James Murphy’s research while affiliated with University of Memphis and other places

The tension reduction hypothesis posits that people consume alcohol to alleviate stress and negative emotions. Prior experimental studies supported this hypothesis by showing stress-induced increases in alcohol's absolute value. However, from a value-based decision-making perspective alcohol's value relative to alternatives should be more relevant for drinking decisions. We developed a novel experimental paradigm that subjected the hypothesis to a more stringent test by examining whether acute stress causes individuals to choose alcohol over appealing non-alcoholic alternatives, and the cognitive mechanisms that underpin this. Participants (N=160) rated various drinks and made repeated choices between alcoholic and non-alcoholic options before and after randomized manipulations of stress and alcohol intoxication (BrAC=.06%). Using drift diffusion modeling, we decomposed choices into three potential mechanisms: how carefully people make decisions, their sensitivity to prior drink prefer-ences, and their bias to choose alcohol regardless of preference. Results showed that stress moder-ately increased choices for alcohol, but only in sober participants. Stress primarily affected decision-making by inducing a bias toward alcohol during evidence accumulation, without impacting decision carefulness or general evidence sensitivity. This computational bias was stronger than observed in raw choice behavior, suggesting that while stress consistently biases evaluation toward alcohol, this bias only sometimes overcomes competing considerations. The boundary condition that stress effects appeared only in sober participants suggests the tension reduction hypothesis may primarily explain initial drinking decisions rather than choices about continued consumption. These findings advance our mechanistic understanding of how stress influences alcohol-related decision-making and suggest interventions might focus on strengthening competing motivations during high-stress moments, particularly during decisions about initiating drinking episodes.

The tension reduction hypothesis posits that people consume alcohol to alleviate stress and negative emotions. Prior experimental studies supported this hypothesis by showing stress-induced increases in alcohol's absolute value. However, from a value-based decision-making perspective alcohol's value relative to alternatives should be more relevant for drinking decisions. We developed a novel experimental paradigm that subjected the hypothesis to a more stringent test by examining whether acute stress causes individuals to choose alcohol over appealing non-alcoholic alternatives, and the cognitive mechanisms that underpin this. Participants (N=160) rated various drinks and made repeated choices between alcoholic and non-alcoholic options before and after randomized manipulations of stress and alcohol intoxication (BrAC=.06%). Using drift diffusion modeling, we decomposed choices into three potential mechanisms: how carefully people make decisions, their sensitivity to prior drink prefer-ences, and their bias to choose alcohol regardless of preference. Results showed that stress moder-ately increased choices for alcohol, but only in sober participants. Stress primarily affected decision-making by inducing a bias toward alcohol during evidence accumulation, without impacting decision carefulness or general evidence sensitivity. This computational bias was stronger than observed in raw choice behavior, suggesting that while stress consistently biases evaluation toward alcohol, this bias only sometimes overcomes competing considerations. The boundary condition that stress effects appeared only in sober participants suggests the tension reduction hypothesis may primarily explain initial drinking decisions rather than choices about continued consumption. These findings advance our mechanistic understanding of how stress influences alcohol-related decision-making and suggest interventions might focus on strengthening competing motivations during high-stress moments, particularly during decisions about initiating drinking episodes.

Purpose: The co-occurrence of cannabis use and internalizing problems, such as depression and anxiety, during emerging adulthood (18-25 years) is well documented, but the directionality is unclear. This study investigates the bidirectional longitudinal relationships between cannabis use (frequency and consequences) and internalizing problems (depressive and anxiety symptoms) among high-risk emerging adults. Methods: Data came from 7 assessment waves collected over a 2-year period among 900 (56% female) high-risk emerging adults participating in two longitudinal cohorts (Ontario, Canada; Tennessee, USA). Latent curve models with structured residuals were used to explore bidirectional between- and within-person relationships between cannabis-related variables and internalizing problems. Results: At baseline, higher cannabis use frequency and consequences were both associated with higher internalizing symptoms. In between-person models, both cannabis-related factors (frequency and consequences) and internalizing symptoms (depressive and anxiety symptoms) decreased across emerging adulthood. In within-person models, bidirectional relationships were observed (partially supporting both symptom-driven and substance-induced pathways), but cannabis consequences, rather than frequency, were most consistently related to subsequently higher internalizing problems (particularly depressive symptoms). These bidirectional relationships were more pronounced among females and those surpassing clinical thresholds for internalizing problems at baseline. Conclusions: This study found evidence of bidirectional relationships between cannabis involvement and internalizing problems across emerging adulthood, although the prevailing direction was from negative cannabis consequences to increases in internalizing. These findings highlight the importance of cannabis intervention in emerging adults, both to reduce consequences and prevent internalizing symptoms, especially targeting females and those with clinically elevated internalizing problems.

