Social cognition in young adults who endorse a cannabis use disorder

Article type: Research Article

Keywords: Cannabis, Cannabis use disorder, Social cognition, Dosage, Emotion

Authors: Gabrielle Abbott, Lisa-Marie Greenwood, Jessica G. Bartschi, Suraya Dunsford, Isabella Goodwin, Anastasia Paloubis, Marianna Quinones-Valera, Eugene McTavish, Antonio Verdejo-Garcia, Janna Cousijn, Gary C. K. Chan, Nadia Solowij, Valentina Lorenzetti ⚬

Affiliations: https://ror.org/00jtmb277grid.1007.60000 0004 0486 528XSchool of Psychology, Faculty of the Arts, Social Sciences, and Humanities, University of Wollongong, Wollongong, Australia; https://ror.org/019wvm592grid.1001.00000 0001 2180 7477School of Medicine and Psychology, The Australian National University, Canberra, Australia; https://ror.org/008n7pv89grid.11201.330000 0001 2219 0747School of Psychology, Faculty of Health, University of Plymouth, Plymouth, UK; https://ror.org/008n7pv89grid.11201.330000 0001 2219 0747Brain Research and Imaging Centre, Faculty of Health, University of Plymouth, Plymouth, UK; https://ror.org/04cxm4j25grid.411958.00000 0001 2194 1270Neuroscience of Addiction and Mental Health Program, Healthy Brain and Mind Research Centre, School of Health and Behavioural Sciences, Faculty of Health Sciences, Australian Catholic University, Level 5 Daniel Mannix Building, 115 Victoria Parade, Fitzroy, VIC 3065 Australia; https://ror.org/02bfwt286grid.1002.30000 0004 1936 7857School of Psychological Sciences and Turner Institute for Brain and Mental Health, Monash University, Melbourne, Australia; https://ror.org/057w15z03grid.6906.90000 0000 9262 1349Neuroscience of Addiction Lab, Department of Psychology, Education & Child Studies, Erasmus University Rotterdam, Rotterdam, The Netherlands; https://ror.org/00rqy9422grid.1003.20000 0000 9320 7537National Centre for Youth Substance Use Research, University of Queensland, Brisbane, Australia; https://ror.org/02jx3x895grid.83440.3b0000 0001 2190 1201Clinical Psychopharmacology Unit, Research Department of Clinical, Educational and Health Psychology, University College London, London, UK

License: © The Author(s) 2025 CC BY 4.0 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Article links: DOI: 10.1007/s00213-025-06890-z  |  PubMed: 40958032  |  PMC: PMC13035583

Relevance: Moderate: mentioned 3+ times in text

Full text: PDF (846 KB)

Introduction

Around 50 million individuals experience a cannabis use disorder (CUD) worldwide (Leung et al. ref. 2020). The global prevalence of CUD has increased by 32% from 1990 to 2019 (Shao et al. ref. 2023), and currently CUDs account for 49% of all substance use disorders worldwide (UNODC ref. 2024). These statistics are concerning, as CUD can be characterised by recurrent use despite difficulties with interpersonal, occupational, and/or social functioning as a consequence of cannabis consumption (DSM-5-TR ref. 2022). In line with cognitive theories of addiction, the psychosocial difficulties entrenched with CUD have been partly ascribed to altered social cognition (Koob ref. 2015; Koob and Volkow ref. 2016). Emerging evidence suggests a bidirectional relationship between altered social cognition and CUD, whereby impaired social cognition, such as diminished emotion recognition or differentiation, may contribute to persistent use which then further exacerbates these deficits, thus reinforcing the cycle of addiction (Goldstein and Volkow ref. 2011; Hammond et al. ref. 2022).

Social cognition encompasses a range of functions, from recognising faces to identifying and differentiating emotions (Arioli et al. ref. 2018). Social cognitive functions underpin the ability to understand others’ internal states (e.g., thoughts, emotions) and build/maintain interpersonal relationships (Arioli et al. ref. 2018; Hess ref. 2016). Notably, impairments in social cognition (e.g., emotion recognition/differentiation, face memory) have been implicated in the development and maintenance of substance use disorders, including CUD (Koob ref. 2015; Koob and Volkow ref. 2016) and mental health problems associated with CUD (e.g., anxiety, depression) (Baez et al. ref. 2023; Demenescu et al. ref. 2010; Onaemo et al. ref. 2021).

Early evidence shows altered social cognition in cannabis users (MacKenzie et al. ref. 2023). For example, cannabis users compared to controls, took more time to identify sadness, happiness, and anger (i.e., Dynamic Emotional Expression Recognition Task; DEER-T) (Platt et al. ref. 2010) and were less accurate as measured by hit rate and sensitivity (i.e. Emotion Processing Task) (Hindocha et al. ref. 2014). Similarly, individuals with CUD compared to controls have shown worse identification of negative emotions (e.g., sadness, shame) and worse discrimination of positive and negative emotions (Bayrakçı et al. ref. 2014) during the Facial Emotion Discrimination Test (FEDT) and Facial Emotion Identification Test (FEIT) (Bayrakçı et al. ref. 2014). Whereas another study found no group differences between CUD and controls in domains of social cognition including face memory and emotion recognition (Koenis et al. ref. 2021).

