We examined the effects of computer-provided reading activities on language performance in chronic aphasic patients. Fifty-five aphasic adults were assigned randomly to one of three conditions: computer reading treatment, computer stimulation, or no treatment. Subjects in the computer groups used computer 3 hours each week for 26 weeks. Computer reading treatment software consisted of visual matching and reading comprehension tasks. Computer stimulation software consisted of nonverbal games and cognitive rehabilitation tasks. Language measures were administered to all subjects at entry and after 3 and 6 months. Significant improvement over the 26 weeks occurred on five language measures for the computer reading treatment group, on one language measure for the computer stimulation group, and on none of the language measures for the no-treatment group. The computer reading treatment group displayed significantly more improvement on the Porch Index of Communicative Ability "Overall" and "Verbal" modality percentiles and on the Western Aphasia Battery Aphasia "Quotient" and "Repetition" subtest than the other two groups. The results suggest that (a) computerized reading treatment can be administered with minimal assistance from a clinician, (b) improvement on the computerized reading treatment tasks generalized to non-computer language performance, (c) improvement resulted from the language content of the software and not stimulation provided by a computer, and (d) the computerized reading treatment we provided to chronic aphasic patients was efficacious.