Despite widespread use of antiretroviral therapy, nearly half of the 1.2 million Americans living with HIV experience neurocognitive impairments (NCI) that negatively impact daily functioning. HIV-associated NCI is estimated to be the most common form of midlife NCI worldwide. Drug abuse, a highly prevalent comorbidity in HIV+ persons, increases the risk of HIV-associated NCI. Unimodal MRI studies have shown that HIV and drug abuse are each linked with a distinct set of structural and functional alterations in the brain, but how they interact is not well understood. Conventional unimodal analyses are limited in their ability to characterize complex neuropsychiatric diseases because each modality provides an incomplete view of the brain. In contrast, we propose to use innovative fusion approaches that exploit the richness of multimodal data to discover covariations across multiple neuroimaging modalities and cognitive measures simultaneously. This proposal capitalizes on existing data from 3 completed NIDA-funded projects on the neurobehavioral effects of HIV and drug abuse. When combined, the dataset will consist of 217 unique cases (108 HIV+ and 109 HIV-) with in-depth substance use histories, multimodal brain images (i.e., structural, diffusion, and resting-state functional MRI), comprehensive neurocognitive batteries, and a wide range of other phenotypic data. The central hypothesis is that HIV-specific co-alterations in gray matter volume and white matter integrity shape neural functioning and in turn NCI, and that cocaine use worsens these alterations due to its long-term effects on neural circuitry. Using complementary multimodal analyses (supervised fusion and connectomics), we aim to: (1) identify neural biomarkers of HIV neurological disease and associated NCI; and (2) test cocaine as a moderator of HIV-specific co-alterations in brain structure and function and associated NCI. In addition, the project will develop new, advanced methods to improve MRI data quality for cross-study harmonization and to optimize the prediction of NCI based on multimodal indices. This amalgamated dataset provides an unprecedented opportunity to test new and clinically important hypotheses about the neural substrates of HIV- associated NCI in the context of drug abuse. The proposal responds directly to the need for research to ?characterize brain morphology or function that is aberrant as a consequence of chronic drug use? and ?[its role] in the evolving dynamics of HIV/AIDS pathogenesis, treatment, prevention, and service delivery? [PAR- 16-234]. This research also addresses high priority topics for AIDS-designated funding, including investigation of neurological complications [NOT-15-137]. Building upon a sound premise and robust preliminary data, this innovative project has strong potential to identify appropriate neural biomarkers of neuroHIV that may facilitate diagnosis and treatment monitoring in active drug users and serve as targets for clinical interventions to mitigate the burden caused by HIV-associated NCI.

Public Health Relevance

Nearly half of HIV-infected Americans experience neurocognitive impairments that impact daily functioning, and the prevalence is even higher among persons who are addicted to stimulant drugs. Using innovative multimodal data fusion approaches, this study will identify the distinct and shared neural signatures of HIV neurological disease and cocaine use disorder that are predictive of neurocognitive impairment. Capitalizing on existing datasets, this project has strong potential to identify appropriate neural biomarkers to facilitate diagnosis and treatment of HIV-associated neurocognitive impairments in active drug users.

National Institute of Health (NIH)
National Institute on Drug Abuse (NIDA)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Lin, Yu
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Duke University
Schools of Medicine
United States
Zip Code