Alcohol craving plays a central role in sustaining alcohol use and contributing to relapse. Noninvasive neuromodulation techniques such as transcranial photobiomodulation (tPBM) and transauricular vagus nerve stimulation (taVNS) may target craving-related neural circuits through complementary cortical and subcortical mechanisms. However, their relative and combined effects on craving and related symptoms remain unclear. To compare the effects of tPBM, taVNS, and combined tPBM + taVNS on alcohol craving, alcohol use, and psychological outcomes in individuals with subclinical alcohol use. Sixty participants were randomized to one of three intervention groups: tPBM, taVNS, or combined tPBM + taVNS. Participants self-administered the assigned stimulation five times per week for five weeks using portable home devices (15 min/session, total 25 sessions). Change in craving was the primary outcome. Both tPBM and combined tPBM + taVNS produced significant reductions in craving (PACS: p < 0.001) and alcohol use (AUDIT: p ≤ 0.011) compared with taVNS alone. ANCOVA revealed a significant group effect on craving (F(2,54) = 5.24, p = 0.008). Logistic regression showed higher odds of craving response with tPBM (OR = 8.09) and of drinking-behavior improvement with combined stimulation (OR = 16.73). tPBM and combined tPBM + taVNS effectively reduced craving, alcohol use in individuals with subclinical alcohol use.

Impulsive decision-making is strongly associated with substance use disorders (SUD). Using computational modeling, we investigated similarities and differences in decision-making profiles between stimulant and opioid users. Additionally, we used a sibling comparison design to determine if impulsive decision-making could serve as a computational marker of familial liability for addiction. Participants included individuals with “pure” SUD (opioids: N=157, stimulants: N=140), healthy controls (N=240), and unaffected siblings of opioid (N=67) and stimulant (N=52) users. We used the Iowa Gambling Task (IGT), Cambridge Gambling Task (CGT), Balloon Analogue Risk Task (BART), and Delayed Reward Discounting Task (DRDT) to measure decision-making. Task behaviors were analyzed using hierarchical Bayesian modeling and comparing group-level decision-making computational parameters. Both substance-using groups exhibited higher delay discounting on the DRDT and CGT. On the IGT, they weighed more recent over distant choices and tended to prioritize past choices over past rewards, as indicated by high perseverance weight and low outcome frequency weight. These patterns were also observed in their unaffected siblings, suggesting that these patterns may reflect computational markers of familial vulnerability to addiction. Opioid users and their siblings displayed reduced loss sensitivity on the IGT and CGT, suggesting that this may be a computational marker associated with familial vulnerability specific to opioid addiction. This study highlights common and specific computational markers of decision-making vulnerabilities associated with addictions to different types of drugs. Some of these markers were observed in unaffected siblings, suggesting that they may reflect familial liability and serve as potential endophenotypes for different types of addiction.

Research on the neurocomputational mechanisms of moral judgment has typically focused on contrasting "utilitarian" preferences to impartially maximize aggregate welfare and "deontological" preferences that judge the morality of actions based on rules. However, there has been little work to decompose the cognitive subcomponents of deontological preferences. Here, we investigated the neurocomputational mechanisms underlying two types of deontological preferences (Rawlsian and Kantian) and their contrast with utilitarian preferences in an incentivized moral dilemma task. Participants repeatedly decided how to allocate harm between a single individual ("the one") and a group of three to four individuals ("the group"). The task distinguished preferences for Rawlsian, Kantian, and utilitarian strategies by quantifying trade-offs among active harm, concern for the worst-off individual, and overall utility. Behaviorally, participants favored the Rawlsian strategy, preferring to impose more harm overall rather than disproportionately harm the one individual. Computational modeling revealed two dissociable dimensions of individual variability in Rawlsian preferences: (i) minimizing the maximum amount of harm delivered to a single person and (ii) subjective threshold of acceptable amount of harm imposed on one person. The combination of univariate and multivariate functional MRI analyses revealed the engagement of distinct brain regions in these two dimensions of Rawlsian preferences, which respectively mapped onto activity in mentalizing and valuation networks. Our results reveal the neurocomputational mechanisms guiding trade-offs between the welfare of one versus a larger group and highlight distinct roles for the mentalizing and valuation networks in shaping Rawlsian moral preferences.

