Objective: Findings on the influence of age and HIV on brain and cognition remain equivocal, particularly in aviremic subjects without other age or HIV-related comorbidities. We aimed to (a) examine the effect of HIV status and age on structural brain measurements and cognition, and (b) apply the machine learning technique to identify brain morphometric and cognitive features that are most discriminative between aviremic subjects with HIV on stable combination antiretroviral therapy (cART) and healthy controls.
Method: Fifty-three HIV-seropositive patients and 62 healthy controls underwent neuropsychological testing (executive functions, attention, memory, learning, psychomotor speed, fluency) and volumetric MRI scans. Voxel-based morphometry, ANCOVAs, machine learning, and multivariate regression were conducted to determine the between group differences in terms of relationship of HIV status, age, and their interaction on neurocognitive and structural brain measures.
Results: Volume and gray matter (GM) thickness of the caudate, parahippocampus, insula, and inferior frontal gyrus were smaller in seropositive subjects in comparison with healthy controls (HC). They also performed worse in complex attention and cognitive fluency tasks. Support vector machine (SVM) analysis revealed that the best between-groups classification accuracy was obtained based on cognitive scores encompassing complex attention and psychomotor speed, as well as volumetric measures of white matter and total gray matter; third, fourth, and lateral ventricles; amygdala; caudate; and putamen. Both voxel-based morphometry (VBM) and regression analysis yielded that HIV and aging independently increase brain vulnerability and cognitive worsening.
Conclusion: Patients with HIV on effective cART demonstrate smaller volumetric measures and worse cognitive functioning relative to seronegative individuals. There is no interaction between HIV infection and aging. (PsycINFO Database Record (c) 2019 APA, all rights reserved).