All activations are reported at peak level and are in standard MNI stereotactic space

We hypothesized that amygdala engagement to aversive stimuli would show a positive correlation with accumulated lifetime ecstasy use; a negative correlation with SERT binding, as assessed with positron emission tomography imaging, in the amygdala; and a negative correlation with time since last use of ecstasy, suggesting recovery of amygdala response.Fourteen users of ecstasy and 12 non-using control subjects were recruited by flyers, advertisements posted on relevant websites, and word of mouth. Potential candidates were invited to a face to-face screening that involved assessment of history of alcohol, tobacco, and illicit drug use and Lifetime Drinking History, as well as screening of current and previous psychiatric symptoms using the Schedules for Clinical Assessment in Neuropsychiatry interview . All participants included in the present study, which focuses on fMRI investigations not presented previously, constitute of individuals who took part in a simultaneously conducted study with PET measurements of serotonergic markers—the PET results obtained in the larger sample have been published previously . To avoid acute drug effects in the ecstasy users, use of drugs was not allowed seven days before the PET and MRI scans, which were conducted on separate days at the Neurobiology Research Unit at Rigshospitalet and at the Danish Research Centre for Magnetic Resonance at Hvidovre Hospital, respectively. Abstinence was confirmed by urine screen on the PET and MRI days. Control individuals were excluded if they reported more than 15 lifetime exposures to cannabis or had any history of use of other illegal drugs,indoor garden table and urine screen on the day of the scan was carried out. Demographic and drug data are presented in Table 1.

No present or prior neurological or psychiatric disorders were allowed for any of the subjects, and all subjects had a normal neurological examination and were lifetime naïve to antidepressants and antipsychotics. The study was approved by the local Ethics Committee, Copenhagen and Frederiksberg, Denmark 01-124/04 with amendment 11-283038, and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.Regression analysis was used to investigate possible correlations between reaction times and log2 to the lifetime intake of ecstasy tablets or log2 to the number of days since the last use of ecstasy. Due to our previous findings that the effects of cumulative ecstasy lifetime intake and time since last use of ecstasy on SERT binding follow a log2distribution, we expected a log2-linear relationship between RT and lifetime intake and time since last use . Analysis and results that include lifetime intake and/or number of days since last use of ecstasy will in this manuscript refer to “log2 to” the number of accumulated lifetime doses/time since last use, but in order to improve readability, in the Results and Discussion sections, we have left out “log2 to” from the text. Differences in RT were assessed using a repeated measures analysis of variance model, with group as between-subject factor and emotion as within-subject factor . The significance level was set to α=0.05. Post hoc t-tests were carried out in order to test paired comparisons between facial expressions. The Greenhouse–Geisser method was used to correct for non-sphericity.The degree of serotonin depletion as reflected by cerebral SERT binding was assessed with PET and the selective SERT radioligand, 11C-DASB, as described in more detail by Erritzoe et al. . In short, PET data were acquired on a GE-Advance scanner as a dynamic 90-minute emission recording after intravenous injection of the radiotracer, 11C-DASB. There was maximum of 1.6 months between the 11C-DASB PET and the fMRI experiment ; approximately half of the group had PET before MRI, and vice versa. The vast majority was scanned within one to two weeks; only two were scanned with an interval of more than a month. For each participant, a mean SERT binding value from the amygdala was calculated. To confirm previous results obtained in the overlapping study sample , we used linear regression analysis to test whether the log2 to the accumulated lifetime intake of ecstasy tablets correlated with SERT binding from the amygdala.

Since effects of lifetime ecstasy intake on SERT become smaller with increasing abstinence from ecstasy , we controlled for this by also including the log2 to the number of days since the last use of ecstasy in the analysis. Also using a linear regression analysis, we tested whether the log2 to the number of days since the last use of ecstasy correlated with SERT binding from the amygdala . The possible difference between ecstasy users and controls in amygdala SERT binding was tested with an independent samples t-test.Functional data were preprocessed and analyzed using the statistical parametric mapping software package . Preprocessing included spatial realignment, co-registration to the anatomical image, segmentation, and normalization to the standard Montreal Neurological Institute template, and smoothing using a symmetric 6-mm Gaussian kernel. Subject-level models were constructed using five emotional face regressors , together with regressors modulating the events by their RT values. In addition, the subject-level models included a regressor for incorrect sex discrimination answers, and 24 nuisance regressors to correct for movement artifacts, including first- and secondorder movement parameters and spin history effects . T-contrasts comparing BOLD responses to emotional and neutral images were created for each participant. These maps were used in group-level models assessing: the main effects of emotional face processing across all participants, differences in BOLD responses between ecstasy users and control subjects, correlations between BOLD responses and the log2 to the accumulated lifetime ecstasy intake in ecstasy users, correlations between brain activity and regional SERT binding in the amygdala in ecstasy users and control subjects, and correlations between BOLD responses and the log2 to the number of days since the last use of ecstasy in ecstasy users. Drug use data used for correlations with MRI data and PET data were acquired in MRI and PET scan days, respectively. In and , one subject was excluded due to lack of precise information about the number of days of abstinence prior to the MRI investigation day, but the person was eligible for other analysis due to a negative urine sample. Our a priori region of interest was the amygdala.

