Conflict monitoring and emotional processing in 3,4-methylenedioxymethamphetamine (MDMA) and methamphetamine users - A comparative neurophysiological study

Opitz and colleagues frame the study around rising MDMA and methamphetamine (METH) use and the need to understand how chronic use of these substituted amphetamines affects impulse control in social–affective contexts. Earlier research indicates both overlapping and substance-specific impairments in executive functions among stimulant users: chronic METH is more strongly associated with dopaminergic dysfunction, addictive potential, impaired social cognition and aggression, whereas chronic MDMA is more associated with serotonergic alterations and, acutely, prosocial effects. Both groups, however, show elevated impulsivity and deficits in inhibitory control on some tasks. The monoamines dopamine, serotonin and noradrenaline are highlighted as likely mediators of these effects, but noradrenergic consequences of chronic use remain under-studied. This study set out to compare chronic but recently abstinent MDMA and METH users with matched healthy controls using an emotional face–word Stroop paradigm while recording event‑related potentials (ERPs). The investigators aimed to test whether conflict control differs between substance groups depending on emotional valence (anger versus happiness), and to identify neurophysiological correlates, in particular later ERP components associated with conflict monitoring and response selection (the conflict slow potential, CSP, and the P3). They hypothesised that METH users would show greater impairments in angry contexts than MDMA users, and expected substance-related modulations in the CSP and P3 rather than in early perceptual components such as P1 or N1.

This was a cross-sectional neurophysiological comparison of chronic but recently abstinent MDMA users (n = 42), METH users (n = 38) and matched healthy controls (MDMA controls n = 42; METH controls n = 41). Participants were recruited at two sites (Zurich for MDMA groups; Dresden for METH groups) and were aged 18–45 years. Inclusion for user groups required at least 25 lifetime uses of the primary substance and recent use within the past 6 (MDMA) or 12 (METH) months; controls were amphetamine‑naïve or had very limited prior illicit use. Full eligibility criteria are reported in the supplement. Exclusion criteria included severe somatic, neurological or psychiatric disorders likely to impair task performance, recent illicit substance intake within 48 hours (urine screen) and pregnancy or breast‑feeding. The study received ethical approval and all participants gave written informed consent. Clinical and substance use assessments included structured diagnostic interviews (SCID‑5 for substance use disorders), questionnaires for depressive symptoms (CESD‑R), ADHD symptoms (ADHD‑SR), impulsivity (BIS) and trait anger (STAXI‑2), plus a standard interview for lifetime psychotropic drug consumption. Objective measures included multi‑drug urine screening at appointment and hair sampling for long‑term exposure analysed by LC‑MS/MS. Craving was assessed for METH users with an adapted brief craving instrument; it was not assessed in the MDMA group given typically low craving reports. The experimental task was an adapted emotional face–word Stroop. Each stimulus combined a greyscale face (happy or angry) with a red word across the face (German words for anger or happiness). Trials were congruent (matching face and word) or incongruent (mismatching), and participants categorised the facial expression while ignoring the word. Sixteen stimuli (four faces × four word/face combinations) were presented in pseudo‑random order; a stimulus lasted 450 ms, response windows and intertrial jittering are described in the methods. EEG was recorded with 60 electrodes (500 Hz sampling), referenced to Fpz, impedances below 5 kΩ. Preprocessing used Automagic and EEGLAB, with segmentation of correct trials from −2000 to +2000 ms around stimulus onset, band filtering above 20 Hz removed, automated artifact rejection, current source density transformation and baseline correction (−200 to 0 ms). ERP components were quantified by inspection and topography: P1 (P7/P8, 100–130 ms), N1 (P7/P8, 155–180 ms), N2 (Cz/CPz, 205–300 ms), P3 (P3/P4, 355–410 ms) and CSP (CPz, 490–630 ms). Behavioural analyses used accuracy and reaction time (RT) restricted to correct responses within 200–1200 ms. Separate mixed ANOVAs were run for behavioural and ERP amplitudes. Between‑subject factors were subsample (MDMA site vs METH site) and group (substituted amphetamine users [MDMA + METH] vs controls), with within‑subject factors congruency (congruent vs incongruent), valence (happy vs angry) and electrode where applicable. Post hoc tests were Bonferroni‑corrected; Bayesian analyses were used to interpret non‑significant trends (p ≤ 0.10). A sensitivity analysis with G*Power suggested a target sample of n = 200 to detect effect sizes f = 0.2 at 95% power; the final analysed sample was n = 163.

