Individuals with Methamphetamine Use Disorder Show Reduced Directed Exploration and Learning Rates Independent of an Aversive Interoceptive State Induction
===========================================================================================================================================================

* Carter M. Goldman
* Toru Takahashi
* Claire A. Lavalley
* Ning Li
* Samuel Taylor
* Anne E. Chuning
* Rowan Hodson
* Jennifer L. Stewart
* Robert C. Wilson
* Sahib S. Khalsa
* Martin P. Paulus
* Ryan Smith

## Abstract

Methamphetamine Use Disorder (MUD) is associated with substantially reduced quality of life. Yet, decisions to use persist, due in part to avoidance of anticipated withdrawal states. However, the specific cognitive mechanisms underlying this decision process, and possible modulatory effects of aversive states, remain unclear. Here, 56 individuals with MUD and 58 healthy comparisons (HCs) performed a decision task, both with and without an aversive interoceptive state induction. Computational modeling measured the tendency to test beliefs about uncertain outcomes (directed exploration) and the ability to update beliefs in response to outcomes (learning rates). Compared to HCs, those with MUD exhibited less directed exploration and slower learning rates, but these differences were not affected by aversive state induction. These results suggest novel, state-independent computational mechanisms whereby individuals with MUD may have difficulties in testing beliefs about the tolerability of abstinence and in adjusting behavior in response to consequences of continued use.

## Introduction

Methamphetamine Use Disorder (MUD) is characterized by biological, cognitive, and behavioral changes that can be detrimental at both the individual and societal level. Though outcomes vary widely, common psychological consequences include psychosis, suicidality, hostility, anxiety, depression, and psychomotor dysfunction1,2. Despite its growing prevalence in the United States and worldwide3, cognitive mechanisms governing the onset, maintenance, and recurrence of MUD remain unclear.

One means by which MUD may be maintained depends on the expected negative outcomes of abstinence and associated withdrawal states, which can motivate avoidance when combined with negative reinforcement processes4. In particular, methamphetamine use may attenuate symptoms of depression or somatic anxiety that are brought on or exacerbated by withdrawal. Deficits in interoceptive processing may further contribute to maladaptive behavior, as previous work has shown that individuals with MUD exhibit attenuated neural responses (e.g., Insula, Anterior Cingulate Cortex) to aversive somatic states5. Countering withdrawal avoidance instead requires that individuals “test out” abstinence as a means of learning whether they are capable of enduring its short-term consequences to improve longer-term quality of life. In computational neuroscience, the abstract structure of this decision problem is captured by so-called “explore-exploit” decision tasks6,7. In these tasks, one can either exploit current (limited) knowledge to maximize short-term reward, or one can first test the outcomes of different options (explore) to make better informed choices in the long-term. Importantly, there are different exploratory strategies, which depend on distinct computational processes, and some may be more clinically relevant than others8–10. *Directed* exploration (DE), for example, requires keeping track of one’s relative uncertainty about different action outcomes, and then choosing the action for which one has the greatest uncertainty (i.e., as this leads to the most information gain). In contrast, so-called *random* exploration (RE) requires keeping track of one’s total uncertainty across action options, where greater total uncertainty should increase the chance of selecting options that do not currently appear most rewarding (i.e., as one might learn that past experiences were misleading). In the example of withdrawal avoidance mentioned above, DE may be more relevant, as the individual must recognize that they have greater uncertainty about the outcomes of abstinence than about those of continued use. Given previous work showing a positive relationship between DE and cognitive reflectiveness10, individuals with methamphetamine use disorder (iMUDs) might also be less likely to implement DE, as less reflective tendencies associated with impulsivity have been observed in this population and positively relate to severity of use11.

Several studies support this in suggesting that iMUDs display lower levels of exploration and altered belief-updating. One study showed that participants with Amphetamine Use Disorder engaged in less information-seeking than healthy controls in a decision-making task that had no cost associated with exploration12. This difference may be partly attributable to effects of the drug itself, given that chronic amphetamine use has been shown to deplete intracellular dopamine13, and that lower tonic dopamine levels have been linked to lower exploratory behavior in individuals with substance use disorders14. Furthermore, a longitudinal study of iMUDs found that participants who decreased methamphetamine use over a period of six weeks showed higher levels of DE by the end of that period15, supporting the idea that methamphetamine use may affect exploratory behavior. In addition to lower levels of exploration, iMUDs have been shown to engage in maladaptive belief-updating, which is often operationalized in terms of altered learning rates within computational models16–21. This overall pattern of altered sensitivity to, and learning from, choice outcomes may help explain continued use and high relapse rates despite reduced quality of life22.

Thus, it appears plausible that methamphetamine use is linked to reduced exploration, and that this might in turn deter abstinence. However, while methamphetamine use is motivated by avoidance of the aversive interoceptive states associated with abstinence4, the effect of these states on decision-making itself remains unclear. In this study, we therefore had two main aims. First, we aimed to test whether, compared to healthy comparisons (HCs), iMUDs would show reduced DE and altered learning, as suggested by the literature reviewed above. Second, we sought to test how an aversive interoceptive state (i.e., a somatic anxiety induction) may affect these mechanisms. To do so, we fit a computational model with both exploration and learning rate parameters to choice behavior on an established explore-exploit decision task in iMUDs and HCs – both with and without a breathing-based aversive interoceptive state induction. We then tested for both group differences and effects of induced somatic anxiety. As a supplementary aim, we also sought to replicate exploratory results demonstrating a relationship between DE and cognitive reflectiveness, and extend this to iMUDs10.

## Results

To acquire a comprehensive clinical phenotype for iMUDs, participants completed several cognitive and clinical scales (see **Methods**). Compared to HCs, iMUDs showed elevated symptoms of anxiety and depression, higher impulsivity, reduced cognitive performance (working memory), and lower cognitive reflectiveness (see **Table 1**).

View this table:
[Table 1.](http://medrxiv.org/content/early/2024/05/21/2024.05.17.24307491/T1)

Table 1. Descriptive Characteristics of HCs and iMUDs.