In addition, early evidence shows that level of problematic cannabis use may be associated with worse social cognition; with greater scores on the Cannabis Use Disorder Identification Test (CUDIT) being associated with poorer emotion recognition of sadness, fear, and anger (Blair et al. ref. 2021). Yet, other research in cannabis users found no significant correlations between frequency, dosage, duration of cannabis use, and emotion identification or emotion differentiation (Bayrakçı et al. ref. 2014). Minimal other research has examined the effect of dosage on social cognition which may be influenced by the lack of consensus for measuring quantities of use (e.g., joints per week, hits past year) (Bayrakçı et al. ref. 2014; Becker et al. ref. 2018).

There are other methodological limitations in the available evidence on social cognition in CUD. First, only a few studies to date have examined social cognition in regular cannabis users with CUD (Bayrakçı et al. ref. 2014; Blair et al. ref. 2021; Koenis et al. ref. 2021). Thus, the evidence base is inadequate to determine whether social cognition is a key feature of CUD or a byproduct of cannabis intoxication and subacute effects that should be accounted for in preventative interventions and treatment settings (Kroon et al. ref. 2020). Second, few studies have examined participants from the general community who endorse current CUD and who are currently using cannabis, which comprise the majority of people who use cannabis, as less than 15% are estimated to seek or receive treatment (Kerridge et al. ref. 2017). Indeed, CUD samples in previous studies were treatment seekers and often abstinent (Bayrakçı et al. ref. 2014; Blair et al. ref. 2021; Koenis et al. ref. 2021) which calls into question the generalisability of these findings to current users.

Third, few studies have examined other metrics of cannabis use in CUD such as whether level of cannabis related problems or dosage affects social cognition in those who endorse a CUD (Bayrakçı et al. ref. 2014; Blair et al. ref. 2021). Given previous early evidence that greater severity of problematic cannabis use (i.e., CUDIT) affects emotion recognition (Blair et al. ref. 2021), it is important to understand if this notion is replicable and extends to other aspects of social cognition (i.e., emotion differentiation) to identify who are the most vulnerable consumers. Finally, confounding variables entrenched with cannabis use that impact cognition independently and in interaction with cannabis use have not been consistently accounted for (e.g., alcohol and nicotine use, anxiety symptoms, hours since last cannabis use) (Copersino ref. 2017; Hofmann et al. ref. 2012; Koob ref. 2015; Koob and Volkow ref. 2016; Verdejo-Garcia et al. ref. 2019); and it is unclear if alterations are due to CUD or confounders entrenched with use. Importantly social cognitive deficits may negatively impact interpersonal relationships and increase the likelihood of social isolation (Tracy and Wallace ref. 2016). Therefore, new evidence in CUD is required to clarify the association between CUD and social cognition, to better understand how to support individuals who experience a CUD.

The primary aim of the current study was to determine whether distinct aspects of social cognition (i.e., emotion identification, emotion differentiation, immediate and delayed face memory) differ between regular cannabis users who endorse a CUD and controls, accounting for alcohol and nicotine use, and trait anxiety. The secondary aim of the study was to explore whether level of problematic cannabis use and dosage (i.e., CUDIT-R scores, cannabis grams/past month) is associated with social cognition (i.e., emotion identification, emotion differentiation, immediate and delayed face memory) in the CUD group, accounting for alcohol and nicotine use, trait anxiety, and hours since last cannabis use. Based on emerging evidence, it was hypothesised that the presence of a CUD, CUD-related problems and cannabis dosage, would be associated with worse performance in emotion differentiation and recognition. We explored the association between CUD, CUD-related problems and cannabis dosage and immediate/delayed face memory.

Methods

Recruitment

This study occurred within a larger project. Participants (N = 136) were recruited via advertisements posted around University of Wollongong campuses and online social media platforms. Informed consent was obtained from all participants and ethics was approved by the University of Wollongong Human Research Ethics Committee (HREC: 2017/389). Participants were reimbursed $40 AUD for completing the face-to-face testing session.

Participant inclusion and exclusion criteria

Inclusion criteria

Participants were required to be aged 18–55, fluent in English, and have an intelligence quotient (IQ) ≥ 80. Cannabis users were further required to (i) endorse Diagnostic and Statistical Manual for Mental Disorders – Fifth Edition (DSM-5-TR ref. 2022) criteria for a CUD (i.e., scoring ≥ 2 on the Structured Clinical Interview for DSM-5 – Research Version; SCID-5-RV) (First et al. ref. 2015), and (ii) currently use cannabis ≥ 4 days per week, and reported using at this level on average for 12-months prior to testing.

