## Introduction

### Study procedure

Participants completed an initial screening and orientation visit during which they provided informed consent and were familiarized with laboratory procedures and study protocol. Participants completed drug administration sessions at UC and then attended a separate fMRI visit at UIC 1–3 weeks later. Participants were asked to abstain from drugs, including alcohol, for 24 h prior to each visit, which was verified by self-report, breath alcohol, and urine screens.

During the four drug administration sessions participants received d-amphetamine (20 mg) and placebo in alternating order. D-amphetamine and placebo were each administered twice to minimize the influence of day-to-day variability [37]. Sessions took place from 09:00 to 13:00 h in comfortable, living-room-like rooms, and were separated by at least 48 h. Participants were tested individually. Participants were instructed not to eat after midnight before each session and were given a light snack upon arrival to the lab. To minimize drug expectancies they were told they could receive one of the following: stimulant, sedative, or placebo. On each session subjects first completed pre-drug ARCI, and had blood pressure and heart rate measured. At 09:20 h, d-amphetamine (20 mg oral; Desoxyn tablets with dextrose filler in size 00 opaque capsules) or placebo (dextrose only) was administered under double-blind conditions. Participants completed the ARCI every 30 min following capsule administration. Sessions ended at 13:00 h, after confirmation that blood pressure and heart rate had returned to baseline.

The fMRI session took place on a separate day, at least 1 week after the last drug administration session. Participants were tested for recent drug use and screened for MRI safety before completing the GRT during the scan. After completing all sessions, participants were debriefed and compensated for their participation.

### fMRI data acquisition

Functional MRI data were collected using a 3 T GE magnetic resonance scanner at the UIC Center for Magnetic Resonance Research. Functional images were acquired using a gradient-echo echo-planar images (2 s TR, 25 ms TE, 82° flip, 64 × 64 matrix, 200 mm FOV, 3 mm slice thickness, 0 mm gap, with 44 axial slices).

### fMRI data analyses

Imaging data were inspected and individuals with >2 mm displacement in any direction were not included in the analysis. The remaining subjects met criteria for high quality and scan stability. Preprocessing of fMRI data were conducted using Statistical Parametric Mapping software (SPM12, Wellcome Department of Imaging Neuro-Science, London, UK). Images were spatially realigned, slice-time corrected, warped to Montreal Neurological Institute (MNI) space using the participant’s mean functional image, resampled to 2 mm3 voxels, and smoothed (4 mm3 kernel). The general linear model was applied to the time series, convolved with the canonical hemodynamic response function and with a 128-s high-pass filter. Condition effects were modeled with event-related regressors representing the occurrence of anticipation for Win or Loss. Effects were estimated at each voxel, and for each subject. Individual contrast maps for Win trials (>Loss trials) were created for each person. Individual motion parameter files were included in the first levels models as regressors-of-no-interest.

## Results

Reward anticipation relative to loss (Win > Loss) significantly activated a large contiguous cluster of frontal and mesolimbic reward regions, including bilateral nucleus accumbens, caudate, and putamen (peak MNI [−6, 16, −10], k = 3376 voxels, Z = 7.30, p < .05, corrected). Mesolimbic activation during reward is illustrated in Fig. 2. Significant peak clusters within the Neurosynth mask are shown in Table 2.

### Subjective effects of amphetamine

As previously reported in this dataset [40, 41], d-amphetamine significantly increased ratings of euphoria [mean amphetamine score = 4.6 (sd = 4.7), mean placebo score = −0.1 (sd = 2.3); paired samples t-test: t(60) = 8.1, p < 0.001] and stimulation [mean amphetamine score = 3.5 (sd = 2.9), mean placebo score = −0.1(sd = 1.6), paired samples t-test: t(60) = 9.2, p < 0.001], relative to placebo. Individuals differed in their responses to d-amphetamine, with peak difference scores (d-amphetamine minus placebo) ranging from −5.5 to 17 for euphoria and −4 to 12.5 for stimulation.

### Association between neural activation during reward and subjective response to amphetamine

More BOLD activation during Win > Loss anticipation was associated with greater drug-induced euphoria (peak change difference ARCI-MBG score) in a large contiguous cluster that included the bilateral putamen and caudate (right: MNI peak [16, 6, 6], k = 266 voxels, Z = 3.62, p < .05, corrected; left: MNI peak [−20, 12, 2], k = 203 voxels, Z = 3.45, p < .05, corrected) (Fig. 3)Footnote 1,Footnote 2. Drug-induced stimulation (peak change difference ARCI-A score) was not associated with BOLD activation within the mask during Win > Loss anticipation.

### Exploratory analyses

To better understand whether Win events or Loss events contributed to the association between Win > Loss anticipation and drug-induced euphoria, we examined if Win > Fixation and Loss > Fixation were separately related to drug-induced euphoria. For these exploratory analyses, individual drug-induced ARCI-MBG scores were entered as a regressor of interest for the Win > Fixation anticipation model and for the Loss > Fixation anticipation model. Neither model was significant, indicating that the original findings were due to the relative difference between Wins vs. Losses.

To determine whether any of the subjects’ responses were related to prior drug use, we examined the relationships among drug-induced euphoria, parameter estimates/β-weights of BOLD peak activation during Win > Loss anticipation, and current and lifetime substance use measures. Neither drug-induced euphoria nor BOLD peak activation during Win > Loss anticipation were related to current (past month) alcohol, cigarette, caffeine, or marijuana use (Spearman Rho correlations p-values >.05), or to whether participants had ever (vs never) used marijuana, hallucinogens, stimulants, opiates, MDMA, or sedatives in their lifetime (t-test p-values >.05).