### The Aversive State Induction Successfully Increased Anxiety During Task Performance Across All Participants

Participants completed the Horizon Task8 twice, where one of the runs included a breathing resistance to induce state anxiety. This task requires participants to repeatedly choose between two options to maximize points when given either equal or unequal information about previous reward outcomes from each option (i.e., either two outcomes per side, or one vs. three; based on four initial forced choices). A greater propensity to choose the more uncertain option on subsequent free choices reflects an *information bonus*, while the propensity to choose the less rewarding option reflects *decision noise*. These propensities are moderated by both learning rates from forced-choice outcomes and expected number of future choices (one vs. six; H1 vs. H6 conditions). Namely, information bonus and decision noise should increase from H1 to H6 – where this increase reflects DE and RE, respectively – as exploration could guide future choice in H6 only (see **Methods**).

In a linear mixed-effects model (LME) predicting self-reported anxiety based on group, resistance condition (baseline, task run without resistance, task run with resistance), and their interaction, all effects were significant (*ps*<.001; see **Supplemental Table S3** for full model results), indicating the breathing resistance successfully induced anxiety during the task and this effect was magnified for iMUDs (see **Figure 1**).

![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/05/21/2024.05.17.24307491/F1.medium.gif)

[Figure 1.](http://medrxiv.org/content/early/2024/05/21/2024.05.17.24307491/F1)

Figure 1. A)
The silicon mask used during both runs of the Horizon Task. **B)** An example resistor attached to the mask via a plastic tube (not depicted), causing participants to experience resistance during inhalation. A resistance of 40 cmH20/L/sec was used for one of the runs of the Horizon Task to induce somatic anxiety. The other run was completed without breathing resistance. **C)** Participants’ self-reported anxiety scores at baseline, during the task run without breathing resistance, and during the task run with breathing resistance. **D)** Horizon Task: Participants first observed outcomes of four forced choices before they were allowed to make either one or six free choices between options to maximize the total number of points received. Games with one or six free choices are referred to as Horizon 1 (H1) and Horizon 6 (H6) games, respectively. The forced choices in each game were either equally informative (two forced choices for each slot machine) or unequally informative (three forced choices for one slot machine and one for the other).

### Individuals with MUD Show Lower Task Performance than Healthy Comparisons

As an initial assessment of task performance, we tested an LME predicting first free-choice accuracy (i.e., choice of the option with the higher average reward value) based on relevant task and experimental conditions. Details are provided in **Supplemental Materials**. In brief, accuracy was higher in H1 than H6 games (as expected), and iMUDs showed lower accuracy than HCs overall. Further, the change in accuracy between H1 and H6 games was less in iMUDs than HCs, consistent with less exploration in iMUDs.

To confirm expected improvements in accuracy over time, and potential modulation of this effect by group or anxiety induction, we tested a subsequent LME predicting accuracy on the six free choices of H6 games based on group, information condition, choice number (1-6), breathing resistance, and the three-way interactions of choice number, group, and breathing resistance, as well as between choice number, group and information condition (including respective two-way interactions). We observed that accuracy was again higher in HCs (estimated marginal mean [EMM]=.81) than iMUDs (EMM=.69), higher in the equal (EMM=.77) than unequal (EMM=.73) information condition, and increased as a function of choice number (see **Table 2**). There was also an interaction between group and resistance such that breathing resistance increased accuracy for HCs (Resistance–No resistance=.008) but decreased accuracy for iMUDs (Resistance–No Resistance=–.007; see **Figure 2**). All effects remained significant controlling for working memory in a subset of participants for which these scores were available (*p*s<.032).

![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/05/21/2024.05.17.24307491/F2.medium.gif)

[Figure 2.](http://medrxiv.org/content/early/2024/05/21/2024.05.17.24307491/F2)

Figure 2. 
H1 and H6 accuracy for each choice by group, resistance, and information condition. Error bars show 95% confidence intervals for accuracy at each choice number. As expected, accuracy was lower in H6 than H1 games for first free choice and improved with further choices in H6 games. The change in first free choice accuracy between H1 and H6 games was also greater in HCs than iMUDs, consistent with greater exploration in HCs.

View this table:
[Table 2.](http://medrxiv.org/content/early/2024/05/21/2024.05.17.24307491/T2)

Table 2. LME Results Predicting Accuracy across H6 Free Choice Trials

### Individuals with MUD Show Less Directed Exploration, Random Exploration, and Slower Learning Rates than Healthy Comparisons

Inter-correlations between model parameter estimates across participants were low (*r*s<.36; see **Supplemental Figure S1**). In our primary model-based analyses, LMEs were used to predict model parameter values based on group, resistance level, and their interaction (see **Table 3**). Result showed higher values in HCs for DE (EMMs: HCs=6.05; iMUDs=4.54), RE (EMMs: HCs=1.65; iMUDs=0.99), α (initial learning rate), and α∞ (asymptotic learning rate; i.e., the value to which a participant’s learning rate would theoretically converge if the game were played indefinitely; EMMs: HCs=0.30; iMUDs=0.22). However, the group difference in RE was no longer significant after two potential outliers were removed using a Grubb’s test (*F*(1,110.0)=1.83, CI=[-0.506, 0.095], *p*=.179, ![Graphic][1]</img>, *b*=-0.205). Because α values showed a bimodal distribution across participants (see **Figure 3**), we instead performed a k-means clustering analysis and divided participants into those with high and low values, and then used cluster membership as a categorical outcome variable in logistic mixed regressions in place of LMEs. In these regressions predicting α, we observed a significant group difference (proportion in high-value cluster: HCs=.67; iMUDs=.32). There was no main effect of breathing resistance or interaction with group for any parameter. When working memory was included as an additional covariate in the subset of participants with available data, all effects of group remained significant (*p*s<.032), except in relation to DE, which increased slightly above the threshold for significance (*F*(1,103.0)=3.32, *p=.*071, ![Graphic][2]</img>, *b=*–0.634, CI=[-1.324, 0.056]; see **Supplemental Table S7**). However, the effect of working memory was not a significant predictor of any model parameter (*ps>*.195).

![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/05/21/2024.05.17.24307491/F3.medium.gif)

[Figure 3.](http://medrxiv.org/content/early/2024/05/21/2024.05.17.24307491/F3)

Figure 3. 
Bar plots showing model parameter values from the Horizon Task by group and resistance condition. **p<*.05, ***p<*.01, \***|*p<*.001. Error bars indicate the standard error for each estimate.

