Erratum: Xie et al., “Neural Network Connectivity Following Opioid Dependence Is Altered by a Common Genetic Variant in the µ-Opioid Receptor, OPRM1 A118G”

Validation of FOS automated quantification method in randomly selected regions and experimental groups. Each (X, Y) value in a data point is the detected number of cells (auto, manual) in ROI on a randomly selected slice containing that region. The automatic quantification method effectively captured the relative cell counts compared with manual identification; however, there is interregional wise variance in the overall detection ratio compared with that in the manual (by slope). To avoid this, we used percentage change (increase over baseline) data for interregional comparison instead of raw c-Fos density expression (specified in the equation above: detect ratio is different among regions).

Fos percentage change over baseline of cortical regions in each genotype, treatment, and sex group. PL, prelimbic area; ILA, infralimbic area; ACAd, dorsal anterior cingulate cortex area; ACAv, ventral anterior cingulate cortex area; AId, agranular insular dorsal part; AIv, agranular insular ventral part; CLA, claustrum; CP, caudate putamen; NAc/ACB, nucleus accumbens; CeA, central amygdala; BLA, basolateral amygdala; BST, basal striatal terminalis; GPe, globus pallidus external segment; GPi, globus pallidus internal segment; PALv, ventral pallidum; hippocampus, hypothalamus; PVT, paraventricular nucleus of thalamus; MH, medial habenula; LH, lateral habenula; DG, dentate gyrus; LHA, lateral hypothalamus; VTA, ventral tegmental area; PAG, periaqueductal grey; SNr, substantial nigra reticular part; SNc, substantial nigra compact part. Mean ± SD; Tukey's post hoc: *p < 0.05, **p < 0.01,***p < 0.001, ****p < 0.0001.

Interregional characteristics of A112G FOS percentage change. A, Correlation matrix based on FOS percentage change of expression over baseline and grouped by macroanatomical regions in acute morphine exposure (top row) and chronic morphine exposure (bottom row). Each grid in the matrix represents Pearson’s correlation coefficient between two regions. B, Correlation network (2D, force-directed) depicts positive pairwise correlations where correlation strength decides the dragging force between two regions. ROIs are colored according to their mesoscale structure (blue, cortex; brown, striatum; gold, thalamus; red, amygdala; violet, pallidum; green, midbrain; yellow, hippocampus and hypothalamus). C, Three-way ANOVA on the small-world coefficient of bootstrapped correlation network showed significant treatment, sex, and genotype effects. Tukey's post hoc multiple comparison test showed significant between-group difference. Small-world coefficient sigma, calculated by the avg clustering coefficient and avg shortest path length of the network with respect to random graphs (N = degree of correlation network). D, All correlation networks show sigma > 1: small-world property is maintained (mean ± SD; **p < 0.01, ****p < 0.0001).

Brain state transition analysis. A, Minimum control energy required to transition from an acute morphine-exposed state to a chronic dependence state when all regions are controlled for each genotype and sex after bootstrapping (red line, median). GG females showed significantly higher minimum energy requirements to transition to an opioid-dependent state compared with AA females (ANOVA; R² = 0.7753; F_(3,159,996)= 184,016; p < 0.0001; Tukey; p < 0.0001; Q = 803.2). GG males showed lower minimum energy requirements compared with AA males (Tukey; p < 0.0001; Q = 64.57). B, Minimum control energy required to transition from acute morphine to morphine-dependent brain state after suppressing input to cortical regions (dACA, vACA, AId, AIv, CLA, PL, ILA) for each genotype and sex. C, Percent increase in minimum control energy following suppression of control input to each region. D, Minimum control energy for state transition after weighting control inputs by expression density map of μ-, δ-, and κ-receptors. E, Minimum control energy required to maintain each brain state (median ± 95% CI; *p < 0.05, ***p < 0.001, ****p < 0.0001). F, Pearson’s correlation of minimum control energy required to transition from an acute to a chronic state with small-world coefficient sigma on bootstrapped (N = 8) and z-transformed data. A significant negative correlation was identified (mean ± 95% prediction bands).