Exclusion criteria

Participants were excluded for the following reasons: (i) self-reported or met DSM-5 criteria on the Mini International Neuropsychiatric Interview (MINI-5) (Sheehan et al. ref. 1998) for current or past diagnosis of psychotic illness, bipolar disorder I and II, obsessive-compulsive disorder, current severe alcohol or other substance use disorder, or self-reported past treatment for alcohol use or drugs other than cannabis; (ii) self-reported high nicotine dependence on the Fagerström Test for Nicotine Dependence (FTND, ≥ 8) (Fagerström ref. 2012), or high risk for alcohol related harm on the Alcohol Use Disorder Identification Test (AUDIT, ≥ 20) (Babor et al. ref. 2001); (iii) currently pregnant or breastfeeding; (iv) lifetime use of illicit substances exceeding recreational levels (other than cannabis for CUD group), measured by the Drug History Questionnaire (DHQ) (Sobell et al. ref. 1995), defined as: weekly use for ≥ 3 months within the past 5 years; cumulative lifetime episodes > 50; and cumulative lifetime use > 5 for methamphetamine or any drug administered intravenously; (v) use of any illicit substance (other than cannabis for CUD group) within 30-days prior to testing; and (vi) known contraindications for cognition data interpretation (e.g., history of neurological illness such as encephalitis or moderate to severe brain injury). Controls were excluded if they reported (i) cannabis use within 3-months prior to testing; (ii) ≥ 50 lifetime cannabis use occasions, or (iii) a history of weekly or more frequent cannabis use over a ≥ 3-month period.

Participants abstained from using cannabis and alcohol 12-hours prior to testing to elucidate non-intoxicated residual effects of cannabis on cognition. A urine sample was collected for toxicology assessment (ECO II – Multi-Drug Screen: amphetamine, benzodiazepine, cocaine, cannabis, methamphetamine, opiate) to ensure the presence of delta-9-tetrahydrocannabinol (Δ9-THC) metabolites in the CUD sample, to detect other illicit substance use, and for pregnancy testing (Alere hCG Cassette). A breath analyser (Alcolizer Easy Check – HH1) was used to corroborate self-reported alcohol abstinence (BAC 0.00%). Any discrepancies between test results and reports of recent drug or alcohol use were clarified with participants. All inclusion and exclusion criteria were assessed prior to testing (described below) and confirmed at face-to-face testing.

Of the 136 participants recruited; seven endorsed cannabis use levels (i.e., frequency and duration of use) below the threshold for inclusion, and 14 did not complete the full face-to-face testing session due to identification of exclusionary issues or did not complete cognitive testing due to technical difficulties. This resulted in a final sample of 115 participants included for analysis. Twenty-nine participants reported past use of illicit substances (other than cannabis for the CUD sample) above defined recreational levels and/or within 30-days prior to testing (Supplementary Table 1). Visual and statistical inspection of distributions of social cognition outcome variables suggested these participants were not outliers or contributing to skewed data. As a result, these participants were retained in analyses to represent members of the community who consume cannabis on a regular basis and other illicit substances occasionally (Hasin and Walsh ref. 2020).

Procedure

Interested members of the community completed an online questionnaire and telephone screening to assess eligibility prior to the face-to-face testing session (see Supplementary Material section 1.1and 1.2). Eligible participants completed a 3-hour testing session at the research facilities located in the School of Psychology at the University of Wollongong. Participants completed semi-structured interviews that asked questions about current and past cannabis use, lifetime history of other substance use using the DHQ (Sobell et al. ref. 1995), recent drug use using the Timeline Follow-Back (TLFB) (Robinson et al. ref. 2014; Sobell and Sobell ref. 1992), and general demographic questions. The absence of exclusionary current and past psychopathology was reconfirmed using the MINI-5 (Sheehan et al. ref. 1998). A computer-based self-report questionnaire battery was administered to reassess levels of nicotine dependence (FTND; scores ≥ 8) (Fagerström ref. 2012) and problematic alcohol use (AUDIT; scores ≥ 20) (Babor et al. ref. 2001) and ensure they did not violate exclusion criteria, and to assess additional sample characteristics including schizotypal traits (i.e., Schizotypal Personality Questionnaire; SPQ) (Raine ref. 1991) and psychotic experiences (i.e., Community Assessment of Psychic Experiences; CAPE) (Stefanis et al. ref. 2002). The State-Trait Anxiety Inventory (STAI-Trait) (Spielberger ref. 1989; Spielberger et al. ref. 1983) assessed general trait anxiety and was used as a covariate in all analyses.

All cannabis users were re-administered the ‘alcohol and other substance use disorders’ module of the SCID-5-RV (First et al. ref. 2015) during the semi-structured interview to determine CUD symptoms. A semi-structured interview was used to ascertain cannabis use history and characterise current usage levels. From this information we extracted participants’ age at onset of regular use (i.e., age when consuming cannabis at least 4 days per week for 3-months or more). The TLFB was used to assess recent cannabis use metrics, specifically to determine total cannabis grams and use days in the month prior to testing, as well as hours since last cannabis use. The Cannabis Use Disorder Identification Test – Revised (CUDIT-R) (Adamson et al. ref. 2010; Adamson and Sellman ref. 2003) assessed problematic cannabis use patterns and was used as a predictor variable for analyses in the CUD sample. Full-scale IQ was measured via the Wechsler Abbreviated Scale of Intelligence – Second Edition (WASI-II), including the matrix reasoning and vocabulary subtests (Wechsler ref. 2011) which was analysed as a demographic variable. Following completion of interviews and self-report measures, all participants underwent social cognitive testing.