View this table:
[Table 3.](http://medrxiv.org/content/early/2024/05/21/2024.05.17.24307491/T3)

Table 3. Results Testing Effects of Group and Resistance on Primary Computational Measures.

To better interpret these group differences, effects on parameters in H1 and H6 were then examined separately. In an LME predicting the information bonus parameter by horizon, group, and their interaction, the interaction was significant (*F*(1,340.0)=13.09, *p*<.001, ![Graphic][3]</img>, *b*=-0.389, CI=[-0.601, - 0.178]), reflecting the group difference in DE (see **Figure 3**). Post-hoc contrasts showed that HCs had higher information bonus values than iMUDs in H6 (HCs–iMUDs=1.67, *t*(139)=2.67, *p=*.0085), but not in H1 (HCs–iMUDs=.117, *t*(139)=0.19, *p=*.852). Notably, this interaction remained significant when accounting for working memory (*F*(1,325.0)=11.16, *p<.*001, ![Graphic][4]</img>, *b*=-0.366, CI=[-0.581, -0.150]), supporting an independent group difference in DE.

In an analogous LME predicting decision noise by horizon, group, and their interaction, the interaction was also significant, as expected (*F*(1,340.0)=4.08, *p=.*044, ![Graphic][5]</img>, *b*=-0.124, CI=[-0.245,- 0.003]). Post-hoc contrasts here instead suggested group differences in RE were driven by greater decision noise in iMUDs than HCs in H1 (HC–iMUD=–.433, *t*(234)= –2.07, *p=*.039), with no difference in H6 (HC–iMUD=.063, *t*(234)=0.30, *p=*.762; see **Figure 3**). Effects remained unchanged when accounting for working memory (*F*(1,325.0)=4.53, *p=.*034, ![Graphic][6]</img>, *b*=–0.135, CI=[-0.260, -0.010]).

Follow-up LMEs were run predicting model parameters based on substance use symptoms in iMUDs (i.e., DAST, DSQ, MAWQ; tested separately). In all models, we observed no significant main effects. However, there were two notable interactions between substance use symptoms and resistance condition, suggesting that the relationship between substance use symptoms and model parameters (i.e., DE and α∞) was negative for the task run without resistance but positive for the task run with resistance (see **Supplemental Tables S8** and **S10**). However, as these results were not hypothesized and somewhat difficult to interpret, we simply note them here for the interested reader and for purposes of future hypothesis generation.

LMEs testing potential parameter differences within iMUDs based on comorbid psychopathology (present/absent), continuous measures of psychopathology (PHQ-9, UPPS-P Total, and STAI-Trait Scores; tested separately), medication status (medicated/unmedicated), time since last methamphetamine use, and time since starting treatment did not show any significant effects (*ps* >.184).

### Greater Exploration and Faster Learning Rates each Improved Performance

To evaluate the theoretical significance of observed group differences, we tested whether some values for each model parameter might be considered more optimal than others with respect to task performance after the first free choice in H6. In brief, LMEs revealed higher values of *α*, and of *α*∞ predicted greater accuracy throughout the task, and higher values of DE and RE predicted steeper improvement in accuracy over time (*p*s<.039; see **Methods** and **Supplemental Tables S11** and **S12**).

### Directed Exploration and Learning Rates were Predicted by Cognitive Reflectiveness

As a secondary aim, we sought to replicate and extend our prior results linking exploration to cognitive reflection10. To do so, we tested LMEs predicting model parameters based on Cognitive Reflection Test (CRT) scores, accounting for potential effects of group and resistance. Across all participants, CRT score significantly predicted DE (*F*(1,107.0)=3.96, *p=.*049, ![Graphic][7]</img>, *b=*0.343, CI=[0.001, 0.685]) and *α* (*χ*2(1)=7.03, *p=.*008, *b=*0.554), but not *α*∞ (*F*(1,109.6)=0.10, *p=.*753, ![Graphic][8]</img>, *b=*0.002, CI=[-0.011, 0.015]) or RE (*F*(1,107.0)=0.07, *p=.*792, ![Graphic][9]</img>, *b=*0.024, CI=[-0.153, 0.200]). When additionally controlling for working memory, CRT remained a significant predictor of *α* (*χ*2(1)=7.33, *p=.*007, *b=*0.491) and was just over threshold for DE (*F*(1,101.0)=3.83, CI=[-0.005, 0.698], p=.053, ![Graphic][10]</img>, b=0.347).

For analogous models restricted to the iMUD sample, CRT score did not significantly predict DE, RE, or *α*∞ (*p*s>.177), and was just over threshold for *α* (*F*(1,51.0)=3.62, *p=.*063, ![Graphic][11]</img>, *b=*0.065). There were also no observed effects of resistance (*p*s>.193). Note that, unlike in the full sample, in iMUDs alone the distribution of *α* values was sufficiently normal to use an LME in place of logistic regression.

## Discussion

In the present study, we compared how treatment-seeking (currently abstinent) individuals with Methamphetamine Use Disorder (iMUDs) and healthy comparisons (HCs) differed in information-seeking and learning under uncertainty, both with and without a somatic anxiety induction. This allowed us to distinguish the causal effect of state anxiety from potential effects of other factors linked to psychopathology. As expected, we found that HCs outperformed iMUDs on the task. Computational modeling revealed that iMUDs had lower values of directed exploration (DE), random exploration (RE), initial learning rate (*α*), and asymptotic learning rate (*α*∞; controlling for *α*), while these parameters themselves were only weakly correlated.

The differences observed in DE and RE support previous research finding that individuals with substance use problems exhibit reduced exploration23,24. Importantly, however, unlike several previous studies, the Horizon Task allowed us to distinguish directed from random strategies, where measures of DE and RE in this task are also sensitive to beneficial vs. suboptimal engagement in exploratory behavior (i.e., with vs. without future choices that could benefit from information gain). Here, DE differences in iMUDs appeared to reflect an attenuated ability to increase exploration when it was beneficial (i.e., in games with longer horizon), whereas differences in RE were instead attributable to less reward sensitive choices when this was not beneficial (i.e., in games with shorter horizon). However, group differences in RE were no longer significant after potential outliers were removed; these results should therefore be treated with caution. Nevertheless, these findings offer insights into more specific cognitive mechanisms that might contribute to maladaptive choice and potentially withdrawal avoidance.