Abstract Background Emotion dysregulation is a core feature of bipolar disorder (BD), yet its behavioral expression during depressive episodes, and potential differences between its types, BD-I and BD-II, remain unclear. This study used automated facial-expression analysis during naturalistic affective film viewing to examine subtype-specific and context-dependent emotional responding in bipolar depression. Methods The sample included 135 participants: 69 healthy controls and 66 patients with BD (BD-I, 23; BD-II, 43). Participants viewed nine emotionally evocative film clips spanning negative, positive, neutral, and socially threatening contexts, while their facial expressions were continuously recorded and quantified using computer vision-based facial-expression analysis. Results Patients with BD-I showed a distinct, context-dependent facial-expression profile, characterized by greater negative responses across multiple contexts than other groups. Specifically, they showed increased sadness during sad, reward, and amusing clips, and elevated anger during sad and neutral clips. In socially threatening contexts, BD-I participants showed a multivalent pattern of elevated anger, fear, and joy, suggesting poorly coordinated or context-incongruent affective expression. In contrast, BD-II participants did not differ significantly from healthy controls on any emotion, despite depressive symptom severity comparable to BD-I participants. Conclusions These findings suggest that facial-expression patterns in bipolar depression differ across subtypes. BD-I may be characterized by heightened negative reactivity and altered context-appropriate modulation of emotional expression, whereas BD-II may not show comparable alterations in overt facial output. Automated facial-expression analysis during naturalistic stimulation may provide a useful behavioral marker for characterizing subtype-specific affective disturbance in bipolar depression and related psychopathology.

Introduction: Non-invasive brain stimulation (NIBS) techniques have emerged as promising interventions for addiction, yet their effects on the cognitive processes underlying addictive behaviors remain inconclusive. Methods: Following PRISMA 2020 guidelines, we systematically reviewed sham-controlled studies identified through PubMed, PsycInfo, and EMBASE (search through November 2025) examining the effects of NIBS on cognitive processes in substance use and behavioral addictions. 39 studies were included and narratively synthesized across four cognitive domains: attentional and behavioral bias, inhibitory control, valuation and decision-making, and reward processing and learning. Results: Across studies, the dorsolateral prefrontal cortex (DLPFC) was the predominant stimulation target, with excitatory protocols (e.g., anodal tDCS, high-frequency rTMS) most commonly applied. Effect direction analysis indicated domain-specific patterns: NIBS most consistently modulated reward-related learning processes, whereas attentional bias measures showed little evidence of change. Findings for inhibitory control and decision-making tasks were mixed. Discussion: Results suggest that some cognitive processes may be more sensitive to cortical neuromodulation than others. A major limitation of the current literature is the reliance on aggregate task performance measures, which obscure the specific cognitive computations influenced by stimulation. Future research may benefit from applying computational modeling approaches, particularly reinforcement learning frameworks, to decompose behavioral performance into latent parameters with clear psychological interpretations. Such approaches may enable more precise mechanistic understanding of NIBS effects and support the development of targeted neuromodulation interventions for addiction.

