Association of Cortico-Striatal Engagement During Cue Reactivity, Reappraisal, and Savoring of Drug and Non-Drug Stimuli With Craving in Heroin Addiction

Abstract

Objective: The authors investigated cortico-striatal reactivity to drug cues (as compared with neutral and food cues), drug cue reappraisal, food cue savoring, and their correlations with heroin craving in individuals with heroin use disorder compared with healthy control subjects.

Methods: Cross-sectional changes in functional MRI blood-oxygen-level-dependent signal during a novel cue reactivity task were assessed in 32 individuals with heroin use disorder (mean age, 40.3 years; seven women) and 21 age- and sex-matched healthy control subjects (mean age, 40.6 years; eight women).

Results: Drug cue reactivity (vs. neutral cues) was significantly higher in the nucleus accumbens in the heroin use disorder group compared with the control group and nominally significantly higher in the orbitofrontal cortex (OFC); ventromedial prefrontal cortex (vmPFC) activity positively correlated with drug craving. Drug cue reactivity (vs. salient food cues) was also higher in the inferior frontal gyrus (IFG) in the heroin use disorder group compared with the control group. Drug reappraisal and food savoring (vs. passive viewing) showed increased IFG and supplementary motor area activity in all participants; in the heroin use disorder group, higher IFG/dorsolateral PFC (dlPFC) activity during drug reappraisal and rostral anterior cingulate cortex (ACC) activity during food savoring were associated with lower drug cue-induced craving and longer treatment, respectively. A direct comparison of regulation of reactivity to both salient cues revealed widespread group differences such that drug reappraisal activity was higher in the heroin use disorder group and food savoring activity was higher in the control group in both cortical (e.g., OFC, IFG, ACC, vmPFC, and insula) and subcortical (e.g., dorsal striatum and hippocampus) regions. Higher drug reappraisal versus food savoring in the dlPFC was associated with higher self-reported methadone dosage in the heroin use disorder group.

Conclusions: The results demonstrate cortico-striatal upregulation during drug cue exposure and impaired reactivity during processing of alternative non-drug rewards in the heroin use disorder group. Normalizing cortico-striatal function by reducing drug cue reactivity and enhancing natural reward valuation may inform therapeutic mechanisms for reducing drug craving and seeking in heroin addiction.

Keywords: Addiction Psychiatry; Opioids; Substance-Related and Addictive Disorders.

FIGURE 1.. The fMRI cue reactivity task paradigm in a study of cortico-striatal reactivity to drug and non-drug cues in individuals with heroin use disorder^a

^a The study procedures are depicted in panel A. During the fMRI session, participants completed three task runs, preceded and followed by pre- and posttask ratings, respectively. After MRI, participants completed computerized picture ratings. Panel B is a depiction of task procedures within a task run. In each run, participants were instructed to passively view blocks of drug, food, and neutral images during the look condition, reduce emotional reactivity to blocks of drug images, and increase emotional reactivity to blocks of food images during the reappraise and savor conditions, respectively. Each 16-second image block consisted of four images presented consecutively for 4 seconds each, with image blocks separated by 10-second intervals. The three conditions were presented in a randomized order.

FIGURE 2.. Drug cue brain reactivity in a study of cortico-striatal reactivity to drug and non-drug cues in individuals with heroin use disorder^a

^a Panel A shows clusters with increased activity in individuals with heroin use disorder compared with healthy control subjects during the look drug>look neutral contrast in the left nucleus accumbens (NAcc), extending to the right subgenual anterior cingulate cortex (sgACC)/ventromedial prefrontal cortex (vmPFC) and right orbitofrontal cortex (OFC). Panel B shows a cluster with increased activity in the right inferior frontal gyrus (IFG) in the heroin use disorder group compared with the control group during the look drug>look food contrast. Panel C shows a cluster with significant whole-brain voxel-wise positive correlations between posttask drug craving and the right vmPFC during the look drug>look neutral contrast (ld-ln) within the heroin use disorder group. Panel D shows a cluster with significant whole-brain voxel-wise positive correlations between Heroin Craving Questionnaire scores and the right OFC during the look drug>look neutral contrast within the heroin use disorder group. For visualization purposes, in all panels, bar graph and scatterplots depict blood-oxygen-level-dependent signal and correlations derived via 3-mm radius masks (7 voxels) centered on Montreal Neurological Institute coordinates from peak activity (white circles represent the approximate peak coordinates). Error bars represent standard error. Voxel resolution=2.1×2.1×2.1 mm.

FIGURE3.. Brain activation for reappraisal and savoring in a study of cortico-striatal reactivity to drug and non-drug cues in individuals with heroin use disorder^a

^a Panel A shows cortical regions of interest with significant between-group comparisons with the reappraise drug>savor food contrast and correlations within the heroin use disorder group. Shown are clusters with significant increased activity in the heroin use disorder group compared with the control group during the reappraise drug>savor food condition in the right orbitofrontal cortex (OFC), right rostral anterior cingulate cortex (rACC), and left anterior prefrontal cortex (aPFC)/inferior frontal cortex (IFG). In the heroin use disorder group, a cluster showed a significant whole-brain voxel-wise negative correlation between post-MRI cue-induced drug craving and the activation of the left IFG/dorsolateral prefrontal cortex (dlPFC) during the reappraise drug>look drug contrast (rd-ld). In this correlation, two outlier data points are depicted in red; a robust regression including the two outliers supported the significant effect when assuming a normal t distribution (t=−22.91, β=−20.515, SE5=0.177, p=0.007). In the heroin use disorder group, another cluster showed a significant whole-brain voxel-wise positive correlation between treatment length and the activation of the left rostral ACC during the savor food>look food contrast (sf-lf) (N=31; treatment length data were missing for one participant). In this correlation, an outlier data point is depicted in red; a robust regression including the outlier supported the significant effect when assuming a normal t distribution (t=4.39, β=0.004, SE=0.001, p<0.001). In the heroin use disorder group, a third cluster showed a significant whole-brain voxel-wise positive correlation between self-reported methadone dosage and the activation of the right dlPFC during the reappraise drug>savor food contrast (rd-sf) (N=22; methadone dosage data were missing for four participants). Panel B shows subcortical regions of interest with significant between-group comparisons for the reappraise drug>savor food contrast. Shown are clusters with significant increased activity in the heroin use disorder group compared with the control group in the right hippocampus and left caudate. Panel C shows a cortical cluster with (reappraise drug>look drug)>(savor food>look food) activity in the heroin use disorder group compared with the control group in the left insula. For visualization purposes, in all panels, bar graphs and scatterplots depict blood-oxygen-level-dependent signal and correlation derived via 3-mm radius masks (7 voxels) centered on the Montreal Neurological Institute coordinates from peak activity (white circles represent the approximate peak coordinates). Error bars represent standard error. Voxel resolution=2.1×2.1×2.1 mm.