A placebo-controlled, partial cross-over, double-blind, randomized study was performed with 46 adults with sleep apnea-hypopnea syndrome (SAHS) to determine the effect of therapeutic and subtherapeutic (0-1 cm H(2)O) nasal continuous positive airway pressure (CPAP) treatment on polysomnographic and neuropsychological testing. The following neuropsychological tests were administered: Geriatric Depression Scale, Trail Making A and B, Digit Span Test Forward and Backward, Epworth Sleepiness Scale, SteerClear, Digit Symbol, Controlled Oral Word Association, and Complex Figure Recall. Compared with results without CPAP, subtherapeutic CPAP did not affect any measured polysomnographic parameter. Comparison of neuropsychological test results obtained between the initial periods of effective treatment (Group 1, 16.1 d; Group 2, 19.6 d; p = NS) in all subjects showed significant improvements in Digit Symbol, Digit Span Backward, and Complex Figure tests. However, there were no group differences in changes in test results during the period when one group was on effective CPAP and the other on ineffective CPAP (Group 1, 16.1 d; Group 2, 13.9 d; p = NS). The results indicate the feasibility and importance of using ineffective CPAP as a placebo treatment and the importance of including a placebo control in studies evaluating the effect of treatment on neuropsychological function in SAHS.