Abstract Body

HIV-associated neurocognitive impairment (NCI) remains a prevalent comorbidity that impacts daily functioning and increases morbidity. While HIV infection is known to cause widespread disruptions in the brain, different MRI modalities have not been effectively integrated. This study applied 3-way supervised fusion to investigate how structural and functional co-alterations affect cognitive function.

Participants completed comprehensive neuropsychological testing and had a multi-modal MRI scan to acquire high-resolution anatomical, diffusion-weighted, and resting-state functional images. Pre-processed data was reduced using voxel-based morphometry (gray matter volume), probabilistic tractography (fractional anisotropy), and regional homogeneity (intrinsic functional connectivity), respectively. We applied multi-modal canonical correlation analysis with reference plus joint independent component analysis (MCCAR+jICA), using global neurocognitive functioning as the reference.

The sample includes 70 HIV+ and 69 HIV- adults who were matched on age (M=38.7), gender (70% male), and race (68% African-American). HIV+ participants had lower global neurocognitive functioning (p=.005), with differences in the domains of learning (p=.006), memory (p=.006), and executive function (p=.026). Figure 1 shows the independent joint component that significantly correlated with global neurocognitive functioning in all three modalities. Gray matter regions included thalamus and cerebellum. White matter regions included the cingulum tract of the cingulate and hippocampus, inferior and superior longitudinal fasciculus, and uncinate fasciculus. Functional regions included posterior parietal, lateral prefrontal, orbitofrontal, anterior cingulate, precuneus, and insular cortices. HIV+ status was associated with lower gray matter volume (p=.038) and lower fractional anisotropy (p=.028) in this component. Duration of HIV disease and nadir CD4 cell count were also associated with gray matter volume and functional connectivity in identified independent components.

These results suggest that linked structural and functional deficits in several brain networks are related to HIV-associated NCI. As MRI becomes more commonplace in HIV care, multimodal fusion may provide neural biomarkers to support diagnosis and treatment of NCI.