This study explores the relationship of immune dysfunction to the neuropsychological performance of individuals infected with HIV-1. Fifty-five HIV-positive homosexual men and 37 negative homosexual controls were evaluated using neuropsychological measures, physical exams, and measures of immune functioning. There were no significant differences favoring HIV-negative subjects over HIV-positive subjects. HIV-positive subjects, in fact, performed slightly better on attention and memory procedures. The HIV-positive subjects were then stratified according to the Centers for Disease Control symptom groupings (Group II, asymptomatic, n = 19; Group III, lymphadenopathy, n = 17; and Group IVA or C-2, symptomatic, non-AIDS, (n = 19). There were no significant neuropsychological differences among the three CDC groups. The HIV-positive subjects were also stratified on two measures of immune functioning: absolute CD4 counts (< 200, 201-400, > 400) and beta 2-microglobulin (beta 2M) (> or = 5.0, 3.0-5.0, < 3.0). Individuals with greater immune compromise, as measured by beta 2M, were more impaired on measures of attention and memory and had greater overall neuropsychological impairment (p < 0.05). Furthermore, 57% of the subjects who were abnormal on beta 2M were also impaired on measures of attention and memory, whereas only 14% of those with normal beta 2M were impaired on these same measures (p < 0.05). These results suggest that HIV-positive asymptomatics without evidence of immune compromise do not appear to be at greater risk of cognitive impairment than HIV-negative controls. However, for those HIV-positive individuals who are immune-compromised (even while asymptomatic), there is increased risk of neuropsychological impairment. These results also suggest that knowledge of serostatus and the use of the CDC classification system alone are insufficient in exploring the development of neuropsychiatric changes in HIV-1 infection.