At first, we therefore restricted the correction for multiple comparisons to the amygdala ROI, as defined by the SPM anatomy toolbox ; p-values are provided as p . We set the significance level for activated voxels at p<0.05 corrected for multiple comparisons using the family-wise error correction . The entry threshold was set to p<0.001 uncorrected with an extent threshold of five contiguous voxels. Second, a whole-brain analysis was performed. The significance level for activated voxels was set at p<0.05 corrected for multiple comparisons . The threshold was the same as in the ROI analysis.To test whether associations between ecstasy usage and BOLD response were mediated by SERT,microgreens grow rack a path analysis was used to decompose the total effect of MDMA usage into direct and indirect effects. The direct effect of ecstasy exposure on the mean BOLD response across the amygdala is the conditional effect adjusting for SERT binding. The indirect effect of MDMA on the mean BOLD response is the difference in the ecstasy effect between a model, where SERT BP is controlled for compared to when it is not. This difference in effect is equivalent to the product between the effect that ecstasy has on SERT and the effect that SERT has on BOLD response. Linearity assumptions were assessed graphically. Standard errors of the indirect effect were calculated by the delta method and were validated by comparison with 95% quantiles from a parametric bootstrap.To our knowledge, this is the first study to examine the effects of long-term ecstasy use on the neural responses to emotional face expressions. Relative to neutral face stimuli, main effects of emotional processing were found bilaterally in the amygdala, showing increased neural activity, especially in response to fearful and angry faces. This concurs well with a number of studies showing that viewing emotional faces, fearful faces in particular, activates the amygdala . While there was no ecstasy effect on task performance, ecstasy users did, as hypothesized, show higher amygdala activity with increased lifetime ecstasy use during angry face processing; that is, the more ecstasy tablets the ecstasy users had taken during their lifetime, the more activation they displayed in amygdala when watching angry faces. In the ecstasy user group, SERT binding correlated negatively with amygdala activity in response to angry faces. Non-significant statistical trends for activity during processing of angry and sad face processing suggested that amygdala activity waned with increasing time since the last intake of ecstasy. Neither the analyses of emotional expressions other than anger nor the whole-brain analysis revealed any significant results. Thus, our results support the hypothesis that long-term ecstasy use alters the neural basis of emotional face processing. This effect is dose-dependently related to lifetime consumption of ecstasy and appears to be reduced with increased time since last use. Interestingly, the linear relationship was consistently expressed for angry faces but not for other aversive facial expressions. This observation is in line with the results of Bedi et al. who found that acute MDMA intake alters the amygdala response to angry, but not fearful, facial expressions. The limited sample size of this study does, however, not allow us to conclude that there is not an effect of lifetime ecstasy intake on processing of other aversive facial expressions.

While acute MDMA intake has been shown to diminish amygdala activation , we found the opposite effect in long-term ecstasy users. This supports our hypothesis that long-term ecstasy users are in a chronic, albeit potentially reversible, serotonin-depleted state and therefore in accordance with studies showing that serotonin depletion, as induced by acute tryptophan depletion, leads to elevated amygdala activity when processing negative facial expressions . When including the lifetime amphetamine use in the model, the effect of the lifetime intake of ecstasy tablets on amygdala activity was no longer significant. This may be due to high correlation between ecstasy and amphetamine use . Since the lifetime amphetamine use in itself did not have a significant effect on amygdala activity during angry face processing, our interpretation of the results is that the effect of ecstasy use on emotional processing would be present also in the absence of amphetamine use. The present study was carried out on a sub-sample of our previous study sample of chronic ecstasy users , and we confirmed a negative correlation between SERT binding and accumulated ecstasy use. Hence, it could be speculated that our present fMRI results, showing a positive correlation between lifetime use of ecstasy tablets and left amygdala activity, was mediated by SERT density; that is, a larger lifetime intake of ecstasy tablets was associated with lower SERT binding levels , possibly leading to a higher degree of amygdala activation during angry face processing. In the ecstasy user group, SERT binding was indeed negatively correlated with amygdala reactivity to angry faces, which is in line with Rhodes et al. , showing a negative correlation in the left amygdala between SERT density and activity during emotional face processing. Post hoc mediation analysis did, however, not support the mediation hypothesis, although these results need to be interpreted with caution given the small sample size and hence the low statistical power. In short, our study suggests that there are functional consequences of a chronically depleted serotonin system as indexed by lowered SERT. Of note, an augmented amygdala response to angry faces has also been observed in mood disorders and could within a population with reduced serotonergic tone represent a sub-clinical vulnerability marker for such conditions. In line with several other studies , we have recently reported that recovery of subcortical—but not cortical—SERT availability takes place after termination of ecstasy use. Importantly, here, we found trends showing that days of abstinence from ecstasy correlated negatively with left amygdala activity during angry face processing and with right amygdala activity during sad face processing. Since lifetime use of ecstasy tablets correlated positively with amygdala activity during angry face processing, the trend toward a negative correlation between days of abstinence from ecstasy and amygdala activity during angry face processing might be a potential sign of functional reversibility.There are limitations to the current study. Because of the cross-sectional nature of our study, it cannot be ruled out that the exaggerated amygdala response to angry faces and/or the low cerebral SERT among heavy ecstasy users represents preexisting traits associated with an increased preference for the use of ecstasy.