Background and Aims: Alcohol use disorder (AUD) involves day-to-day fluctuations in cognitive, psychological, and contextual factors that are difficult to fully capture using retrospective or cross-sectional assessment alone. Computational markers, such as discounting rate, ambiguity tolerance, and risk preference, may provide useful indicators of day-level drinking risk, but have rarely been assessed repeatedly in real-world settings or examined together with daily app-based self-reports and passive smartphone-derived measures. This study examined whether these multimodal day-level measures were associated with same-day nighttime drinking, focusing on whether within-person changes in computational markers remained associated with drinking after accounting for concurrent self-reported and contextual measures. Design: A 28-day longitudinal observational study using repeated smartphone-based assessments and passive sensing. Setting: Daily-life smartphone-based monitoring of community-recruited participants in the Republic of Korea. Participants: The final analytic sample included 143 adults meeting DSM-5 criteria for AUD. Measurements: The primary outcome was same-day nighttime drinking, assessed using daily app-based self-reports. Daily predictors covered three domains. App-based self-reports included mood, alcohol craving, meal intake, and sleep-related variables. Computational predictors were derived from delay discounting and choice under risk and ambiguity tasks. Passive smartphone-derived predictors included time spent at home, calls, steps, and foreground app use. Generalized linear mixed-effects models examined associations between nighttime drinking and multimodal predictors decomposed into within-person and between-person components. Findings: A multimodal model integrating computational parameters, daily app-based self-reports, and passive smartphone-derived measures showed the highest explanatory power for same-day nighttime drinking (marginal R² = 0.22). In this final model, higher within-person delay discounting [log(k)], alcohol craving, and positive mood, as well as lower meal intake and percent time at home, were associated with a higher likelihood of nighttime drinking. At the between-person level, only lower average meal intake was associated with a higher likelihood of nighttime drinking. Within-person fluctuations in log(k) were largely uncorrelated with other day-level predictors and showed no clear next-day association. Conclusions: Associations with same-day nighttime drinking were observed primarily at the within-person level, suggesting that day-level changes may be more informative than stable individual differences. Within-person increases in delay discounting remained associated with nighttime drinking after accounting for other modalities, suggesting that multimodal monitoring may help identify periods of elevated drinking risk.

Addiction involves rapidly fluctuating affective and value-based decision-making processes that undermine cessation efforts, yet capturing these dynamics at clinically meaningful timescales remains challenging. Standard cognitive decision-making tasks are time-intensive and require many trials to achieve reliable parameter estimates, limiting their use longitudinally and in real-world settings. Here, we developed a rapid, smartphone-based framework that integrates ecological momentary assessment (EMA) with adaptive design optimization (ADO), a Bayesian method that enables reliable estimation of computational decision-making parameters from as few as 20-30 trials per task. We tested this framework in N = 79 individuals undergoing a 5-6-week smoking cessation program, who completed daily ADO-based delay discounting and risk/ambiguity tasks alongside EMA surveys assessing smoking behavior, craving, stress, mood, anxiety, and medication adherence. At the day-to-day level, elevated craving, depressive symptoms, and ambiguity tolerance predicted increased smoking the following day, whereas lower discounting rates and reduced craving and stress predicted cessation success at treatment completion. For the latter, models using data from the first week achieved meaningful predictive performance (mean AUC = 0.76), approaching the upper-bound performance observed in models incorporating both task and survey data from the full study period (mean AUC = 0.83). Together, these findings demonstrate that rapid, low-burden, ADO-based delivery of decision-making tasks via EMA can capture clinically relevant, dynamic vulnerability states during smoking cessation treatment. This methodology offers a promising approach for identifying cognitive markers that may facilitate or inhibit cessation success and informing personalized, time-sensitive intervention strategies for nicotine addiction and related psychiatric disorders.

Purpose of Review Despite the growing number of studies on interoception in addiction research, a comprehensive review of these findings is still lacking. The goal of this paper was to conduct a systematic review of empirical studies to assess the evidence for interoceptive deficits in individuals with substance use disorders (SUDs), identify gaps in the existing literature, and propose directions for future research. Recent Findings From 1,544 records identified in APA PsycINFO, PubMed, CINAHL, and Embase using predefined search terms, 39 studies met eligibility criteria. Of these, 14 examined alcohol use disorder, 12 heterogeneous substance use, 6 opioid use disorder, 4 smokers, 2 cocaine use disorder, and 1 cannabis use. With respect to interoceptive domains (accuracy, sensibility, awareness), 12 studies assessed accuracy only, 19 assessed sensibility only, and 8 measured both; only one study additionally evaluated interoceptive awareness. Behavioral tasks (e.g., heartbeat counting) and self-report questionnaires (e.g., Multidimensional Assessment of Interoceptive Awareness; MAIA) were the most common measures of interoceptive accuracy and sensibility, respectively. Neuroimaging work was scarce, with only four studies examining neural correlates. Summary Across studies of interoception in SUDs, the most consistent pattern is reduced interoceptive accuracy relative to healthy controls, particularly in alcohol use disorder where impairments were most robust, though findings for other substances showed greater variability. Evidence for interoceptive sensibility is mixed, and work on interoceptive awareness remains sparse. The literature is constrained by a narrow emphasis on two dimensions and by heavy reliance on the heartbeat counting task, which has been widely criticized for validity concerns. Future research should assess multiple interoceptive dimensions across modalities, develop and validate more robust measures, and address potential confounders such as smoking status.

