2026 ECNP Congress

PS01-1252

Computational modelling reveals long-lasting decision-making impairment in patients with cocaine use disorder

Mia Beastall — T.V. Lim, K.D. Ersche · University of Cambridge, Department of Psychiatry, Cambridge, United Kingdom

2026 ECNP Congress

Introduction: Patients with cocaine use disorder (CUD) make disadvantageous decisions with adverse personal and social consequences, as reflected in impaired performance on laboratory-based decision-making tasks [1]. Despite the substantial societal costs of these poor decisions [2], it remains unclear whether decision-making deficits recover following prolonged drug abstinence. While standard measures of lab-based decision-making paradigms reflect observable performance, computational analysis of behaviour enables us to characterise more precisely the cognitive subprocesses underlying decision-making impairments in CUD patients [3]. We therefore applied computational modelling to the Cambridge Gambling Task [4] to investigate whether these processes provide cross-sectional evidence of recovery in abstinent CUD patients. Based on preclinical longitudinal evidence that structural alterations in brain regions implicated in decision-making remain following prolonged abstinence [5], we hypothesised that decision-making impairments in cocaine addiction persist following abstinence.

Methods: We recruited CUD patients actively using cocaine (N=28), CUD patients abstinent from cocaine for on average two years (M=23.2 months; SD=29; N=21), and healthy volunteers without a history of cocaine use or any substance use disorder (N=53). All participants completed the Cambridge Gambling Task (https://cambridgecognition.com/), making choices between colour options with known outcome probabilities and placing bets on their choices. Trial-by-trial task performance was fitted to a validated computational model [3] deconstructing decision-making into six latent parameters measuring core decisional components related to probability weighting, value (i.e., loss aversion, bet determinism, colour bias) and decisional impulsivity (i.e., risk preference, delay aversion). Parameters were estimated within a hierarchical Bayesian framework using RStan. Group differences on each parameter were evaluated using mean differences (d) with 95% highest density intervals (HDI); a 95% HDI excluding zero (P≥0.95) indicated credible differences.

Results: Groups were matched for age and sex. The two groups of CUD patients did not differ in their duration of regular cocaine use (t(29.96)=-1.41,p=0.168). Compared with healthy volunteers, all CUD patients, regardless of abstinence status, showed reduced probability weighting (dcurrent=−2.01, 95%HDI [−2.98,−1.00], P>0.99; dabstinent=−2.14, 95%HDI [−3.20,−1.07], P>0.99), impairments in valuation of options (reduced loss aversion: dcurrent=−0.56, 95%HDI [−1.09,−0.04], P=0.97; dabstinent=−0.56, 95%HDI [−1.12,−0.01], P=0.96; reduced bet determinism: dcurrent=−8.02, 95%HDI [−11.2,−4.85], P>0.999; dabstinent=−7.36, 95%HDI [−10.8,−3.95], P>0.99), and increased decisional impulsivity (increased risk preference: dcurrent=0.38, 95%HDI [0.15,0.64], P>0.99; dabstinent=0.198, 95%HDI [−0.003,0.414], P=0.95; increased delay aversion: dcurrent=0.09, 95%HDI [0.01,0.17], P=0.98; dabstinent=0.10, 95%HDI [0.02,0.19], P=0.98). Colour bias did not differ credibly in either CUD group (0∈95%HDI). Importantly, no credible differences were observed between current and abstinent CUD patients across any parameter (0∈95%HDI). Correlational analysis showed that the longer patients were abstinent from cocaine, the more sensitive they were to losses (r=0.625,pFDR=0.012).

Conclusion: Computational modelling indicates that decision-making impairments in cocaine addiction show only partial recovery with prolonged abstinence, suggesting that additional support may be needed for patients during recovery. The observed positive association between loss sensitivity and the duration of abstinence further points to a drug-related effect on the commonly reported blunted response to negative feedback in chronic cocaine users. Overall, these findings highlight the added value of computational modelling over standard analytical approaches and support its role in advancing precision psychiatry.

References

  1. Verdejo-García, A., Chong, T.T.-J., Stout, J.C., Yücel, M., London, E.D., 2018. Stages of dysfunctional decision-making in addiction. Pharmacol. Biochem. Behav. 164, 99-105.
  2. Schwartz, E.K.C., Wolkowicz, N.R., De Aquino, J.P., MacLean, R.R., Sofuoglu, M., 2022. Cocaine use disorder (CUD): current clinical perspectives. Subst. Abuse Rehabil. 13, 25-46.
  3. Lim, T.V., Cardinal, R.N., Ziauddeen, H., Regenthal, R., Sahakian, B.J., Robbins, T.W., Ersche, K.D., 2025. Atomoxetine reduces decisional impulsivity in human cocaine addiction. Biol. Psychiatry 97 (6), 627-636.
  4. Rogers, R.D., Everitt, B.J., Baldacchino, A., Blackshaw, A.J., Swainson, R., Wynne, K., Baker, N.B., Hunter, J., Carthy, T., Booker, E., London, M., Deakin, J.F., Sahakian, B.J., Robbins, T.W., 1999. Dissociable deficits in the decision-making cognition of chronic amphetamine abusers, opiate abusers, patients with focal damage to prefrontal cortex, and tryptophan-depleted normal volunteers: evidence for monoaminergic mechanisms. Neuropsychopharmacology 20 (4), 322-339.
  5. Jedema, H.P., Song, X., Aizenstein, H.J., Bonner, A.R., Stein, E.A., Yang, Y., Bradberry, C.W., 2021. Long-term cocaine self-administration produces structural brain changes that correlate with altered cognition. Biol. Psychiatry 89 (4), 376-385.