It is also noteworthy that iMUDs showed lower initial learning rates and asymptotic learning rates than HCs, suggesting that they updated their beliefs less after observing each new outcome. This could be taken to suggest a greater expectation that mean reward values will remain stable, or, relatedly, the belief that each observed outcome is less informative (i.e., more noisy, less trustworthy) regarding the true value of the generative mean25. In real-world contexts, lower learning rates could prevent individuals from changing their behavior, despite experiencing harmful consequences. However, it remains to be shown whether such results generalize to learning in daily life.

Contrary to our hypothesis, somatic anxiety did not affect any computational measure. Here, it is notable that, while previous research has shown negative correlations between anxiety and DE9,10, to our knowledge this is the first study to causally manipulate somatic state anxiety and differentiate its influence from trait factors. While this could indicate that state anxiety does not account for differences in information-seeking, it is possible that the resistance level, which was chosen to maintain tolerability, did not induce sufficiently high anxiety. Self-reported anxiety was higher for the task run with resistance (Mean=3.05) than without resistance (Mean=1.46), and this effect was greater for iMUDs than HCs, but anxiety scores were still well below the maximum score of 10. Future work might therefore aim to induce higher levels of somatic anxiety in a feasible manner and reassess its potential effects.

In line with our secondary aim, results also successfully replicated prior findings10 linking cognitive reflectiveness to DE, and also showed a novel association with initial learning rates. Given the positive relationship between model parameters and task choice accuracy, this suggests reflectiveness may facilitate sensitivity to uncertainty and promote adaptive information-seeking in uncertain environments. Notably, however, these relationships were not observed in iMUDs alone. This could be due to insufficient sample size, the lower values and restricted range of reflectiveness scores in iMUDs, or perhaps a mechanism whereby substance use decouples these variables. Future work should examine whether improving reflectiveness could promote more adaptive information seeking and learning, and whether this might be clinically beneficial. This possibility is supported by previous work showing that cognitive reflectiveness can be improved with training26,27.

It is important to consider limitations of the present study when interpreting these results. First, our sample size was only moderate and unable to reliably detect small effect sizes. We also could not determine whether observed group differences represent a preexisting vulnerability factor or a consequence of methamphetamine use. No relationships were found with length of abstinence, days since starting treatment, or medication status, perhaps suggesting that group differences were better explained by pre-existing factors or were insensitive to recovery; but longer recovery times will need to be examined. The presence of other comorbid affective and substance use disorders also did not appear to account for any results.

With these limitations in mind, we found that individuals with methamphetamine use disorder exhibited lower levels of exploration and reduced learning rates when making decisions under uncertainty. Contrary to expectation, we did not observe an effect of aversive interoceptive state induction (and resulting increases in somatic anxiety) on model parameters or other behavioral metrics, suggesting trait factors may be of more central importance. Overall, these results highlight directed exploration and learning rates as possible mechanisms of maladaptive choice in individuals with Methamphetamine Use Disorder and could point to novel treatment targets that could be tested in future work.

## Methods

### Participants

Participants included 56 inpatient treatment-seeking iMUDs and 58 HCs. Individuals with MUD were currently abstinent (mean time since methamphetamine use=50.04 days, mean time since starting treatment=34.47 days) and recruited from two recovery homes in the Tulsa, Oklahoma area: (1) GRAND and (2) Women in Recovery (WiR). All iMUDs met criteria for a DSM-5 diagnosis of Current Amphetamine Use Disorder due to use of Methamphetamine, which was assessed by clinical interview (Mini International Neuropsychiatric Interview 728). Due to high rates of comorbidity, iMUDs were not excluded based on the presence of other substance use disorders or depression/anxiety disorders (for a list of comorbid disorders in the MUD sample, see **Supplemental Table S1**). However, individuals with bipolar disorder, personality disorders, eating disorders, schizophrenia, or obsessive-compulsive disorder were excluded. Current use of psychotropic medications was permitted for iMUDs, as these are frequently utilized by providers in acute substance use treatment. HCs did not have any history of psychiatric illness and were not on any psychotropic medication.

### Protocol

After providing informed consent to participate in a larger study protocol, participants completed a drug test and breathalyzer assessment to confirm eligibility for the study. Next, iMUDs completed the Desire for Speed (Methamphetamine) Questionnaire (DSQ)29 to assess baseline craving levels.

Following completion of these questionnaires, participants were fit with a silicon mask (see **Figure 1**), which would later be used for anxiety induction during performance of the Horizon Task (described below). This breathing-based anxiety induction apparatus has been used safely and effectively in several previous studies30–33. Here, filters are used to add inspiratory resistance (i.e., requiring more effort to breathe in, but no added effort to breathe out), which creates a sensation of air hunger and elevates somatic anxiety. This initial fitting period was part of a sensitivity protocol designed to confirm sufficient comfort with the mask and allow us to assess how anxiety changed as a function of resistance level. During the preliminary sensitivity protocol, participants breathed through the mask while being exposed to six levels of resistance (0, 10, 20, 40, 60, and 80 cmH2O/L/sec) in ascending order for one minute each, with a short break in between each. After each exposure, they were instructed: “Please rate how much anxiety you felt while breathing from 0 to 10” (where 0 indicates no anxiety and 10 indicates maximum possible anxiety). We refer to this as *self-reported anxiety scores*. After completing this protocol, participants removed the mask. iMUDs then completed the DSQ a second time to assess whether craving levels had changed due to anxiety induction.

After this sensitivity protocol, participants completed neuropsychological testing and additional self-report questionnaires as part of the larger study protocol. This ensured participants were able to return to baseline arousal state before performing the Horizon Task. Participants were then re-fit with the mask before task performance and indicated their baseline level of anxiety (using both the self-report item mentioned above and the STAI State scale34). Next, they completed two runs of the Horizon Task, where one of the runs included a breathing resistance of 40 cmH2O/L/sec (counterbalanced order across participants). After each run, they again completed the STAI State scale and indicated their self-reported anxiety during task performance.

