Cognitive control of drug craving inhibits brain reward regions in cocaine abusers

Abstract

Loss of control over drug taking is considered a hallmark of addiction and is critical in relapse. Dysfunction of frontal brain regions involved with inhibitory control may underlie this behavior. We evaluated whether addicted subjects when instructed to purposefully control their craving responses to drug-conditioned stimuli can inhibit limbic brain regions implicated in drug craving. We used PET and 2-deoxy-2[18F]fluoro-d-glucose to measure brain glucose metabolism (marker of brain function) in 24 cocaine abusers who watched a cocaine-cue video and compared brain activation with and without instructions to cognitively inhibit craving. A third scan was obtained at baseline (without video). Statistical parametric mapping was used for analysis and corroborated with regions of interest. The cocaine-cue video increased craving during the no-inhibition condition (pre 3+/-3, post 6+/-3; p<0.001) but not when subjects were instructed to inhibit craving (pre 3+/-2, post 3+/-3). Comparisons with baseline showed visual activation for both cocaine-cue conditions and limbic inhibition (accumbens, orbitofrontal, insula, cingulate) when subjects purposefully inhibited craving (p<0.001). Comparison between cocaine-cue conditions showed lower metabolism with cognitive inhibition in right orbitofrontal cortex and right accumbens (p<0.005), which was associated with right inferior frontal activation (r=-0.62, p<0.005). Decreases in metabolism in brain regions that process the predictive (nucleus accumbens) and motivational value (orbitofrontal cortex) of drug-conditioned stimuli were elicited by instruction to inhibit cue-induced craving. This suggests that cocaine abusers may retain some ability to inhibit craving and that strengthening fronto-accumbal regulation may be therapeutically beneficial in addiction.

Figure 1

Measures for self-reports of craving and for the Cocaine Craving Questionnaire and for the cardiovascular measures (heart rate, systolic and diastolic blood pressure) obtained prior to (lighter bars) and 30 minutes after initiation of the video (darker bars) or for the corresponding time period for the baseline. Craving measures were higher for cocaine-cue no inhibition (NI) than for cognitive inhibition (CI) or for the baseline condition whereas cardiovascular measures significantly increased for both NI and CI but not for baseline. *** p < 0.005, * p < 0.05.

Figure 2

A. SPM results for the comparison of the baseline (no video) versus the cocaine-cue video with no-inhibition (NI) conditions. B. SPM results for the comparison of the baseline versus the cocaine-cue video with cognitive inhibition (CI) conditions. The cocaine-cue video for both conditions induced relative increases in metabolism in occipital cortex. In addition the cocaine-cue video with CI induced relative decreases in right cingulate gyrus, right medial orbitofrontal cortex (OFC), right and left nucleus accumbens (NAcc), right posterior insula, and left fusiform gyrus (BA 28, BA 36). Significance corresponds to p < 0.001 not corrected, cluster > 100 voxels.

Figure 2

A. SPM results for the comparison of the baseline (no video) versus the cocaine-cue video with no-inhibition (NI) conditions. B. SPM results for the comparison of the baseline versus the cocaine-cue video with cognitive inhibition (CI) conditions. The cocaine-cue video for both conditions induced relative increases in metabolism in occipital cortex. In addition the cocaine-cue video with CI induced relative decreases in right cingulate gyrus, right medial orbitofrontal cortex (OFC), right and left nucleus accumbens (NAcc), right posterior insula, and left fusiform gyrus (BA 28, BA 36). Significance corresponds to p < 0.001 not corrected, cluster > 100 voxels.

Figure 3

A. SPM results for the comparison of the cocaine-cue video with cognitive inhibition (CI) versus the cocaine-cue video with no-inhibition (NI) and location of the ROI for the NAcc (red circles) and the medial OFC (blue oval) shown for the contralateral hemisphere. The SPM showed significantly lower metabolism in right NAcc and right medial OFC (mOFC) in CI when compared with NI, which was corroborated by ROI analysis. There were no areas where metabolism was higher for CI versus NI. Significance corresponds to p < 0.005 not corrected, cluster > 100 voxels. B. Regression slope for the differences in metabolism between CI versus NI in right NAcc and in right inferior frontal cortex BA 44 (r = −0.62, p < 0.005).

Figure 3

A. SPM results for the comparison of the cocaine-cue video with cognitive inhibition (CI) versus the cocaine-cue video with no-inhibition (NI) and location of the ROI for the NAcc (red circles) and the medial OFC (blue oval) shown for the contralateral hemisphere. The SPM showed significantly lower metabolism in right NAcc and right medial OFC (mOFC) in CI when compared with NI, which was corroborated by ROI analysis. There were no areas where metabolism was higher for CI versus NI. Significance corresponds to p < 0.005 not corrected, cluster > 100 voxels. B. Regression slope for the differences in metabolism between CI versus NI in right NAcc and in right inferior frontal cortex BA 44 (r = −0.62, p < 0.005).

Figure 4

A. SPM results for the voxel-wise correlations between metabolic activity during NI and the self-reports of craving (p < 0.005, cluster size > 100 voxels). B. SPM results for the voxel-wise correlation between metabolic activity during CI and the changes in the self-reports of craving (CI vs NI).

Figure 4

A. SPM results for the voxel-wise correlations between metabolic activity during NI and the self-reports of craving (p < 0.005, cluster size > 100 voxels). B. SPM results for the voxel-wise correlation between metabolic activity during CI and the changes in the self-reports of craving (CI vs NI).