In this study, we present behavioral, computational, and neuroimaging evidence that social stress enhances intuitive prosocial value processing while impairing self-reward processing. When deciding on monetary rewards for individuals at various social distances, participants who exhibited elevated cortisol levels following a social stress task were more inclined to choose a disadvantageous unequal option. Neuroimaging data revealed that participants more likely to choose the disadvantageous unequal option exhibited increased encoding of other-regarding rewards in the ventral medial prefrontal cortex (mPFC), whereas the dorsal mPFC exhibited a decrease in encoding. Mediation analyses further indicated that both the ventral and dorsal mPFC indirectly mediated the relationship between heightened cortisol levels and a greater likelihood of choosing a disadvantageous unequal option. Additionally, effective connectivity analysis results demonstrated that cortisol has an excitatory effect on the dorsal mPFC via the ventral striatum, while simultaneously sending inhibitory signals to the dorsal mPFC via the dorsal striatum. These findings provide empirical evidence to clarify the ambiguity surrounding the effects of stress on prosocial decision-making, suggesting that social stress disrupts deliberative decision-making while simultaneously promoting intuitive prosocial motivation through the differential modulation of hierarchically organized cortico-striatal loops.

BackgroundTobacco smoking continues to be a leading cause of preventable morbidity and mortality globally, with the success rate of unaided cessation remaining consistently low. Understanding the neurobiological mechanisms of smoking cessation is crucial for improving quit rates. However, there has been a lack of studies examining brain network changes associated with smoking cessation over time. In this study, we aimed to investigate longitudinal changes in the functional connectivity (FC) of large-scale brain networks underlying smoking cessation outcomes using resting-state functional magnetic resonance imaging (fMRI).MethodsA total of 98 treatment-seeking smokers participated in a 5-week cessation program and underwent resting-state fMRI scans before and after the intervention. Independent component analysis identified the salience network (SN), executive control network (ECN), and default mode network (DMN) components, and region of interest (ROI)-to-ROI FC was compared between successful and unsuccessful quitters using a group × time mixed-effects model. Correlations with smoking-related measures were explored.ResultsSignificant group-by-time interaction effects were found in FC, particularly involving connections between SN and ECN, as well as between the SN and DMN. Specifically, successful quitters exhibited greater baseline FC in the SN-ECN and SN-DMN circuits, which tended to decrease and converge toward levels observed in unsuccessful quitters during the cessation process. Exploratory correlational analyses revealed trends suggesting that stronger pre-quit connectivity between the SN and ECN was associated with greater withdrawal severity and longer smoking history in successful quitters.ConclusionTaken together, the reduction of initially elevated pre-quit FC in SN-ECN and SN-DMN circuits may reflect an adaptive neural process that supports successful withdrawal management and attentional reallocation during cessation. The identification of these neural substrates not only enhances our mechanistic understanding of smoking cessation over time but also underscores the need for targeted interventions that focus on these neural circuits to enhance quit outcomes.

