Recent work regarding the analysis of brain imaging data has focused on examining functional connectivity of the brain. In this paper, we developed and applied a novel framework to examine the modulation of functional connectivity networks during resting state or experimental stimuli. Hierarchical clustering within anatomically parcellated brain regions is used to cluster voxels into anatomically local, functionally homogenous groups. Parcellation allows for the assessment of connectivity at a region of interest level, rather than a voxel-specific level. The framework utilizes a mixed-effects model of the Fisher Z-transformed correlation between pairs of spatially distinct brain regions to assess the modulation of, or differences in, connectivity networks across multiple sessions or between groups of patients. A permutation test was employed to control family-wise type I error. We applied our method to evaluate the effects of a non-selective opioid receptor antagonist, naloxone, on resting state functional connectivity networks with multi-session multi-subject (n = 10) functional magnetic resonance imaging data. Naloxone induced a significant reduction in resting state connectivity between the right thalamus and the right parahippocampal gyrus when compared to saline solution.