Sample and data inclusion: The final analysed sample comprised 163 participants: 42 MDMA users (12 meeting MDMA‑related SUD criteria), 42 matched MDMA controls, 38 METH users (all meeting METH‑related SUD criteria), and 41 matched METH controls. Positive urine results occurred in some participants in both subsamples; these individuals were retained and additional analyses of urine effects are reported in the supplement. Behavioural outcomes: There were no significant interactions between subsample and group, indicating that MDMA and METH users did not differ from each other on the behavioural measures. Across the full substituted amphetamine user group versus controls, users were generally less accurate (86.7% ± 0.6) than controls (90.7% ± 0.6), and the MDMA subsample showed slightly higher overall accuracy than the METH subsample (89.7% ± 0.6 vs 87.7% ± 0.7). A group × valence interaction indicated that users were less accurate in angry trials (85.7% ± 0.9) than in happy trials (87.9% ± 0.7), whereas controls showed no valence difference; the valence effect (happy minus angry) was larger in users (2.1% ± 0.5) than in controls (0.6% ± 0.4). For RTs, substituted amphetamine users responded faster overall (497 ms ± 10.1) than controls (533 ms ± 9.9). Both groups exhibited the expected Stroop effect (slower responses for incongruent than congruent trials), but the Stroop effect was smaller in users (4 ms ± 1.4) than in controls (10 ms ± 1.4). When depressive symptoms were included as a covariate, the group × valence interaction for accuracy was no longer significant. Within MDMA users, higher MDMA hair concentrations and greater weekly MDMA use over the prior 12 months were associated with slower responses on some trial types. Neurophysiological outcomes: The investigators focused on ERP effects that aligned with behavioural findings. For the early P1 (P7/P8) there was a group × congruency × electrode interaction; in substituted amphetamine users the Stroop effect at P8 exceeded that at P7, but this did not map onto differential behavioural performance between MDMA and METH users. N1 showed non‑significant trends and Bayesian analyses favoured the null for some electrode/subsample interactions. The N2 results in the extracted text are not fully reported and the details are unclear from the provided extraction. The CSP (CPz, 490–630 ms) was larger in the substituted amphetamine users (6.2 μV/m² ± 0.5) than in controls (3.6 μV/m² ± 0.5), and there was also a subsample effect described (details partially obscured in the extraction). For the P3 (P3/P4, 355–410 ms) a significant group × valence × electrode interaction emerged. At electrode P4, substituted amphetamine users showed larger P3 amplitudes for angry (22.1 μV/m² ± 1.2) than happy faces (21.1 μV/m² ± 1.1), whereas controls did not; the valence effect at P4 was larger in users (1.0 μV/m² ± 0.4) than in controls (0.1 μV/m² ± 0.3). There was a non‑significant trend for an interaction between group, congruency and electrode for P3; exploratory analyses at P4 indicated users had larger P3 amplitudes in congruent than incongruent trials (22.0 μV/m² ± 1.2 vs 21.2 μV/m² ± 1.1), while controls did not, and the Stroop effect in P3 trended to be slightly larger in users (0.8 μV/m² ± 0.2) than in controls (0.2 μV/m² ± 0.2). No other group‑related ERP effects corresponding to behavioural results reached significance in the extracted text.

Opitz and colleagues interpret their findings as showing substance‑independent alterations in cognitive–emotional processing among chronic users of substituted amphetamines: contrary to their hypothesis, MDMA and METH users did not differ from each other, but both groups differed from controls. Behaviourally, users exhibited a reduced modulation of performance by cognitive–emotional conflict (a smaller emotional Stroop effect) and selective deficits in processing anger (lower accuracy for angry versus happy faces). Neurophysiologically, these behavioural patterns were accompanied by stronger modulations in later ERP components, notably larger CSP amplitudes in users and P3 modulations reflecting sensitivity to angry stimuli and to congruency at electrode P4. The authors link the P3 to stimulus–response mapping and decision processes, and suggest that altered catecholaminergic function (particularly noradrenaline, which both substances acutely release) may underlie the similar effects across MDMA and METH users, since dopaminergic and serotonergic differences alone do not readily explain the parallel findings. The discussion notes several caveats acknowledged by the investigators. First, the emotional Stroop design used faces as targets and words as distractors; because faces are highly salient and innate stimuli, this configuration may have reduced the magnitude of conflict and could explain why users showed smaller behavioural Stroop effects. Second, most participants were active users who only temporarily abstained for the study, so post‑acute or tolerance‑related effects might have influenced results; differences between short‑ and long‑term abstinence remain unresolved. Third, the two user groups differed in clinical severity (all METH users met SUD criteria while fewer MDMA users did), and the sample was heterogeneous; the control groups were not strictly substance‑naïve for recreational substances such as cannabis, which increases ecological validity but may introduce confounds. The authors report they evaluated a wide range of potential confounders and found no effects on the main interactions except for depressive symptoms affecting one valence interaction. Finally, they call for further work to clarify the role of noradrenergic dysfunction after chronic substituted amphetamine use, to examine whether different task designs (for example emotional words as targets) yield larger conflict effects, and to explore possible non‑linear relationships between behavioural and neurophysiological changes across stages of use and abstinence.

This comparative study is presented as the first to directly contrast chronic MDMA and chronic METH users on cognitive–emotional conflict control and its ERP correlates. The investigators conclude that MDMA and METH users do not differ substantially from each other in conflict monitoring or emotional processing, but both groups differ from healthy controls: they show a reduced behavioural modulation by cognitive–emotional conflict and selective impairments in processing anger. Stronger P3 modulations underpinned these effects and are interpreted as markers of altered stimulus–response mapping and decision making in substituted amphetamine users. The authors emphasise that the precise neurobiological mechanisms remain uncertain, and highlight the noradrenergic system as an important target for future research.