Social cognitive testing

Social cognition was assessed for a range of domains (i.e., emotion recognition, emotion differentiation, immediate and delayed face memory) using the adult Pennsylvania Computerised Neurocognitive Battery (Penn CNB) (Gur et al. ref. 2001, ref. 2010). The adult Penn CNB includes a number of subtests that assess various domains of cognition; four subtests capturing aspects of social cognition (Moore et al. ref. 2015) were administered as part of the current study. The four subtests were: (i) emotion recognition (ER): a measure of emotion identification where participants were shown a series of 40 faces and were asked to determine what emotion the face was showing by selecting one of the five emotion labels (i.e., happy, sad, anger, fear, no emotion) (ii) measured emotion differentiation (MED): a measure of emotion differentiation where participants were shown 40 pairs of faces and asked to determine whether one of the faces expressed the named emotion (i.e., happy or sad) more intensely or whether they were equal in emotional intensity, and (iii/iv) immediate face memory (FM) and delayed (FM-d): a measure of facial episodic memory where participants were shown 20 faces and were then asked to identify those faces in a series of 40 faces (i.e., 20 target stimuli and 20 novel) at an immediate and delayed (15–30 min later) recall interval.

The Penn CNB was administered in a standardised lab environment with a consistent set-up for desk height, and mouse, keyboard and monitor size (i.e., 62 centimetres, measured diagonally from bottom left corner to top right corner). For all Penn CNB subtests, accuracy (i.e., total number or percent correct) and speed scores (i.e., median response time, milliseconds) were transformed to their standardised equivalent (i.e., z-score) and combined to calculate an efficiency score, consistent with the literature (Merikangas et al. ref. 2017; Moore et al. ref. 2015; Van Pelt et al. ref. 2021). All z-scores for median response time were multiplied by −1 to produce a speed value where, as for accuracy, higher scores reflect better performance.

Statistical analyses

All analyses were performed using IBM SPSS Statistics (Version 29) (IBM Corp., ref. 2022), R (Version 4.4.0) (R Core Team, ref. 2024), and Jamovi (Version 2.6.0) (The Jamovi Project, 2024). The level of significance was set at p <.05. Normality checks of descriptive and outcome data, data handling, and assumption testing is outlined in Supplementary Materials section 1.4.

Primary aim: social cognition differences between CUD and control group

We ran a series of one-way analyses of covariance (ANCOVA) using group (CUD vs. control) as a predictor, social cognition efficiency scores as outcome variables that were normally distributed, including emotion recognition (i.e., ER); emotion differentiation (i.e., MED); and immediate face memory (i.e., FM). A robust regression was used to examine group differences for delayed face memory (i.e., FM-d) which had a skewed distribution. All analyses were run with and without adjusting for covariates including nicotine use (i.e., FTND), alcohol use (i.e., AUDIT), and trait anxiety (i.e., STAI-Trait).

Secondary aim: CUDIT-R and cannabis grams/past month predicting social cognition in the CUD group

In the CUD group, regression analyses were used to determine whether the predictor variables, level of problematic cannabis use (i.e., CUDIT-R scores) and cannabis dosage (i.e., cannabis grams/past month), predicted outcome variables that were normally distributed including: emotion recognition (i.e., ER); emotion differentiation (i.e., MED); immediate face memory (i.e., FM). A robust regression was run to examine the effect of CUDIT-R and cannabis grams/past month as predictors on delayed face memory (i.e., FM-d) due to its skewed distribution. All analyses were run with and without adjusting for covariates including nicotine use (i.e., FTND), alcohol use (i.e., AUDIT), trait anxiety scores (i.e., STAI-Trait), as well as number of hours since last cannabis use.

Sensitivity analyses

To confirm the robustness of results, sensitivity analyses were conducted to examine the primary and secondary aims, excluding outliers defined as +/- 3 standard deviations from the mean (see Supplementary Table 2).

Results

Sample descriptives

Sample characteristics are shown in Table 1. This study included a sample of 115 participants (82 male, 33 female) with a mean age of 23.2 years. Of these, 83 participants endorsed a CUD and 32 were controls. The CUD group compared to controls had significantly fewer years of education, lower IQ, and showed significantly higher trait anxiety, schizotypal personality traits, and psychotic experiences. The CUD group also had significantly greater FTND and AUDIT scores, as well as reported consuming a higher number of standard alcoholic drinks and cigarettes in the previous month.

Cannabis use levels

The CUD group reported First using cannabis at 16.7 years. Regular use, defined as 4 or more days per week, commenced at 19.9 years. The CUD group also reported using cannabis on an average of 24.5 out of 30 days per month, consumed an average of 20 g of cannabis in the past month, and average THC-COOH urine levels of 64 ng/ml confirm the presence of Δ9-THC metabolites. The CUD group reported that their last cannabis use occurred 14.5 h (median) prior to testing.

Group comparison (CUD vs. controls) in social cognition

As shown in Table 2, groups did not differ significantly for social cognition outcomes, accounting for nicotine and alcohol use, and trait anxiety. Specifically, CUD and controls performed similarly for emotion recognition (p =.859, d = 0), where results remained non-significant after removing outliers (p =.988). Similarly, there were no significant differences between the CUD and control group for emotion differentiation (p =.121, d = 0.314, outlier removal sensitivity analysis: p =.194) or immediate face memory (p =.635, d = 0.090, outlier removal sensitivity analysis: p =.494). The robust regression examining delayed face memory found no significant differences between the CUD and control group (p =.273) (see Supplementary Material section 2.1.1).