## Discussion

One predictor of risk for drug use and abuse is sensitivity to the rewarding effects of drugs, and perhaps also sensitivity to rewards in general. Here, we studied brain responses to a monetary reward in relation to sensitivity to the subjective rewarding effects of amphetamine in healthy young adults. We found novel evidence that greater activation in mesolimbic reward regions during anticipation of monetary reward in a drug-free state was associated with greater euphoria after d-amphetamine administration. Specifically, individuals who reported greater subjective euphoria after amphetamine during the behavioral phase of the study exhibited greater neural activation during reward anticipation in the bilateral caudate and right putamen. Notably, the relative difference between Win and Loss events, rather than Win events or Loss events separately (although see limitations of these analyses below), was significantly related to subjective drug response. This is the first study to show a relationship between neural correlates of monetary reward (in the absence of drug) and sensitivity to the subjective rewarding effects of a stimulant drug in humans.

The finding that activation in the caudate and putamen was related to subjective drug reward is consistent with what is known about both brain reward circuitry and risk factors involved in drug use and abuse. The striatum, including the caudate and the putamen, are involved in many aspects of reward evaluation and incentive-based learning (see ref. [20]), and the striatum has been linked to both the initiation and maintenance of substance use behaviors [13]. A recent study found that greater dopamine release was associated with both an earlier age of first drunkenness and greater neural activation to monetary reward in high-risk youth (Weiland et al., 2017), supporting the notion that more striatal activation to non-drug, monetary reward may be a neural profile for vulnerability for drug use problems. Similarly, subjective response to drug reward has been shown to be an important risk factor for several indicators of substance use problems, including choosing to take the drug again, progression of substance use, and development of SUD [2,3,4,5,6,7,8,9,10,11]. Therefore, the current study establishes an important link between activation in the striatum, a known neural mechanism involved in both reward and substance abuse, and euphoric response to drug reward among healthy, non-dependent young adults.

In addition, the current study bridges two separate, but related, literatures on neural response to non-drug reward and neural response to acute drug administration. It extends observations that individuals who score high on personality measures of reward sensitivity show increased neural activation in brain reward circuitry during non-drug reward [21, 22], by showing that greater subjective reward to a drug challenge is related to greater neural activation in the striatum during non-drug reward. Further, previous studies showed that acute amphetamine and methylphenidate increased dopamine release in the striatum [14,15,16,17,18,19], while our study showed that greater subjective rewarding effects of amphetamine are also related to increased activation in the striatum during non-drug reward (when individuals are not acutely intoxicated). Taken together, these findings demonstrate a similar pattern of increased activation in the striatum in response to drug and non-drug reward.

In this sample of relatively light drug users, previous substance use was not related to drug-induced euphoria or to BOLD peak activation during Win > Loss anticipation. This suggests that the relationships reflect stable, individual differences that are not the result of prior drug exposure. The findings suggest that a proportion of relatively light drug users may be at elevated risk for future onset of drug use problems. These relationships may be even stronger among individuals who were not tested here because they had already developed problems. Alternatively, these individuals may indeed be at risk, but were protected by other factors that prevent the development of excessive drug use. Future prospective studies will be needed to determine whether either acute drug responses or neural responses to reward predict development of problem use.

These findings have important implications. Greater neural activation to non-drug, monetary reward may represent a risk factor for SUD among young adults. Greater neural activation during reward is linked to greater hedonia and subjective pleasure from drug rewards and this, in turn could lead to more drug-seeking. Therefore, we may be able to use neural activation to monetary reward to detect this broader trait early on, before substance use initiation or progression. This may be particularly useful in adolescent populations, in whom it is not possible to study responses to drugs.

Despite the strengths of the study (e.g., a relatively large sample size and separate measures of drug and non-drug reward), the study also had limitations. The participants were healthy young adults, and it is not clear that the findings would be generalizable to individuals who are at high risk, such as those with a family history of drug use. Another limitation was that the behavioral drug sessions always preceded the fMRI, raising the possibility that exposure to the drug changed brain responses to monetary reward. However, it is unlikely that single low doses of d-amphetamine influenced brain activity 1–3 weeks later. Further, we tested only one dose of amphetamine, and full dose-response data with amphetamine would provide a more complete picture of the pharmacological profile. There were also limitations related to the task used during the imaging session. Because the GRT task does not have a neutral anticipation condition in which subjects expect no money to be won or lost, there is no true control condition to compare against Win and Loss anticipation events. Our attempt to examine the separate effects of Win and Loss events using Fixation is not a comparable “control” condition. It will important for future studies to determine the respective contributions of Win and Loss anticipation using a ‘neutral’ condition. The GRT also did not allow us to examine neural activation during reward receipt, which would be an interesting and relevant construct. Our study did not assess family history of drug use, which might influence activation to reward (see ref. [41]). Although, we typically detect a low prevalence of family history of alcoholism in drug challenge studies in our laboratory, it would be of interest to examine family history of SUD in relation to drug-induced euphoria and neural reward activation. Finally, our study is cross-sectional and we did not measure substance use over time. Future and ongoing studies are needed to replicate the current study in longitudinal samples.

Our study provides novel evidence of a relationship between neural correlates of monetary reward in the absence of drug and sensitivity to the subjective rewarding effects of amphetamine in humans. Specifically, we found that greater activation in mesolimbic reward regions, including the bilateral caudate and putamen, during non-drug reward anticipation was associated with more drug-induced euphoria. These findings extend our understanding of the neural correlates of a well-known risk factor for SUD, sensitivity of drug-induced euphoria, and provide evidence that individual differences to non-drug reward and subjective drug reward are linked and could represent a profile of vulnerability for SUD.