In all, 22 reports of 20 randomized, controlled rehabilitation studies were evaluated. In 18 of these, the design of the trial was parallel, with a cross-over format being employed in the remaining 2 instances. Seven studies related to intensive rehabilitation during the early post-acute period. In six others, specific techniques--sometimes associated with traditional physiotherapy procedures--were compared: biofeedback, perceptual retraining, and amphetamine treatment. Eight experiments evaluated speech therapy in aphasia. Frequently, methodological considerations limited the interpretation of the results. The review showed that: 1) as regards activities of daily life and motor function, differences as between rehabilitation in stroke units on the one hand and non-rehabilitation or rehabilitation in medical wards on the other, were detected in relatively few quality studies and remained particularly inconclusive insofar as life in the home environment was concerned; 2) rehabilitation for aphasia and perceptual dysfunction proved effective for at least several months after acute stroke; 3) in general, examination of the reports cited revealed no differences in the effects of biofeedback and perceptual retraining vs conventional physiotherapy; 4) rehabilitation, whether administered by specialists or amateur caregivers purpose-trained by specialists, proved equally effective for aphasia. These conclusions constitute a valuable basis for the development and evaluation of modern rehabilitation programs for stroke patients.