The removal of covariates resulted in a significant change to the model for emotion differentiation (p =.012, d =.487), whereby the CUD group performed significantly worse than controls (MD = 0.707, SE = 0.275). No covariates were highly correlated with emotion differentiation (ρ = 0–0.21). Results from group comparisons examining other domains of social cognition (i.e., emotion recognition, immediate and delayed face memory) remained non-significant after the removal of covariates (see Supplementary Table 3).

CUDIT-R scores and cannabis dosage as predictors of social cognition in the CUD group

There was no significant effect of CUDIT-R scores and cannabis grams/past month on emotion recognition, emotion differentiation, or immediate and delayed face memory (p >.1). All results remained unaltered with and without covariates or outliers (see Supplementary Material, section 2.2.1 and Supplementary Tables 4 & 5).

Discussion

The current study examined social cognition in non-treatment seeking regular cannabis users who endorsed a CUD and non-using controls. This study also examined social cognition in the CUD group in relation to level of problematic cannabis use (i.e., CUDIT-R scores) and cannabis dosage (i.e., cannabis grams/past month), comprehensively accounting for confounders (i.e., nicotine/alcohol use, trait anxiety, and hours from last cannabis use). In contrast to the study hypotheses, which were based on neuroscientific theories of addiction (Goldstein and Volkow ref. 2011; Koob ref. 2015), there was no significant effect of CUD, level of problematic cannabis use, or dosage on emotion differentiation or recognition, nor other domains of social cognition (i.e., immediate and delayed face memory), with effect sizes ranging from very small to small (i.e., d = 0–0.314). These results may indicate that social cognition is not a core feature of CUD or that the relationship between CUD and social cognition is influenced by moderators such as age, treatment seeking, education level, and IQ. Indeed, the current CUD sample included mostly young adults (M = 23.6 years) who were university students at the time of testing and had above average IQ and no participant was seeking or receiving treatment for their cannabis use. It may be that young age, access to tertiary education, high cognitive function and the lack of seeking/receiving treatment and protected participants against social cognitive alterations. Future work is required to determine if social cognition is affected in adolescents and older adults with a CUD who are seeking or undergoing treatment.

The lack of significant social cognitive differences between the CUD and control group align with a previous study that reported no group differences for some of the same measures examined herein (i.e., immediate and delayed face memory, emotion recognition) (Koenis et al. ref. 2021). However, our results contrast with other research that found worse emotion recognition and differentiation performance in regular cannabis users (Hindocha et al. ref. 2014; Platt et al. ref. 2010) and former cannabis users with CUD (Bayrakçı et al. ref. 2014) compared to controls. It is possible that inconsistent accounting for confounds known to be comorbid with cannabis use and CUD (e.g., alcohol use, nicotine use, depression and anxiety scores) across studies has contributed to mixed findings in the literature. Indeed, in the current study when covariates (i.e., alcohol and nicotine use, trait anxiety) were removed for a sensitivity analysis comparing emotion differentiation between CUD and controls, an effect emerged whereby CUD performance was significantly lower than controls. Further, two of the three studies that reported significantly worse social cognition in people who use cannabis compared to controls failed to account for potential confounding variables in analyses (i.e., depression symptoms) (Bayrakçı et al. ref. 2014; Platt et al. ref. 2010). The third study accounted for sex and schizotypy scores in sensitivity analyses, however, did not covary depression scores or alcohol use which may have produced an effect when entered into statistical models due to being significantly higher in the cannabis group (Hindocha et al. ref. 2014). Our results highlight the importance of accounting for potential confounders when examining social cognition in CUD. Without our conservative approach to analyses, we may have inadvertently overstated cannabis-related effects on emotion differentiation, which may explain findings in previous studies that used less conservative statistical approaches.

Similarly, in the CUD group there was no significant relationship between level of problematic cannabis use (i.e., CUDIT-R scores) or dosage (i.e., cannabis grams/past month) and social cognition. The role of severity of use and cannabis use levels is an understudied area of the literature, which limits the interpretation of findings. The limited previous research found no association between dosage (i.e., joints/week) and emotion identification and discrimination (Bayrakçı et al. ref. 2014), which is consistent with the current findings. However, our findings contrast with another study that found significant associations between CUDIT scores and emotion recognition (i.e., sadness, fear, and anger) in a sample of abstinent individuals with CUD (Blair et al. ref. 2021). It may be that different characteristics between the current CUD sample and those in previous studies, such as treatment seeking, comorbidities (e.g., psychopathology, other substance use) education, and IQ, explain the inconsistent findings. First, the two previous studies that examined associations between cannabis use metrics and social cognition, utilised treatment seeking samples (Bayrakçı et al. ref. 2014; Blair et al. ref. 2021) that endorsed psychiatric and other substance use comorbidities (Blair et al. ref. 2021). It may be that comorbidities may moderate social cognitive alterations however, this was not examined in the current study due to our stringent exclusion criteria, implemented to ensure sound methodology to detect cannabis-specific effects, resulting in a non-treatment seeking sample with minimal comorbidities.

