Evidence Accumulation Efficiency Predicts Substance Use Better Than Inhibitory Control

Redefining the Neurocognitive Markers of Substance Use Risk

Clinicians have long associated substance use disorders with deficits in inhibitory control, often viewing the inability to suppress impulsive actions as a primary driver of addiction [1]. This neurocognitive framework suggests that individuals at high risk for substance misuse exhibit reduced activation in specific inhibitory brain networks, a pattern also observed in comorbid conditions like attention-deficit/hyperactivity disorder [2]. Beyond substance use, similar impairments in motor impulsivity are documented in behavioral addictions, such as pathological gambling, where patients struggle to cancel inappropriate responses [3]. Current neuroimaging research continues to explore how these regulatory failures in the brain's inhibitory circuits overlap with heightened reward salience to escalate drug-seeking behavior [4]. However, the reliability of traditional inhibitory metrics as standalone predictors of clinical outcomes remains a subject of active investigation. A new study now offers fresh insights into which specific cognitive mechanisms most accurately signal vulnerability to substance use.

Limitations of Traditional Inhibitory Metrics

Clinical assessments of addiction risk frequently focus on poor inhibitory control and decision-making, which are established risk factors for substance use and other adverse psychiatric outcomes. To quantify these deficits, clinicians and researchers often rely on the Stop-Signal Task (SST), a widely used protocol designed to index an individual's ability to suppress a motor response after it has been initiated. The conventional metric derived from this task is the stop signal reaction time (SSRT), a calculated value representing the time required for the brain to cancel a planned action. While the stop signal reaction time has served as a standard measure of inhibitory control for decades, the researchers behind the current study argue that previous models of this metric may be too simplistic to capture the complex, multifaceted cognitive processes that occur during task performance. To address these limitations, the study utilized the Racing Diffusion Ex-Gaussian ABCD (RDEX-ABCD) model, a mechanistic framework that provides a more granular decomposition of cognitive performance. Unlike traditional metrics that provide a single score for inhibition, the RDEX-ABCD model captures both specific inhibitory control and task-general decision-making processes by accounting for variables such as the speed of evidence accumulation and the probability of failing to initiate a response. By applying this model to data from the IMAGEN cohort, which included more than 1,000 participants at ages 19 and 23, the authors demonstrated that parameters indexing inhibitory control had no associations with substance use and were only weakly linked to brain connectivity. Instead, the findings indicated that broader decision-making mechanisms, such as the efficiency of evidence accumulation (the rate at which the brain processes information to reach a decision) and the decision threshold (the amount of information required before committing to a response), served as more reliable indicators of vulnerability to cannabis and cigarette use.

Longitudinal Analysis of the IMAGEN Cohort

The researchers conducted a longitudinal analysis using data from the IMAGEN cohort, a large scale multicenter study of adolescent brain development and mental health. This specific analysis included a sample size of more than 1,000 individuals (n > 1000) who were assessed at two distinct developmental stages: age 19 and age 23. To move beyond traditional metrics of impulsivity, the study applied the Racing Diffusion Ex-Gaussian ABCD (RDEX-ABCD) model to data obtained from the Stop-Signal Task. This computational model allows for a more nuanced decomposition of cognitive performance by separating the speed of inhibitory processes from general decision-making efficiency, providing a more mechanistic view of how a patient processes the instruction to stop an action. To identify which cognitive factors most accurately predicted clinical outcomes, the authors utilized Elastic Net regression (a regularized statistical method that identifies the most significant predictors within a large dataset while minimizing the risk of overfitting). This analysis examined the relationship between specific model parameters and the participants' reported substance use. Furthermore, the researchers employed connectome-based predictive modeling (a technique that uses whole-brain functional connectivity patterns to predict individual behavioral traits) to identify the specific neural networks that underpin these cognitive parameters. By mapping these functional connections, the study sought to determine if specific brain architectures could serve as biological markers for addiction risk. The study then examined the association between these identified brain networks and substance use patterns. The findings revealed that parameters indexing inhibitory control had no associations with substance use and were only weakly linked to brain connectivity. In contrast, the neural networks associated with general decision-making processes, such as the efficiency of evidence accumulation and the probability of go failure (the likelihood of failing to initiate a response when required), were significant predictors of cannabis and cigarette use. For the practicing clinician, these results suggest that a patient's vulnerability to substance misuse may be more closely tied to how efficiently their brain processes information to reach a decision rather than their ability to suppress a motor impulse.

