This placebo-controlled, double blind, cross-over study tested the efficacy of two different doses of Peptide T in the treatment of nine intravenous drug users with early AIDS dementia who were also receiving methadone and AZT. Subjects received Peptide T doses of either 15 or 1.5 mg daily for four weeks. Neuropsychological performance improved in four of five patients treated with the high dose, but at the lower dose, three of four patients showed no improvement on Peptide T when compared with placebo. When subjects who received the high dose were compared with those who received the low dose, a significant dose effect was found only during the active phase of the trial even after correction for differences in level of functioning at baseline.