Objective: Impulsive decision-making is strongly associated with substance use disorders (SUD). We used a battery of decision-making tasks to investigate similarities and differences in decision-making patterns between stimulant (amphetamine) and opioid (heroin) users. Additionally, we explored whether impulsive decision-making could serve as a potential endophenotype for addiction using a sibling comparison design. Methods: Participants included individuals with “pure” SUD (opioids: N=157, stimulants: N=140), healthy controls (N=240), and unaffected siblings of opioid (N=67) and stimulant (N=52) users. We used the Iowa Gambling Task (IGT), Cambridge Gambling Task (CGT), Balloon Analogue Risk Task (BART), and Delayed Reward Discounting Task (DRDT) to measure decision-making. Task behaviors were analyzed using hierarchical Bayesian modeling and we compared the group-level decision-making computational parameters. Results: Both drug groups had higher temporal discounting of rewards on the DRDT and CGT; tended to prioritize their past choices over past rewards on the IGT, indicated by high perseverance weight and low outcome frequency weight; and weighed more recent choices over distant ones on the IGT. These patterns were also observed in their unaffected siblings, indicating potential endophenotypes. Opioid users and their siblings displayed reduced loss sensitivity on the IGT and CGT, indicating that this might be an endophenotype specific to opioid addiction. Conclusions: This study provides insights into common and specific decision-making vulnerabilities of different types of drug addictions, some of which were observed in unaffected siblings, suggesting potential endophenotypes. Such computational decision-making profiles could increase the precision of neurocognitive assessment and aid in developing targeted intervention strategies.

Theories of individual differences are foundational to psychological and brain sciences, yet they are traditionally developed and tested using superficial summaries of data (e.g., mean response times) that are disconnected from our otherwise rich conceptual theories of behavior. To resolve this theory-description gap, we review the generative modeling approach, which involves formally specifying how behavior is generated within individuals, and in turn how generative mechanisms vary across individuals. Generative modeling shifts our focus away from estimating descriptive statistical "effects" toward estimating psychologically interpretable parameters, while simultaneously enhancing the reliability and validity of our measures. We demonstrate the utility of generative modeling in the context of the "reliability paradox," a phenomenon wherein replicable group effects (e.g., Stroop effect) fail to capture individual differences (e.g., low test-retest reliability). Simulations and empirical data from the Implicit Association Test and Stroop, Flanker, Posner, and delay discounting tasks show that generative models yield (a) more theoretically informative parameters, and (b) higher test-retest reliability estimates relative to traditional approaches, illustrating their potential for enhancing theory development. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Background:Inability to manage conflicts between multiple decision-making systems is associated with various mental disorders including addiction. The imbalance between the Pavlovian and instrumental systems may contribute to smoking cessation failure and relapse, as excessive Pavlovian motivation may override goal-directed quitting attempts. However, the exact role of the Pavlovian bias in smoking cessation and its neurocognitive mechanisms remain poorly understood. Methods:Eighty-six smokers participated in a smoking cessation clinic, completing the well-established orthogonalized go-nogo task before and after the intervention. We applied computational modeling and functional magnetic resonance imaging (fMRI) analyses to investigate the role of Pavlovian bias and its neural correlates in predicting smoking cessation outcomes. Results:Computational modeling revealed that higher levels of Pavlovian bias were associated with treatment failures. Importantly, Pavlovian bias significantly moderated the relationship between clinic participation rate and the likelihood of smoking cessation, suggesting that such bias can undermine motivational effort to quit. Furthermore, individuals who successfully quit smoking exhibited increased Pavlovian bias toward reward following cessation, indicating a potential cognitive mechanism underlying the relapse cycle. Complementary fMRI analyses demonstrated that Pavlovian bias influenced valence encoding in the ventromedial prefrontal cortex (vmPFC), suggesting altered decision-making neural circuits. Conclusions:Pavlovian bias can be a critical risk factor of relapse in the smoking population, while this bias can be increased in response to nicotine abstinence. Moreover, the results highlight the roles of the vmPFC in decision-making under Pavlovian-instrumental conflict, providing insights into the neural underpinnings of relapse vulnerability. Understanding these mechanisms may guide targeted interventions to enhance smoking cessation outcomes.

