Neuromodulation of delay discounting, the reflection effect, and cigarette consumption

Abstract

Cigarette smokers and substance users discount the value of delayed outcomes more steeply than non-users. Higher discounting rates are associated with relapse and poorer treatment outcomes. The left dorsolateral prefontal cortex (DLPFC) exerts an inhibitory influence on impulsive or seductive choices. Greater activity in the prefrontal cortex is associated with lower discounting rates. We hypothesized that increasing activity in the left DLPFC with high frequency repetitive transcranial magnetic stimulation (HF rTMS) would decrease delay discounting and decrease impulsive decision-making in a gambling task as well as decrease cigarette consumption, similar to other studies. In this single-blind, within-subjects design, smokers with no intention to quit (n = 47) and nonsmokers (n = 19) underwent three counterbalanced sessions of HF rTMS (20 Hz, 10 Hz, sham) delivered over the left DLPFC. Tasks were administered at baseline and after each stimulation session. Stimulation decreased discounting of monetary gains (F([3,250]) = 4.46, p < .01), but increased discounting of monetary losses (F([3,246]) = 4.30, p < .01), producing a reflection effect, normally absent in delay discounting. Stimulation had no effect on cigarette consumption. These findings provide new insights into cognitive processes involved with decision-making and cigarette consumption and suggest that like all medications for substance dependence, HF rTMS is likely to be most effective when paired with cognitive-behavioral interventions.

Copyright © 2013 Elsevier Inc. All rights reserved.

Figures

Figure 1

Mean discounting of monetary gains (A; n=66 participants, 325 observations) and losses (B; n=66 participants, 322 observations) plotted by condition and smoking status (nonsmokers = circles, smokers=triangles). The y-axis is the discounting rate expressed as the natural logarithm of k. Dotted error bars are 95% CIs among conditions. Solid error bars are 95% CIs for the differences between nonsmokers and smokers among conditions. A significant difference, at the .05 level, between nonsmokers and smokers is present when the solid bar does not cover both the circle and the triangle (corrected for 4 multiple comparisons with Holm’s Step-down method).

Figure 2

Mean discounting of monetary gains (A; n=66 participants, 325 observations) and losses (B; n=66 participants, 322 observations) plotted by condition. The y-axis is the discounting rate expressed as the natural logarithm of k. The horizontal line is the baseline mean. Dotted error bars are the 95% CIs among conditions. Solid error bars are 95% CIs for the a priori comparisons between sham, 10Hz, and 20Hz conditions and the baseline mean. A significant difference, at the .05 level, is present when the solid bar does not cross the mean baseline value (corrected for 3 multiple comparisons with Dunnett’s method).

Figure 3

Mean discounting of cigarette gains (A; n=47 smokers, 205 observations) and losses (B; n=46 smokers, 201 observations) plotted by condition. The y-axis is the discounting rate expressed as the natural logarithm of k. The horizontal line is baseline mean. Dotted error bars are the 95% CIs among conditions. Solid error bars are 95% CIs for the a priori comparisons between sham, 10Hz, and 20Hz conditions and the baseline mean. A significant difference, at the .05 level, is present when the solid bar does not cross the mean baseline value (corrected for 3 multiple comparisons with Dunnett’s method).

Figure 4

In the risky choices gambling task, mean points earned (A; n=66 participants, 325 observations) and time-to-complete (B; n=66 participants, 322 observations) plotted by condition and smoking status (nonsmokers=circles, smokers=triangles). Dotted error bars are 95% CIs among conditions. Solid error bars are 95% CIs for differences between nonsmokers and smokers among conditions. A significant difference, at the .05 level, between nonsmokers and smokers is present when the solid bar does not cover both the circle and the triangle (corrected for 4 multiple comparisons with Holm’s Step-down method).

Figure 5

The process of decision-making is influenced by neurobiological, environmental, and cognitive factors such as intention and/or motivation to quit and the perceived selection and organization of choices. Neuromodulation can influence decision-making by affecting the inhibitory influences of the dorsolateral prefrontal cortex thereby altering individual’s ability to select smaller sooner versus larger later rewards. This model is adapted from the work of Levasseur-Moreau et al. 2012 and Fecteau et al 2010.