Reduced Glutamate Turnover in the Putamen Is Linked With Automatic Habits in Human Cocaine Addiction

Abstract

Background: The balance between goal-directed behavior and habits has been hypothesized to be biased toward the latter in individuals with cocaine use disorder (CUD), suggesting possible neurochemical changes in the putamen, which may contribute to their compulsive behavior.

Methods: We assessed habitual behavior in 48 patients with CUD and 42 healthy control participants using a contingency degradation paradigm and the Creature of Habit Scale. In a subgroup of this sample (CUD: n = 21; control participants: n = 22), we also measured glutamate and glutamine concentrations in the left putamen using ultra-high-field (7T) magnetic resonance spectroscopy. We hypothesized that increased habitual tendencies in patients with CUD would be associated with abnormal glutamatergic metabolites in the putamen.

Results: Compared with their non-drug-using peers, patients with CUD exhibited greater habitual tendencies during contingency degradation, which correlated with increased levels of self-reported daily habits. We further identified a significant reduction in glutamate concentration and glutamate turnover (glutamate-to-glutamine ratio) in the putamen in patients with CUD, which was significantly related to the level of self-reported daily habits.

Conclusions: Patients with CUD exhibit enhanced habitual behavior, as assessed both by questionnaire and by a laboratory paradigm of contingency degradation. This automatic habitual tendency is related to a reduced glutamate turnover in the putamen, suggesting a dysregulation of habits caused by chronic cocaine use.

Keywords: 7T magnetic resonance spectroscopy; Cocaine use disorder; Contingency degradation; Creature of Habit Scale; Glutamate-to-glutamine ratio.

Figure 1

(A) Experimental design of a varying degradation of contingency. Participants are shown a white vase on the computer screen, which fills with flowers when the subject presses a dedicated button on the keyboard. Button presses are always associated with a 60% probability of receiving 20 pence. Depending on the experimental condition, not pressing the button was also associated with receiving 20 pence. (B) Subjective value of monetary reward. Participants rated on a visual analog scale (0 = never, 100 = always) how likely it was that they would pick up a 20-pence coin lying on the street. The groups did not differ in terms of their subjective value of 20 pence. (C) Response rate (number of button presses per second) across the three different task conditions (nondegraded, partially degraded, and fully degraded). Control participants demonstrated sensitivity to changes in the action-outcome contingencies as they significantly reduced their response rate in the fully degraded condition, whereas patients with cocaine use disorder did not. Their response rates between the partially and fully degraded conditions did not differ. (D) After each condition, participants were asked to indicate how likely they think it is that pressing the button wins them money. While control participants were aware about the changes in the action-outcome relationship and adjusted their behavior accordingly, patients with cocaine use disorder continued to believe in the effectiveness of their actions to receive a financial reward and so they continued responding accordingly. Error bars denote standard error of the mean. ∗p < .05. A, action; ns, not significant; O, outcome; P, probability.

Figure 2

We calculated a ratio score of pairs of nondegraded and partially or fully degraded blocks to estimate the balance between the goal-directed and habitual systems, as described by Vaghi et al. (40). For each pair, the number of responses in the nondegraded block was divided by the sum of responses in both the contingent and the degraded blocks. Values close to 1 reflect high sensitivity to the reinforcement contingency (indicative of goal-directed tendencies) and values close to 0.5 reflect a similar response patterns of the contingent and degraded conditions (indicative of habitual tendencies). (A) In the fully degraded condition, patients with cocaine use disorder showed increased habitual responses the longer they have been using cocaine (r = −3.1, p < .05). (B) The longer the duration of cocaine use, the more automatic habits patients with cocaine use disorder reported on the COHS (r = .33, p < .05). COHS, Creature of Habit Scale.

Figure 3

Goal-to-habit ratio score for the condition in which the action-outcome relationship was fully degraded. The score was computed from the number of responses during the nondegraded condition divided by the sum of responses in the fully degraded and nondegraded conditions. Values close to 1 suggest high sensitivity to the reinforcement contingency (indicative of the goal-directed system) and values close to 0.5 suggest that the response pattern does not differentiate between nondegraded and degraded conditions (indicative of the habit system). (A) When the action-outcome contingency was fully degraded, patients with cocaine use disorder exhibited a less goal-directed response tendency than healthy control participants. (B) COHS automaticity scores (i.e., the degree to which behavioral responses are involuntarily triggered by specific contexts) were associated with more habitual response tendencies on fully degraded trials (r = −.23, p < .05; control participants: r = .12, p = .445, cocaine use disorder: r = .26, p = .069). COHS, Creature of Habit Scale.

Figure 4

(A) Voxel placement and representative spectra from the left putamen. (B) Mean GABA (t₃₉ = −0.66, p = .514; d = 0.21), glutamine (t₃₉ = 0.79, p = .437; d = 0.24), and glutamate (t₃₉ = 2.14, p = .039; d = 0.64) concentration in the left putamen, as well as the glutamate-to-glutamine ratio (t₃₉ = 2.11, p = .33; d = 0.69) in both healthy control participants and patients with cocaine use disorder (error bars denote standard error of the mean; ∗p < .05). (C) Self-reported Creature of Habit traits (COHS automaticity levels) are not associated with the glutamate-to-glutamine ratio in the left putamen in control participants (r = .11, p > .5). (D) In patients with cocaine use disorder, however, there is a strong relationship between COHS automaticity and the glutamate-to-glutamine ratio measured in the left putamen (r = .49, p < .05). COHS, Creature of Habit Scale; GABA, gamma-aminobutyric acid.