Background Tobacco smoking continues to be a leading cause of preventable morbidity and mortality globally, with the success rate of unaided cessation remaining consistently low. Understanding the neurobiological mechanisms of smoking cessation is crucial for improving quit rates. However, there has been a lack of studies examining brain network changes associated with smoking cessation over time. In this study, we aimed to investigate longitudinal changes in the functional connectivity (FC) of large-scale brain networks underlying smoking cessation outcomes using resting-state functional magnetic resonance imaging (fMRI). Methods A total of 98 treatment-seeking smokers participated in a 6-week cessation program and underwent resting-state fMRI scans before and after the intervention. Independent component analysis identified the salience network (SN), executive control network (ECN), and default mode network (DMN) components, and region of interest (ROI)-to-ROI FC was compared between successful and unsuccessful quitters using a group × time mixed-effects model. Correlations with smoking-related measures were explored. Results Significant group-by-time interaction effects were found in FC, particularly involving connections between SN and ECN, as well as between the SN and DMN. Specifically, successful quitters exhibited greater baseline FC in the SN-ECN and SN-DMN circuits, which tended to normalize during the cessation process. Exploratory correlational analyses revealed trends suggesting that stronger pre-quit connectivity between the SN and ECN was associated with greater withdrawal severity and longer smoking history in successful quitters. Conclusions Taken together, the normalization of initially elevated pre-quit FC in SN-ECN and SN-DMN circuits may reflect an adaptive neural process that supports successful withdrawal management and attentional reallocation during cessation. The identification of these neural substrates not only enhances our mechanistic understanding of smoking cessation over time but also underscores the need for targeted interventions that focus on these neural circuits to enhance quit outcomes. ### Competing Interest Statement The authors have declared no competing interest. National Research Foundation of Korea, RS-2018-NR031737, RS-2021-II211343, RS-2024-00435727, RS-2025-00516410 Seoul National University, Future Star Psychologist Award

Nicotine addiction is a complex disorder shaped by factors such as craving, mood, and neurocognitive processes. While ecological momentary assessment (EMA) provides a real-time method for capturing dynamic changes in behavior, traditional tasks and surveys are often too lengthy and demanding for repeated use in clinical settings. Integrating EMA with computational approaches offers a promising solution to predict smoking behavior dynamically while addressing the practical limitations of conventional assays, paving the way for more effective and scalable interventions. To evaluate the predictive value of computational markers derived from decision-making tasks and ecological momentary assessment (EMA) data for short-term (daily smoking behavior) and long-term (cessation success) outcomes, and to assess the timing and amount of data collection needed for prediction. 79 daily smokers (mean age 25.64 years, 83% male) took part in a longitudinal experimental study involving EMA surveys of psychological states and decision-making tasks, delivered daily via a smartphone app, while undergoing a 5–6 week smoking cessation program. Using a machine-learning methodology (adaptive design optimization, ADO) to effectively generate task variables, we estimated computational markers from just 20 to 30 trials per day, reducing task length and participants burden. A time-lagged model incorporating both computational markers and self-reported daily psychological states provided the most accurate prediction of next-day smoking behavior. Higher levels of craving, depression, and ambiguity tolerance in decision-making on the previous day were significantly predictive of increased smoking amount the following day. Smoking cessation status at the end of treatment was most strongly predicted by lower discounting rates, reduced craving and stress, a longer smoking history, and greater engagement in treatment (AUC = 0.76). Notably, models based on data collected during the first week of follow-up, either on the decision-making tasks (AUC = 0.74) or psychological variables (AUC = 0.73), demonstrated comparable predictive accuracy for end-of-treatment smoking cessation. Combining computational markers with EMA data offers a dynamic and efficient approach for predicting smoking behavior and cessation success and holds promise for clinical applications.