### Horizon Task

As in previous studies10, the Horizon Task here consisted of 80 games in which participants chose between two slot machines with different (unknown) average payout values (see **Figure 1** for a depiction of the task). For one of the slot machines, results were sampled from a Gaussian distribution with a mean of either 40 or 60 and a fixed standard deviation of 8. For the other slot machine, the distribution was shifted 4, 8, 12, 20, or 30 points in either direction from the first slot machine.

Participants first observed outcomes of four forced choices before they were allowed to make either one or six free choices between options to maximize the total number of points received. Games with one or six free choices are referred to as Horizon 1 (H1) and Horizon 6 (H6) games, respectively. The forced choices in each game were either equally informative (two forced choices for each slot machine) or unequally informative (three forced choices for one slot machine and one for the other). The different information conditions, decision horizons, and mean slot machine values were all counterbalanced throughout the task.

To minimize potential influences on individual differences in behavior, the observed outcome for each choice was sampled from the underlying Gaussian distributions but fixed across participants and task runs. Thus, two participants who chose the same option on a specific trial always observed the same result. However, after preliminary checking of data in the first five participants (all HCs), unexpected behavior in certain games led us to realize that forced choice outcomes in a few cases were not representative of the underlying distributions, which generated concerns given the number of trials per task condition (i.e., with respect to generative mean differences). To minimize this issue, forced choice results in these cases were re-sampled until they more closely aligned with the true differences between underlying distributions. Any potential effects of task version on behavior were accounted for in subsequent analyses.

### Computational Model

An established computational model was fit to task behavior (i.e., predicting the first free choice across games), as described in detail by Zajkowski, et al. 35. In brief, the probability of choosing the right option was calculated using a logistic choice function (equation 1) that included the difference in expected reward values between options, Δ*R*, the information difference between options, Δ*I*, a potential bias toward the right vs. left choice, *B*, and decision noise *σ*. ![Formula][12]</img>  The information difference (Δ*I*) was equal to +1 when one outcome was shown for the right option, –1 when three outcomes were shown for that option, and 0 when two outcomes were shown for each option. This was then scaled by a free parameter referred to as the information bonus (depicted above as *A*). The expected reward value difference (Δ*R*) was calculated using a Rescorla-Wagner update equation (2), where the learning rate *α* varied as a function of uncertainty (i.e., in relation to the number of previous observations). ![Formula][13]</img>  The initial learning rate *α* was a free parameter fit to participant data. For each subsequent choice, the learning rate updated with the following equation (3): ![Formula][14]</img>  Where *α**d* is derived from the following equation (4): ![Formula][15]</img>  Here, the *α*∞ term is also a free parameter fit to the data. This is the asymptotic learning rate, as it is the value to which the learning rate would theoretically converge if the game were played indefinitely (i.e., due to evolving levels of uncertainty after seeing an increasing number of outcomes). As evident in the equations above, a lower value of *α*∞ (bound between 0 and 1) would lead to lower values of *α**d*, causing the learning rate to decrease more rapidly between trials. Values for these learning rates could be seen to reflect implicit beliefs about the stability of mean rewards within a game and how quickly an individual changes their level of confidence in this expectation.

To get parameter estimates for each participant, a hierarchical Bayesian model36 with 12 free parameters in total was fit using a Markov Chain Monte Carlo (MCMC) method implemented with MATJAGS37. The spatial bias (*B*) and decision noise (*σ*) were fit separately for the four combinations of horizon (H1 or H6) and information condition (equal or unequal); the information bonus (*A*) was fit separately for the two horizon conditions (i.e., this can only be fit for unequal information games); the initial learning rate and asymptotic learning rate were fit across all games together. In the fitting procedure, group-level hyperparameters define prior distributions from which individual parameters were sampled (see **Supplemental Table S2** for the complete specification of these prior distributions). Note that, to avoid potential issues arising from the known covariance between information bonus and starting expected values for each game (i.e., prior to the first force choice), we fixed the initial expected values to 50.

### Measures

Participants completed the following measures to assess relevant clinical symptoms as well as trait and state psychological characteristics.

### Symptom Severity

To measure MUD severity, we used the Drug Abuse Screening Test (DAST), the Methamphetamine Withdrawal Questionnaire (MAWQ), and the Desires for Speed (Methamphetamine) Questionnaire (DSQ). DAST measures overall drug abuse severity and interference with life functioning38; MAWQ measures withdrawal symptoms39; and DSQ measures current craving levels29. These measures were only gathered in iMUDs.

To measure comorbid symptom dimensions associated with MUD, we used questionnaires measuring depression, anxiety, and impulsivity. Overall depressive symptoms were measured using the Patient Health Questionnaire (PHQ-9)40. State and trait anxiety was measured with the State-Trait Anxiety Inventory (STAI-State/Trait)34,41. Impulsivity was measured with the Urgency-Premeditation-Perseverance-Sensation Seeking-Positive Urgency (UPPS-P) Impulsive Behavior Scale Total Score42,43.

### Cognitive Reflectiveness

The Cognitive Reflection Test [CRT-744] measures the tendency to “stop and think” before immediately trusting one’s intuition. The test asks seven short questions designed such that there is an immediately intuitive, but incorrect, answer, and a correct answer that, while not logically difficult, requires the individual to devote effortful cognitive resources instead of immediately choosing the intuitively appealing response. An example item is “If it takes 5 machines 5 minutes to make 5 widgets, how long would it take 100 machines to make 100 widgets?” (intuitive incorrect answer: 100 minutes; correct answer: 5 minutes).

### Working Memory

The List Sorting Working Memory Test from the NIH Toolbox Cognition Battery45 was used to assess working memory. In our analyses, we used participants’ t-scores adjusted for age and sex.

## Statistical Analyses

Between-subject statistical analyses were carried out in R (version 4.2.1) with R Studio. K-means clustering was performed for α using the *kmeans* function of the *stats* package46. Linear mixed-effect models (LMEs) and logistic mixed regressions were run using the *lmer* function and the *glmer* function of the *lme4* package47. Potential outliers were identified using an iterative Grubb’s method (threshold: *p*<.01) using the *grubbs.test* function from the *outliers* package48. Effect sizes were calculated with the *F_to_eta2* function of the *effectsize* package49. All continuous predictors were mean-centered using the *gscale* function of the *jtools* package50. Unless otherwise stated, categorical variables were sum-coded as factors, including group (HCs=–1, iMUDs=1), breathing resistance (absent*=*–1, present*=*1), information condition (equal=–1, unequal=1), horizon (H1=–1, H6=1), sex (male=–1, female=1), and task version (main version=–1, initial version in first 5 participants=1).