Evidence Accumulation as a Superior Predictor

The analysis of the IMAGEN cohort data revealed a notable disconnect between traditional measures of impulse suppression and clinical outcomes. Specifically, the researchers found that parameters indexing inhibitory control had no associations with substance use among the participants at age 19 or 23. This lack of correlation suggests that the speed of a motor stop signal may not be the most relevant metric for assessing addiction risk in a clinical setting. This finding was further supported by the neurological data, which showed that parameters indexing inhibitory control were only weakly associated with brain connectivity, indicating that the functional networks responsible for simple motor inhibition do not strongly overlap with the circuits involved in substance use vulnerability. In contrast to the specific inhibitory metrics, the study identified several broader cognitive factors that served as significant predictors of cannabis and cigarette use. These parameters reflecting general decision-making processes included efficiency of evidence accumulation, a measure of how effectively an individual integrates information to reach a decision. The researchers also found that parameters reflecting general decision-making processes included decision threshold (the level of response caution or the amount of information required before acting), as well as the probability of go failure (the likelihood of failing to initiate a response to a stimulus). These findings suggest that the underlying neurocognitive vulnerability to substance use is rooted in how a patient processes information and sets thresholds for action, rather than a simple deficit in their ability to cancel an ongoing motor response.

Clinical Implications for Risk Stratification

The identification of specific neurocognitive markers provides a more precise framework for assessing addiction risk in clinical practice. The researchers found that general decision-making parameters and their associated brain activity were significant predictors of cannabis use and also served as significant predictors of cigarette use among the cohort of more than 1,000 individuals. These parameters, derived from the Racing Diffusion Ex-Gaussian ABCD model, offer a mechanistic view of how a patient processes information before acting. Specifically, the efficiency of evidence accumulation emerged as a robust predictor of substance use vulnerability, suggesting that the speed and quality of information integration are more central to addiction risk than previously understood. This efficiency is defined as a neurocognitive mechanism that facilitates adaptive decision making across many contexts, allowing individuals to weigh options and potential consequences effectively before committing to a choice. For the practicing clinician, these findings suggest a necessary shift in how cognitive risk is evaluated during patient screenings. The study indicates that general decision-making mechanisms may act as more reliable indicators of vulnerability to substance use than conventional inhibitory control measures, such as the stop signal reaction time. While traditional models focused on a patient's ability to suppress an impulse, this data demonstrates that the underlying vulnerability often lies in the preceding stage of evidence processing. Because the efficiency of evidence accumulation is a fundamental process for adaptive choices in various life scenarios, its impairment may signal a broad susceptibility to substance misuse. Consequently, diagnostic focus should move toward assessing how patients accumulate and process information, as these task-general cognitive processes show a stronger correlation with long-term substance use outcomes and neural connectivity than simple motor inhibition.

References

1. Hildebrandt MK, Dieterich R, Endrass T. Neural correlates of inhibitory control in relation to the degree of substance use and substance-related problems - A systematic review and perspective.. Neuroscience and biobehavioral reviews. 2021. doi:10.1016/j.neubiorev.2021.06.011

2. Adisetiyo V, Gray KM. Neuroimaging the neural correlates of increased risk for substance use disorders in attention-deficit/hyperactivity disorder-A systematic review.. The American journal on addictions. 2017. doi:10.1111/ajad.12500

3. Chowdhury NS, Livesey EJ, Blaszczynski A, Harris JA. Pathological Gambling and Motor Impulsivity: A Systematic Review with Meta-Analysis.. Journal of gambling studies. 2017. doi:10.1007/s10899-017-9683-5

4. Hinojosa CA, Sitar SI, Zhao JC, et al. Functional Domains of Substance Use and their Implications to Trauma: A Systematic Review of Neuroimaging Studies. Chronic Stress. 2024. doi:10.1177/24705470241258752