Despite preventing dementia via antiretroviral therapy, individuals living with chronic HIV show structural brain alterations and cognitive impairments, which vary in aging populations. Consequently, group-level comparisons with uninfected individuals retain heterogeneity. Machine learning algorithms can predict an individual’s age based on neuroimaging features. However, brain-predicted age might be different from the true biological age, a difference known as the brain-age gap. “We have previously seen that an algorithm trained to guess someone’s age from their brain magnetic resonance imaging would overestimate the ages of people with HIV, on average, compared to people without. We also saw huge variability in ‘brain age’ within those groups. Thus, we asked to what extent it is HIV itself or all the challenges that often accompany it that contribute to this wide range of aging patterns,” explains Kalen Petersen, the first author of the study.
Petersen et al. analyzed data from people with and without HIV by using a joint model to detect common factors related to brain aging. They found that, in individuals with HIV, brain aging is best explained not by their serostatus (such as having HIV) or viral load (the amount of virus detectable in the blood) but instead by a combination of clinical, social and comorbid risk factors. The researchers identified co-infection with hepatitis C and cardiovascular disease as important comorbidities, among others. “One of the more novel aspects of the study is the inclusion of geospatial data on neighborhood socioeconomic factors. This is calculated by the latitude and longitude of a participant’s address and then plugged into public databases to get information about such factors as employment, income, and education,” adds Petersen.
This is a preview of subscription content, access via your institution