Background An imbalance between model-based and model-free decision-making systems is a common feature in addictive disorders. However, little is known about whether similar decision-making deficits appear in internet gaming disorder (IGD). This study compared neurocognitive features associated with model-based and model-free systems in IGD and alcohol use disorder (AUD). Method Participants diagnosed with IGD (n=22) and AUD (n=22), and healthy controls (n=30) performed the two-stage task inside the functional magnetic resonance imaging (fMRI) scanner. We used computational modeling and hierarchical Bayesian analysis to provide a mechanistic account of their choice behavior. Then, we performed a model-based fMRI analysis and functional connectivity analysis to identify neural correlates of the decision-making processes in each group. Results The computational modeling results showed similar levels of model-based behavior in the IGD and AUD groups. However, we observed distinct neural correlates of the model-based reward prediction error (RPE) between the two groups. The IGD group exhibited insula-specific activation associated with model-based RPE, while the AUD group showed prefrontal activation, particularly in the orbitofrontal cortex and superior frontal gyrus. Furthermore, individuals with IGD demonstrated hyper-connectivity between the insula and brain regions in the salience network in the context of model-based RPE. Discussion and Conclusions The findings suggest potential differences in the neurobiological mechanisms underlying model-based behavior in IGD and AUD, albeit shared cognitive features observed in computational modeling analysis. As the first neuroimaging study to compare IGD and AUD in terms of the model-based system, this study provides novel insights into distinct decision-making processes in IGD.

A major challenge in assessing psychological constructs such as impulsivity is the weak correlation between self-report and behavioral task measures that are supposed to assess the same construct. To address this issue, we developed a real-time driving task called the “highway task”, where participants often exhibit impulsive behaviors, such as reckless driving, thereby mirroring real-life impulsive traits captured by self-report surveys. Here, we first show that a self-report measure of impulsivity is highly correlated with performance in the highway task, but not with traditional behavioral task measures of impulsivity. By integrating deep neural networks with an inverse reinforcement learning (IRL) algorithm, we inferred dynamic changes of subjective rewards during the highway task. The IRL results indicated that impulsive participants attribute high subjective rewards to irrational or risky driving behaviors and situations. Overall, our results suggest that using real-time tasks combined with IRL can help reconcile the discrepancy between self-report and behavioral task measures of psychological constructs including impulsivity, with IRL being a practical modeling framework for multidimensional data from real-time tasks.

The decision-making framework and computational models have been instrumental in revealing the neurocognitive processes that underlie human behaviors, significantly enhancing our understanding of the human mind. However, traditional approaches often fall short in closely mimicking real-world behaviors due to their oversimplified nature, inability to capture the dynamic aspects of the human mind, and heavy reliance on secondary (e.g., monetary) rewards that have limited ecological utility. This article summarizes the evolution and recent developments of computational approaches in understanding human cognition and decision-making, with a particular focus on addiction. We highlight the transition from conventional laboratory-based methods to innovative approaches that more effectively simulate real-world scenarios. These advanced approaches leverage state-of-the-art computational techniques to capture dynamic psychological processes, contributing to the ongoing development and advances in the field.

Objectives: Addictions have recently been classified as substance use disorder (SUD) and behavioral addiction (BA), but the concept of BA is still debatable. Therefore, it is necessary to conduct further neuroscientific research to understand the mechanisms of BA to the same extent as SUD. The present study used machine learning (ML) algorithms to investigate the neuropsychological and neurophysiological aspects of addictions in individuals with internet gaming disorder (IGD) and alcohol use disorder (AUD). Methods: We developed three models for distinguishing individuals with IGD from those with AUD, individuals with IGD from healthy controls (HCs), and individuals with AUD from HCs using ML algorithms, including L1norm support vector machine, random forest, and L1-norm logistic regression (LR). Three distinct feature sets were used for model training: a unimodal-electroencephalography (EEG) feature set combined with sensor- and source-level feature; a unimodal-neuropsychological feature (NF) set included sex, age, depression, anxiety, impulsivity, and general cognitive function, and a multimodal (EEG + NF) feature set. Results: The LR model with the multimodal feature set used for the classification of IGD and AUD outperformed the other models (accuracy: 0.712). The important features selected by the model highlighted that the IGD group had differential delta and beta source connectivity between right intrahemispheric regions and distinct sensorlevel EEG activities. Among the NFs, sex and age were the important features for good model performance. Conclusions: Using ML techniques, we demonstrated the neurophysiological and neuropsychological similarities and differences between IGD (a BA) and AUD (a SUD).