The variables age and sex were included in all models as potential covariates to ensure they did not explain observed effects. As a small number of participants also completed a slightly different version of the task (i.e., with a different sequence of reward values sampled from the underlying generative means; described above), task version was also included in all models as a potential covariate. After controlling for age, sex, and task version, a follow-up model was run that additionally included working memory capacity, given that general cognitive ability has previously been shown to positively correlate with performance in the Horizon Task10. As working memory data was missing for a subset of participants (N=4), its potential explanatory power was only assessed in the subset of participants with available data in these follow-up analyses (as this would otherwise effectively remove data from these four participants from all analyses). When necessary, significant effects were further interpreted using post-hoc contrasts of estimated marginal trends (EMTs) or estimated marginal means (EMMs) using the *emmeans* package51.

### Protocol Validation

To test whether administration of the moderate breathing resistance level (40 cmH2O/L/sec) used during task performance also successfully increased anxiety, an LME was run predicting self-reported anxiety during the task, with resistance condition (baseline, task run with resistance, task run without resistance), group, and their interaction as predictors. Identical LMEs were also run using STAI state scores as the outcome variable in place of self-reported anxiety to confirm consistency (see **Supplemental Table S4** for the results of this model).

To confirm efficacy of the aversive state induction, we performed another LME to test if administration of the breathing resistance successfully induced anxiety within the pre-task exposure protocol. This model specifically assessed whether self-reported anxiety level was predicted by breathing resistance level (0, 10, 20, 40, 60, and 80 cmH2O/L/sec), group, and/or their interaction (see **Supplemental Table S5** for the results of this model).

### Model-Based Analysis

Model-based measures included: directed exploration (DE), random exploration (RE), initial learning rate (α), and asymptotic learning rate (α∞). DE was calculated by subtracting the information bonus parameter fit to H1 trials from that fit to H6 trials. This allowed us to measure the degree to which participants became more information-seeking as decision horizon increased (i.e., when information became goal-relevant). Note that this only applied to trials in which unequal information was given. RE was calculated by subtracting the decision noise parameter fit to H1 trials from that fit to H6 trials, allowing us to measure the degree to which participants became less value sensitive in their initial choice as decision horizon increased (i.e., which can also serve as an information-seeking strategy). Analyses of RE were here restricted to trials where equal information was given, such that directed information-seeking could not account for any apparent changes in value sensitivity. For analyses with α∞ as an outcome variable, we also included α as a covariate, given that those with the highest initial learning rate tended to experience the greatest decrease in learning rate over time (somewhat analogous to regression to the mean).

To examine potential effects of group and breathing resistance on each of these model parameters, separate LMEs (and logistic mixed regression for α) were run predicting each parameter value from group, resistance condition, and the interaction of those variables. If extreme values of model parameters were identified (testing within each resistance condition separately), analyses were repeated with those data removed, and any discrepancies between results with/without outliers were reported.

To assess whether observed differences in DE and RE were better explained by differences in H1 or H6, separate models were also run with information bonus or decision noise as the outcome variable, respectively, including horizon as an additional predictor. Note that the strength of the interaction between horizon and group can here also be seen as a test of group differences in DE or RE.

To examine the relationship between model parameters and substance use symptoms, separate LMEs were run predicting each parameter value based on a drug-related measure, accounting for the effect of breathing resistance (see **Supplemental Tables S8-S10** for additional details and model results). Here, we were additionally interested in examining if the relationship between parameter values and drug-related symptoms differed between resistance conditions, so the interaction between these variables was also included in the models.

To determine if group differences in model parameters were attributable to comorbid psychopathology in iMUDs, follow-up analyses were run in this sample predicting each parameter value based on the presence of each comorbid affective or substance use disorder (present/absent), accounting for any effect of breathing resistance. Analogous models were also run that included continuous measures of psychopathology (i.e., PHQ-9, STAI-Trait, and UPPS-P) as predictors. To determine if medication status (medicated/unmedicated), time since last use of methamphetamine, and/or time since starting treatment might influence model parameters, similar LMEs were run with those variables (separately).

### Model-Free and Model-Based Behavioral Analyses

To evaluate the effect of experimental condition on task performance, LMEs tested if free-choice accuracy differed by group, information condition, and/or horizon, and whether these effects might be moderated by breathing resistance and information condition. We additionally tested if these variables predicted accuracy across the free choices of the H6 condition. To examine how model parameters influenced subsequent task performance in H6 trials (i.e., to interpret whether some values might be considered more optimal than others), we also examined if accuracy was predicted by each of the model parameters, free choice number (2-6; i.e., excluding the first free choice to which these parameters were directly fit), resistance, and/or group, and whether a given model parameter moderated the improvement in accuracy as choice number increased (see **Supplemental Table S11** and **S12** for full model results).

### Model Parameters Predicting Cognitive Reflectiveness

We also sought to replicate prior results10, and extend them to iMUDs, linking model-based Horizon Task metrics (and DE in particular) to cognitive reflectiveness (i.e., CRT scores). We therefore tested if model parameters could be predicted by the number of correct answers on this measure (accounting for effects of resistance). Group was included as a covariate to ensure that observed effects were not explained by group differences in cognitive reflectiveness or parameter values (see **Table 1** and **Figure 1**). These analyses were repeated with data restricted to the iMUDs sample to determine if any observed relationships were specific to the iMUD population.

## Supporting information

Supplemental Materials [[supplements/307491_file02.pdf]](pending:yes)

## Data Availability

All data produced in the present study are available upon reasonable request to the authors.

## Author Contributions

Writing: C.M.G., T.T., and R.S. Review and editing: all authors. Data analysis: C.M.G., T.T., C.A.L., R.S, and R.H. Methodology: R.S and R.C.W. Conceptualization: R.S., J.L.S, R.C.W., S.S.K., M.P.P. Funding acquisition: R.S. and M.P.P. Data collection: A.E.C., S.T., C.A.L., N.L. Project administration: R.S.

## Data and Code Availability

All data and code used in the formulation of this manuscript are available upon reasonable request to the authors.

