Background:  Functional MRI studies demonstrate brain activation abnormalities in HIV-infected individuals. The purpose of this study was:  to determine reproducibility of functional MRI in HIV-uninfected controls; to compare brain activation in HIV-infected and -uninfected individuals; and to determine changes in brain activation in HIV-infected persons beginning HAART.

Methods:  Controls (n = 9) were HIV-at-risk but -uninfected and scored normally on neuropsychological tests. HIV-infected subjects included: no therapy (NoRx, n = 4), stable HAART (StableRx, n = 5), and beginning HAART (NewRx, n = 9). Subjects had CD4 ≤ 350 cells/uL, were on no antiretroviral therapy for ≥ 12 weeks (NoRx and NewRx) or on stable therapy for > 20 weeks with plasma HIV-1 RNA < 200 copies/mL (StableRx). Controls and subjects were matched for age, gender, and education. At entry and 8 weeks later, functional MRI during a 2-back working memory test with visual stimuli and neuropsychological tests were performed. functional MRI analysis was based on average z-score for each participant in 4 regions of interest:  right and left parietal and dorsolateral prefrontal cortex. Between group comparisons were made with a Wilcoxon rank-sum test and within group comparisons with a Wilcoxon signed rank test.

Results:  At entry, location of brain activation was the same for all 4 groups and there were no group differences in amount of activation. At 8 weeks, functional MRI activation was greater in controls than NewRx in all 4 regions, and StableRx > NewRx in 3 of 4 regions (p < 0.05). Controls also showed significant activation increases in 3 of 4 regions (p < 0.05), and StableRx showed trends toward increases in 3 of 4 regions (p = 0.08). No changes in functional MRI activation or neuropsychological tests results were seen in NewRx subjects. However, when subjects were divided into 3 groups by entry plasma HIV RNA, those with the lowest values (< 100 copies/mL) had significantly greater brain activation in 3 of 4 regions at 8 weeks than the other 2 groups (p < 0.02). Moreover, subjects with low HIV RNA at both entry and 8 weeks had significantly greater activation after 8 weeks than subjects whose plasma HIV RNA went from high to low (p < 0.02). 

Conclusions:  As with neuropsychological tests, functional MRI shows a learning effect for controls. While HAART may result in improved brain function, sustained control of plasma viremia for > 8 weeks may be required to demonstrate this by functional MRI or neuropsychological tests.

Keywords: neuroimaging; treatment response; cognitive impairment