Background:
The functional connectivity of global brain networks, quantified using graph theory metrics, provide insights into the efficiency and other characteristics of brain network functioning at a whole brain level. Such metrics are widely studied in numerous neurological and psychiatric conditions under a variety of task demands but have not yet been evaluated in virally-suppresesed people with HIV (VS-PWH). Here we examined global brain network organization in virally-suppressed people with HIV under three functional MRI conditions including resting state MRI (rsfMRI) and the performance of verbal memory and go/no-go tasks.
Methods:
A total of 56 participants (69.6% VS-PWH; 51.7% female, 69.6% black) completed the MRI session and performed a battery of cognitive tests using the iPad-based platform Brain-Baseline Assessment of Cognition and Everyday Functioning. Graph theory metrics of the global functional brain network were computed in the weighted positive network, determined by positive correlations in the BOLD activity, to yield assortativity, characteristic path length (CPL), clustering coefficient (CC), global efficiency (GE), modularity (Q), and average node strength (ANS). Multiple regressions were performed to determine serostatus effects and associations with cognitive performance, adjusting for sex, age, race, ethnicity and education.
Results:
In resting state, compared to controls, VS-PWH showed lower CC (p=.03), GE (p=.02), and ANS (p=.02) and higher CPL (p=.04) and Q (p=.04). In the verbal memory task, compared to controls, VS-PLWH also showed lower CC (p=.01), GE (p=.02), and ANS (p=.01) and higher CPL (p=.04) and Q (p=.005). No HIV-serostatus effects were observed in the go/no-go task. In the total sample, there were significant associations of CPL, CC, GE, and ANS with processing speed (reaction time on flanker, Stroop, Trail Making, and fine motor skills task), working memory (accuracy on n-back) and visual spatial learning test (total correct across trials).
Conclusions:
During both the resting state and performance of a verbal memory task, VS-PLWH show less efficiently organized network structure (GE), less small-world network structure (CC, CPL), and less distributed connectivity (Q and ANS). These metrics relate to a number of cognitive processes including processing speed, working memory, and visual spatial learning. Here we show graph theory metrics may serve as biomarker for VS-PLWH to monitor disease progress and to evaluate new therapeutics.