You are here

Individualized prediction of HIV-associated neurocognitive impairment using machine learning techniques

Individualized prediction of HIV-associated neurocognitive impairment using machine learning techniques

Even in the era of combination antiretroviral therapy, nearly half of persons living with HIV in the US experience HIV-associated neurocognitive impairment (NCI). The purpose of this study is to discover the neural signatures of HIV-associated NCI using machine learning techniques. The proposed analysis capitalizes on existing data from three projects that collected comprehensive neurocognitive batteries, multimodal MRIs, in-depth substance use histories, and a wide range of phenotypic data, including medical records with biological indicators of HIV disease progression. This data has been harmonized and merged, and the final sample includes 103 HIV+ adults. The input features will be white and gray matter metrics of 375 atlas-derived regions, as well as substance use, demographic, and clinical variables. The goal is to classify the sample based on NCI status.