Second, the current study assessed CUD symptoms and level of problematic cannabis use using robust methodology – including but not limited to SCID-5-RV for confirming the presence of a CUD, CUDIT-R to measure CUD-related problems, and timeline follow-back to measure recent exposure to cannabis. The CUD sample examined were not seeking or receiving treatment and therefore future work is required to elucidate if treatment status moderates the association between CUD and social cognition performance. Finally, our CUD sample had high levels of education and an average IQ (M = 116, SD = 10.7) one standard deviation above the mean which may indicate sampling bias. Although our study had a scope of regular cannabis users aged 18–55 years, our recruitment advertising was mostly concentrated to university campuses and associated online social media pages. It is likely that this contributed to a sample of more highly educated (M = 15 years, SD = 1.88 years), young (M age = 23.6 years), university students whose cannabis use may not be as severely impacting functioning compared to those seeking treatment. Indeed, CUD samples in previous research had substantially lower years of education (e.g., M = 8.3 years, SD = 2.6 years) (Bayrakçı et al. ref. 2014) and IQ (M = 101.69, SD = 11.65) (Blair et al. ref. 2021) compared to the current CUD sample. It may be that cannabis use has minimal effect on social cognition in people with higher IQ and related socioeconomic characteristics; however, this needs to be determined via more research.

Limitations

First, our CUD sample characteristics may indicate the presence of sampling bias. Indeed, the CUD group had above average IQ, were young, and had minimal psychopathology and other substance use comorbidities. It may be that these findings do not reflect characteristics of social cognitive functioning of the wider CUD population with older ages or youth, treatment seekers and polysubstance users. To ensure research is generalisable to CUD populations, future research should recruit participants with a range of educational experiences and psychosocial characteristics (e.g., comorbid psychopathology, other substance use) and examine the impact of such factors on social cognition outcomes. Second, the cross-sectional design of the current study prevented exploration of whether social cognitive alterations emerge or dissipate with continued cannabis use and/or changes in chronicity of use, and severity of CUD; a notion to be confirmed by longitudinal designs. Third, participants endorsed a range of CUD severities (i.e., mild, moderate, severe) and it may be that social cognitive deficits only emerge in severe dependence. This was tested via our secondary aim which found no significant association between CUDIT-R scores and social cognition thus there does not appear to be a significant association between CUD severity and social cognitive outcomes. However, more research is required to replicate these findings. Fourth, our study had an insufficient sample size to examine the moderating effects of other sample characteristics (e.g., schizotypal traits, SPQ and psychotic experiences, CAPE) shown to be associated with social cognition (Weinreb et al. ref. 2022). It may be that other psychopathology variables moderate the association between cannabis use and social cognition. Future studies should confirm the role of which moderators affect the association between CUD and social cognition using larger samples. Fifth, although the current study used a semi-structured interview and validated assessment tool (i.e., TLFB) to measure cannabis use metrics, we did not utilise a standardised questionnaire to measure cannabis use nor did we test samples of the cannabis consumed by participants to determine Δ9-THC concentration. Future research should utilise validated semi-structured interviews such as the Daily Sessions, Frequency, Age of Onset, and Quantity of Cannabis Use Inventory (DFAQ-CU) alongside the TLFB and cannabis sample testing for Δ9-THC concentration to enhance methodological robustness and reproducibility.

Conclusion

The current study found that social cognition (i.e., emotion recognition, emotion differentiation, immediate and delayed face memory), did not differ between CUD and controls, nor did level of problematic cannabis use (i.e., CUDIT-R) or cannabis dosage (i.e., cannabis grams/past month) significantly predict social cognition in the CUD group, accounting for confounds. These findings may indicate that altered social cognition is not a key feature of CUD, or that the relationship between CUD and social cognition is moderated by factors such as age, treatment seeking, education, IQ, and comorbidities (e.g., psychopathology, other substance use) which were unable to be examined in the current sample due to possible sampling bias. Indeed, the current CUD sample were young, non-treatment seeking, endorsed no major comorbidities, and had an above average IQ which may have served as a protective factor against social cognitive alterations; however, these results need to be confirmed by replication studies. Future research is required that examines social cognition in vulnerable CUD samples from the general community and treatment services who endorse a broader range of ages, educational experiences, IQ, and comorbidities. Such new research will provide insight into whether social cognition is a feature of CUD, or more vulnerable CUD populations, which will support the development of clinical interventions to mitigate potential interpersonal and social challenges.