## Competing Interests

The authors have no competing interests to declare.

## Acknowledgements

This project was funded by National Institute of General Medical Sciences (P20GM121312 [R.S. and M.P.P.]) and the Laureate Institute for Brain Research.

## Footnotes

*   * indicates co-first authorship

*   Received May 17, 2024.
*   Revision received May 17, 2024.
*   Accepted May 21, 2024.


*   © 2024, Posted by Cold Spring Harbor Laboratory

This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/)

## References

1.  Paulus, M. P. & Stewart, J. L. Neurobiology, clinical presentation, and treatment of methamphetamine use disorder: a review. JAMA psychiatry 77, 959–966 (2020).
    
    

2.  May, A. C., Aupperle, R. L. & Stewart, J. L. Dark times: the role of negative reinforcement in methamphetamine addiction. Frontiers in psychiatry 11, 504133 (2020).
    
    

3.  Crime, U. N. O. o. D. a. (United Nations, New York, 2022).
    
    

4.  May, A. C., Aupperle, R. L. & Stewart, J. L. Dark times: the role of negative reinforcement in methamphetamine addiction. Frontiers in psychiatry 11, 114 (2020).
    
    

5.  Stewart, J. L. et al. You are the danger: attenuated insula response in methamphetamine users during aversive interoceptive decision-making. Drug and alcohol dependence 142, 110–119 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.drugalcdep.2014.06.003&link_type=DOI) 

6.  Wilson, R. C., Bonawitz, E., Costa, V. D. & Ebitz, R. B. Balancing exploration and exploitation with information and randomization. Curr Opin Behav Sci 38, 49–56, doi:10.1016/j.cobeha.2020.10.001 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cobeha.2020.10.001&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33184605&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 

7.  Gershman, S. J. Deconstructing the human algorithms for exploration. Cognition 173, 34–42, doi:10.1016/j.cognition.2017.12.014 (2018).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cognition.2017.12.014&link_type=DOI) 

8.  Wilson, R. C., Geana, A., White, J. M., Ludvig, E. A. & Cohen, J. D. Humans use directed and random exploration to solve the explore-exploit dilemma. J Exp Psychol Gen 143, 2074–2081, doi:10.1037/a0038199 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1037/a0038199&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25347535&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 

9.  Fan, H., Gershman, S. J. & Phelps, E. A. Trait somatic anxiety is associated with reduced directed exploration and underestimation of uncertainty. Nature Human Behaviour 7, 102–113 (2023).
    
    

10. Smith, R. et al. Lower levels of directed exploration and reflective thinking are associated with greater anxiety and depression. Frontiers in Psychiatry 12, 782136 (2022).
    
    

11. Paulus, M. P. & Stewart, J. L. Methamphetamine use disorder: the next addiction crisis. JAMA psychiatry 77, 959 (2020).
    
    

12. Clark, L., Robbins, T. W., Ersche, K. D. & Sahakian, B. J. Reflection impulsivity in current and former substance users. Biological psychiatry 60, 515–522 (2006).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.biopsych.2005.11.007&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16448627&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000240506000012&link_type=ISI) 

13. Calipari, E. S. & Ferris, M. J. Amphetamine mechanisms and actions at the dopamine terminal revisited. Journal of Neuroscience 33, 8923–8925 (2013).
    
    [FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6Njoiam5ldXJvIjtzOjU6InJlc2lkIjtzOjEwOiIzMy8yMS84OTIzIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMDUvMjEvMjAyNC4wNS4xNy4yNDMwNzQ5MS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

14. Byrne, K. A. et al. Examining the relationship between substance use and exploration–exploitation behavior in young adults. Experimental and Clinical Psychopharmacology (2022).
    
    

15. Robinson, A. H., Chong, T. T. J. & Verdejo-Garcia, A. Computational models of exploration and exploitation characterise onset and efficacy of treatment in methamphetamine use disorder. Addiction Biology 27, e13172 (2022).
    
    

16. Ghaderi, S., Hemami, M., Khosrowabadi, R. & Rad, J. A. The role of reinforcement learning in shaping the decision policy in methamphetamine use disorders. Journal of choice modelling 50, 100469 (2024).
    
    

17. Robinson, A. H., Perales, J. C., Volpe, I., Chong, T. T. J. & Verdejo-Garcia, A. Are methamphetamine users compulsive? Faulty reinforcement learning, not inflexibility, underlies decision making in people with methamphetamine use disorder. Addiction Biology 26, e12999 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/adb.12999&link_type=DOI) 

18. Smith, R. et al. Imprecise Action Selection in Substance Use Disorder: Evidence for Active Learning Impairments When Solving the Explore-exploit Dilemma. Drug and Alcohol Dependence 215, 108208 (2020).
    
    

19. Ghaderi, S., Rad, J. A., Hemami, M. & Khosrowabadi, R. Dysfunctional feedback processing in male methamphetamine abusers: Evidence from neurophysiological and computational approaches. Neuropsychologia, 108847 (2024).
    
    

20. Taylor, S. et al. Active learning impairments in substance use disorders when resolving the explore-exploit dilemma: A replication and extension of previous computational modeling results. Drug and Alcohol Dependence 252, 110945 (2023).
    
    

21. Smith, R. et al. Slower Learning Rates from Negative Outcomes in Substance Use Disorder over a 1-Year Period and Their Potential Predictive Utility. Computational Psychiatry 6, 117–141 (2022).
    
    

22. Rafizadeh, R. et al. Association of clozapine treatment and rate of methamphetamine or amphetamine relapses and abstinence among individuals with concurrent schizophrenia spectrum and amphetamine use disorder: A retrospective cohort study. Journal of Psychopharmacology 37, 1040–1048 (2023).
    
    

23. Morris, L. S. et al. Biases in the explore–exploit tradeoff in addictions: the role of avoidance of uncertainty. Neuropsychopharmacology 41, 940–948 (2016).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/npp.2015.208&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26174598&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 

24. Addicott, M. A., Pearson, J. M., Froeliger, B., Platt, M. L. & McClernon, F. J. Smoking automaticity and tolerance moderate brain activation during explore–exploit behavior. Psychiatry Research: Neuroimaging 224, 254–261 (2014).
    
