Characterization of Resting-State Functional Connectivity Changes in Hypertension by a Modified Difference Degree Test.

Abstract

Hypertension affects over a billion people worldwide, and the application of neuroimaging may elucidate changes brought about by the disease. We have applied a graph theory approach to examine the organizational differences in resting-state functional magnetic resonance imaging (rs-fMRI) data between hypertensive and normotensive participants. To detect these groupwise differences, we performed statistical testing using a modified difference degree test (DDT). Structural and rs-fMRI data were collected from a cohort of 52 total (29 hypertensive, 23 normotensive) participants. Functional connectivity maps were obtained by partial correlation analysis of participant rs-fMRI data. We modified the DDT null generation algorithm and validated the change through different simulation schemes, and then applied this modified DDT to our experimental data. Through a comparative analysis, the modified DDT showed higher true positivity rates (TPR) when compared with the base DDT while also maintaining false positivity rates below the nominal value of 5% in nearly all analytically thresholded trials. Applying the modified DDT to our rs-fMRI data showed differential organization in the hypertension group in the regions throughout the brain including the default mode network. These experimental findings agree with previous studies. While our findings agree with previous studies, the experimental results presented require more investigation to prove their link to hypertension. Meanwhile, our modification to the DDT results in higher accuracy and an increased ability to discern groupwise differences in rs-fMRI data. We expect this to be useful in studying groupwise organizational differences in future studies.