References

1. SJ Adamson, FJ Kay-Lambkin, AL Baker, TJ Lewin, L Thornton, BJ Kelly, JD Sellman. An improved brief measure of cannabis misuse: the cannabis use disorders identification Test-Revised (CUDIT-R). Drug Alcohol Depend, 2010. [DOI | PubMed]
2. SJ Adamson, JD Sellman. A prototype screening instrument for cannabis use disorder: the cannabis use disorders identification test (CUDIT) in an alcohol-dependent clinical sample. Drug Alcohol Rev, 2003. [DOI | PubMed]
3. M Arioli, C Crespi, N Canessa. Social cognition through the lens of cognitive and clinical neuroscience. Biomed Res Int, 2018. [DOI]
4. Babor TF, Higgins-Biddle JC, Saunders JB, Monteiro MG (2001) AUDIT: the alcohol use disorders identification test: guidelines for use in primary health care. World Health Organization
5. S Baez, MA Tangarife, G Davila-Mejia, M Trujillo-Güiza, DA Forero. Performance in emotion recognition and theory of mind tasks in social anxiety and generalized anxiety disorders: a systematic review and meta-analysis. Front Psychiatry, 2023. [DOI]
6. Bayrakçı A, Sert E, Zorlu N, Erol A, Sarıçiçek A, Mete L (2014) Facial emotion recognition deficits in abstinent cannabis dependent patients. Compr Psychiatr 58. 10.1016/j.comppsych.2014.11.008
7. MP Becker, PF Collins, A Schultz, S Urošević, B Schmaling, M Luciana. Longitudinal changes in cognition in young adult cannabis users. J Clin Exp Neuropsychol, 2018. [DOI | PubMed]
8. RJR Blair, J Bashford-Largo, R Zhang, A Mathur, A Schwartz, J Elowsky, P Tyler, CJ Hammond, FM Filbey, M Dobbertin, S Bajaj, KS Blair. Alcohol and cannabis use disorder symptom severity, conduct disorder, and callous-unemotional traits and impairment in expression recognition. Front Psychiatry, 2021. [DOI | PubMed]
9. ML Copersino. Cognitive mechanisms and therapeutic targets of addiction. Curr Opin Behav Sci, 2017. [DOI | PubMed]
10. LR Demenescu, R Kortekaas, JA den Boer, A Aleman. Impaired attribution of emotion to facial expressions in anxiety and major depression. PLoS ONE, 2010. [DOI]
11. DSM-5-TR (2022) Diagnostic and statistical manual of mental disorders: fifth edition, text revision. American Psychiatric Association Publishing
12. K Fagerström. Determinants of tobacco use and renaming the FTND to the Fagerstrom test for cigarette dependence. Nicotine Tob Research: Official J Soc Res Nicotine Tob, 2012. [DOI]
13. RZ Goldstein, ND Volkow. Dysfunction of the prefrontal cortex in addiction: neuroimaging findings and clinical implications. Nat Rev Neurosci, 2011. [DOI | PubMed]
14. RC Gur, JD Ragland, PJ Moberg, TH Turner, WB Bilker, C Kohler, SJ Siegel, RE Gur. Computerized neurocognitive scanning: I. Methodology and validation in healthy people. Neuropsychopharmacology, 2001. [DOI | PubMed]
15. RC Gur, J Richard, P Hughett, ME Calkins, L Macy, WB Bilker, C Brensinger, RE Gur. A cognitive neuroscience-based computerized battery for efficient measurement of individual differences: standardization and initial construct validation. J Neurosci Methods, 2010. [DOI | PubMed]
16. Hammond C, Allick A, Park G, Rizwan B, Kim K, Lebo R, Nanavati J, Parvaz M, Ivanov I (2022) A meta-analysis of fMRI studies of youth cannabis use: alterations in executive control, social cognition/emotion processing, and reward processing in cannabis using youth. Brain Sci 12(10). 10.3390/brainsci12101281
17. D Hasin, C Walsh. Cannabis use, cannabis use disorder, and comorbid psychiatric illness: A narrative review. J Clin Med, 2020. [DOI | PubMed]
18. Hindocha C, Wollenberg O, Leno VC, Alvarez BO, Curran HV, Freeman TP (2014) Emotional processing deficits in chronic cannabis use: A replication and extension. 28:466–471. 10.1177/0269881114527359
19. SG Hofmann, KK Ellard, GJ Siegle. Neurobiological correlates of cognitions in fear and anxiety: A cognitive-neurobiological information-processing model. Cogn Emot, 2012. [DOI | PubMed]
20. IBM Corp (2022) IBM SPSS statistics for windows (Version 29.0) [Computer software]. IBM Corp
21. BT Kerridge, PM Mauro, SP Chou, TD Saha, RP Pickering, AZ Fan, BF Grant, DS Hasin. Predictors of treatment utilization and barriers to treatment utilization among individuals with lifetime cannabis use disorder in the united States. Drug Alcohol Depend, 2017. [DOI | PubMed]
22. MMG Koenis, J Durnez, AL Rodrigue, SR Mathias, AF Alexander-Bloch, JA Barrett, GE Doucet, S Frangou, EEM Knowles, J Mollon, D Denbow, K Aberizk, M Zatony, RJ Janssen, JE Curran, J Blangero, RA Poldrack, GD Pearlson, DC Glahn. Associations of cannabis use disorder with cognition, brain structure, and brain function in African Americans. Hum Brain Mapp, 2021. [DOI | PubMed]
