to attain such results the patients should be gradually, and not suddenly, deprived of the conventional neuroleptic therapy which they have had for many years.Some patients, already ameliorated, interrupted prematurely the treatment and rapidly presented a recidive. It is noteworthy that the patients who did not react to baclofen also had a very bad response to conventional neuroleptics. No side effects of baclofen were noticed.Together with other types of investigation, such as EEG studies(Badr et al. 1977) these results strongly suggest that baclofen could be a putative associative psychotropic agent. Another interesting aspect of this product is that it offers the possibility to reduce the administration of conventional neuroleptics hence decreasing the side-effects of these drugs.