    

25. Piray, P. & Daw, N. D. A model for learning based on the joint estimation of stochasticity and volatility. Nat Commun 12, 6587, doi:10.1038/s41467-021-26731-9 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41467-021-26731-9&link_type=DOI) 

26. Calvillo, D. P., Bratton, J., Velazquez, V., Smelter, T. J. & Crum, D. Elaborative feedback and instruction improve cognitive reflection but do not transfer to related tasks. Thinking & Reasoning 29, 276–304 (2023).
    
    

27. Lachaud, L., Jacquet, B., Bourlier, M. & Baratgin, J. Mindfulness-based stress reduction is linked with an improved Cognitive Reflection Test score. Frontiers in Psychology 14, 1272324 (2023).
    
    

28. Sheehan, D. V. et al. The Mini-International Neuropsychiatric Interview (MINI): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. Journal of clinical psychiatry 59, 22–33 (1998).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=9448666&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 

29. James, D., Davies, G. & Willner, P. The development and initial validation of a questionnaire to measure craving for amphetamine. Addiction 99, 1181–1188 (2004).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.1360-0443.2004.00819.x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15317639&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000223721100018&link_type=ISI) 

30. Berk, L., May, A. C., Stewart, J. L., Paulus, M. P. & Tapert, S. F. in BIOLOGICAL PSYCHIATRY. 310S-310S (ELSEVIER SCIENCE INC 360 PARK AVE SOUTH, NEW YORK, NY 10010-1710 USA).
    
    

31. Berk, L. et al. Under pressure: adolescent substance users show exaggerated neural processing of aversive interoceptive stimuli. Addiction 110, 2025–2036, doi:10.1111/add.13090 (2015).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/add.13090&link_type=DOI) 

32. Berner, L. A. et al. Altered interoceptive activation before, during, and after aversive breathing load in women remitted from anorexia nervosa. Psychol Med 48, 142–154, doi:10.1017/S0033291717001635 (2018).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1017/S0033291717001635&link_type=DOI) 

33. Stewart, J. L. et al. You are the danger: attenuated insula response in methamphetamine users during aversive interoceptive decision-making. Drug Alcohol Depend 142, 110–119, doi:10.1016/j.drugalcdep.2014.06.003 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.drugalcdep.2014.06.003&link_type=DOI) 

34. Spielberger, C. D. Manual for the State-Trait Inventory STAI (Form Y). (Mind Garden, 1983).
    
    

35. Zajkowski, W. K., Kossut, M. & Wilson, R. C. A causal role for right frontopolar cortex in directed, but not random, exploration. Elife 6, e27430 (2017).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.7554/Elife.27430&link_type=DOI) 

36. Lee, M. D. & Wagenmakers, E.-J. Bayesian cognitive modeling: A practical course. (Cambridge university press, 2014).
    
    

37. Plummer, M. (Lyon, France, 2012).
    
    

38. Bohn, M., Babor, T. & Kranzler, H. Validity of the Drug Abuse Screening Test (DAST-10) in inpatient substance abusers. Problems of drug dependence 119, 233–235 (1991).
    
    

39. Zorick, T. et al. Withdrawal symptoms in abstinent methamphetamine-dependent subjects. Addiction 105, 1809–1818 (2010).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.1360-0443.2010.03066.x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=20840201&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000281850600021&link_type=ISI) 

40. Kroenke, K. & Spitzer, R. L. Vol. 32 509-515 (Slack Incorporated Thorofare, NJ, 2002).
    
    

41. Spielberger, C. D., Gonzalez-Reigosa, F., Martinez-Urrutia, A., Natalicio, L. F. & Natalicio, D. S. The state-trait anxiety inventory. Revista Interamericana de Psicologia/Interamerican journal of psychology 5 (1971).
    
    

42. Whiteside, S. P. & Lynam, D. R. The five factor model and impulsivity: Using a structural model of personality to understand impulsivity. Personality and individual differences 30, 669–689 (2001).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0191-8869(00)00064-7&link_type=DOI) 

43. Cyders, M. A. & Smith, G. T. Emotion-based dispositions to rash action: positive and negative urgency. Psychological bulletin 134, 807 (2008).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1037/a0013341&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18954158&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000260279500002&link_type=ISI) 

44. Toplak, M. E., West, R. F. & Stanovich, K. E. Assessing miserly information processing: An expansion of the Cognitive Reflection Test. Thinking & reasoning 20, 147–168 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1080/13546783.2013.844729&link_type=DOI) 

45. Tulsky, D. S. et al. NIH Toolbox Cognition Battery (NIHTB-CB): List sorting test to measure working memory. Journal of the International Neuropsychological Society 20, 599–610 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1017/S135561771400040X&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24959983&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 

46. Team, R. C., Team, M. R. C., Suggests, M. & Matrix, S. Package stats. The R Stats Package (2018).
    
    

47. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software 67, 1–48, doi:10.18637/jss.v067.i01 (2015).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.18637/jss.v067.i01&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23757445&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F05%2F21%2F2024.05.17.24307491.atom) 

48. Komsta, L. outliers: Tests for outliers. R package version 0.14 (2011).
    
    

49. Ben-Shachar, M. S., et al. (2022).
    
    

50. Long, J. A. jtools: Analysis and Presentation of Social Scientific Data. R package version 2.2.0, <[https://cran.r-project.org/package=jtools](https://cran.r-project.org/package=jtools)> (2022).
    
    

51. Lenth, R. emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version 1.8.8, <[https://CRAN.R-project.org/package=emmeans](https://CRAN.R-project.org/package=emmeans)> (2023).

 [1]: /embed/inline-graphic-1.gif
 [2]: /embed/inline-graphic-2.gif
 [3]: /embed/inline-graphic-3.gif
 [4]: /embed/inline-graphic-4.gif
 [5]: /embed/inline-graphic-5.gif
 [6]: /embed/inline-graphic-6.gif
 [7]: /embed/inline-graphic-7.gif
 [8]: /embed/inline-graphic-8.gif
 [9]: /embed/inline-graphic-9.gif
 [10]: /embed/inline-graphic-10.gif
 [11]: /embed/inline-graphic-11.gif
 [12]: /embed/graphic-8.gif
 [13]: /embed/graphic-9.gif
 [14]: /embed/graphic-10.gif
 [15]: /embed/graphic-11.gif