23. GF Koob. The dark side of emotion: the addiction perspective. Eur J Pharmacol, 2015. [DOI | PubMed]
24. GF Koob, ND Volkow. Neurobiology of addiction: a neurocircuitry analysis. Lancet Psychiatry, 2016. [DOI | PubMed]
25. E Kroon, L Kuhns, E Hoch, J Cousijn. Heavy cannabis use, dependence and the brain: a clinical perspective. Addiction, 2020. [DOI | PubMed]
26. Leung J, Chan GCK, Hides L, Hall WD (2020) What is the prevalence and risk of cannabis use disorders among people who use cannabis? A systematic review and meta-analysis. Addict Behav 109. 10.1016/j.addbeh.2020.106479
27. A MacKenzie, A Cservenka, K Dunn, W Stoops. Cannabis and emotion processing: A review of behavioral, physiological, and neural responses. Exp Clin Psychopharmacol, 2023. [DOI | PubMed]
28. Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA (1983) Manual for the state-trait anxiety inventory
29. AK Merikangas, L Cui, ME Calkins, TM Moore, RC Gur, RE Gur, KR Merikangas. Neurocognitive performance as an endophenotype for mood disorder subgroups. J Affect Disord, 2017. [DOI | PubMed]
30. TM Moore, SP Reise, RE Gur, H Hakonarson, RC Gur. Psychometric properties of the Penn computerized neurocognitive battery. Neuropsychology, 2015. [DOI | PubMed]
31. Hess U (2016) Nonverbal communication. In H. S. Friedman (Ed.), Encyclopedia of Mental Health (Second Edition) (pp. 208–218). Academic Press. 10.1016/B978-0-12-397045-9.00218-4
32. VN Onaemo, TO Fawehinmi, C D’Arcy. Comorbid cannabis use disorder with major depression and generalized anxiety disorder: A systematic review with meta-analysis of nationally representative epidemiological surveys. J Affect Disord, 2021. [DOI | PubMed]
33. AE Van Pelt, RS Beidas, JC Scott, TM Moore, CV Ahmed, KH Morales, B Thuto, O Tshume, RC Gur, JH Holmes, M Matshaba, ED Lowenthal. Acceptability of a computerized neurocognitive battery to identify cognitive impairments among children and adolescents in Botswana. Global Implement Res Appl, 2021. [DOI]
34. B Platt, S Kamboj, CJA Morgan, HV Curran. Processing dynamic facial affect in frequent cannabis-users: evidence of deficits in the speed of identifying emotional expressions. Drug Alcohol Depend, 2010. [DOI | PubMed]
35. R Core Team (2024). R: A language and environment for statistical computing. R Foundation for statistical computing, Vienna, Austria. https://www.R-project.org/
36. A Raine. The SPQ: A scale for the assessment of schizotypal personality based on DSM-III-R criteria. Schizophr Bull, 1991. [DOI | PubMed]
37. SM Robinson, LC Sobell, MB Sobell, GI Leo. Reliability of the timeline followback for cocaine, cannabis, and cigarette use. Psychol Addict Behav, 2014. [DOI | PubMed]
38. H Shao, H Du, Q Gan, D Ye, Z Chen, Y Zhu, S Zhu, L Qu, J Lu, Y Li, J Duan, Y Gu, M Chen. Trends of the global burden of disease attributable to cannabis use disorder in 204 countries and territories, 1990–2019: results from the disease burden study 2019. Int J Mental Health Addict, 2023. [DOI]
39. DV Sheehan, Y Lecrubier, KH Sheehan, P Amorim, J Janavs, E Weiller, T Hergueta, R Baker, GC Dunbar. The Mini-International neuropsychiatric interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry, 1998. [PubMed]
40. LC Sobell, E Kwan, MB Sobell. Reliability of a drug history questionnaire (DHQ). Addict Behav, 1995. [DOI | PubMed]
41. Spielberger CD (1989) State-Trait anxiety inventory. In. Consulting Psychologists
42. N Stefanis, M Hanssen, N Smirnis, D Avramopoulos, I Evdokimidis, C Stefanis, H Verdoux, J van Os. Evidence that three dimensions of psychosis have a distribution in the general population. Psychol Med, 2002. [DOI | PubMed]
43. First MB, Williams JBW, Karg RS, Spitzer RL (2015) Structured Clinical Interview for DSM-5 – Research Version (SCID-5 for DSM-5, Research Version); SCID-5-RV, Version 1.0.0)
44. Sobell LC, Sobell MB (1992) Timeline Follow-Back. In: Litten RZ, Allen JP (eds) Measuring alcohol consumption: psychosocial and biochemical methods. Humana, pp 41–72. 10.1007/978-1-4612-0357-5_3
45. K Tracy, SP Wallace. Benefits of peer support groups in the treatment of addiction. Subst Abuse Rehabilitation, 2016. [DOI]
46. A Verdejo-Garcia, G Garcia-Fernandez, G Dom. Cognition and addiction. Dialog Clin Neurosci, 2019. [DOI]
47. Wechsler D (2011) Wechsler Abbreviated Scale of Intelligence – Second Edition. Pearson
48. S Weinreb, F Li, MM Kurtz. A meta-analysis of social cognitive deficits in schizophrenia: does world region matter?. Schizophr Res, 2022. [DOI | PubMed]
49. UNODC (2024) World Drug Report. United Nations Publication, 2024