Background. Behavioral economic research has revealed significant increases in alcohol demand following exposure to alcohol-related cues. Prior research has focused exclusively on non-treatment-seeking heavy drinkers, included only male participants, or used heterogeneous methods. The current studies sought to replicate and extend existing findings in treatment-seeking and non-treatment seeking heavy drinkers while also examining sex effects and moderation by AUD severity. Methods. Study 1 included 117 non-treatment seeking heavy drinkers (51.5% women; M age 34.69; 56.4% AUD+), and Study 2 included 89 treatment-seeking heavy drinkers with AUD (40.4% women; M age = 41.35). In both studies, alcohol demand was measured using a state-based alcohol purchase task (APT) and subjective alcohol craving was measured using visual analog scales. Measures were collected following exposure to neutral (water) cues in a standard room and alcohol cues in a bar lab.Results. Alcohol demand (intensity, Omax, breakpoint, and elasticity) and craving were significantly increased following alcohol cues compared to neutral cues (ps < .005) with effect sizes ranging from small to large (partial eta squared = .074-.480). Participants with AUD (Study 1) or with higher AUD severity (Study 2) reported higher craving and higher demand for most indices (i.e., main effects; ps < .032, partial eta squared = .043-.239), and a larger alcohol cue increase in Omax was found for AUD+ participants in Study 1 compared to non-AUD (p = .028, partial eta squared = .041). There were no significant sex effects. Conclusions. These findings replicate and extend prior research by offering additional insight into alcohol cue effects on the reinforcing value of alcohol and subjective motivation to drink. Results also suggest that the presence of an AUD may amplify cue effects on maximum alcohol expenditure but not for other indices of alcohol demand.

Background and aims: The Marijuana Purchase Task (MPT) is increasingly used to measure cannabis reinforcing value and has potential utility for cannabis etiological and regulatory research. This meta-analysis sought to evaluate for the first time the MPT's concurrent validity in relation to cannabis involvement. Methods: Electronic databases and pre-print repositories were searched for MPT studies that examined the cross-sectional relationship between frequency and quantity of cannabis use, problems, dependence and five MPT indicators: intensity (i.e., unrestricted consumption), Omax (i.e., maximum consumption), Pmax (i.e., price at which demand becomes elastic), breakpoint (i.e., first price at which consumption ceases), and elasticity (i.e., sensitivity to rising costs). Random effects meta-analyses of cross-sectional effect sizes were conducted, with Q tests for examining differences by cannabis variables, meta-regression to test quantitative moderators, and publication bias assessment. Moderators included sex, number of MPT prices, variable transformations, and year of publication. Populations included community and clinical samples. Results: The searches yielded 14 studies [(N= 4,077, median % females: 44.8%: weighted average age= 29.08 (SD: 6.82)], published between 2015-2022. Intensity, Omax and elasticity showed the most robust concurrent validity [|r's|= .147-325, ps <.014] with the largest significant effect sizes for quantity [|r| intensity= .325] and cannabis dependence [|r|Omax = .320,|r|intensity= .305,|r|elasticity= .303]. Higher proportion of males was associated with increased estimates for elasticity-quantity, and Pmax -problems. Higher number of MPT prices significantly altered magnitude of effects sizes for Pmax and problems, suggesting biased estimations if excessively low prices are considered. Methodological quality was generally good and minimal evidence of publication bias was observed. Conclusions: The Marijuana Purchase Task appears to have concurrent validity to quantify cannabis demand, most robustly for intensity, Omax and elasticity. Moderating effects by sex suggest potentially meaningful sex differences in the reinforcing value of cannabis.

Objective: A substantial number of people reduce their consumption of alcohol in the absence of formal treatment; however, less is known about the mechanisms of change. The aim of this study is to explore whether constructs derived from behavioral economics and computational decision-modeling characterize the moderation of alcohol consumption that many heavy drinkers experience without treatment. Method: Case-control, pre-registered design. People who reside in the United Kingdom and who drink heavily (n = 60) or used to drink heavily but now consume alcohol in moderation (n = 60) were recruited. Participants completed self-report behavioral economic measures (alcohol demand and alcohol-related and alcohol-free reinforcement), and a two-alternative forced choice task in which they chose between two alcoholic (in one block) or two soft-drink images (in a different block). A drift-diffusion model was fitted to responses from this task to yield the underlying parameters of value-based choice. Results: Compared to heavy drinkers, moderated drinkers had significantly lower alcohol demand (Omax, p = .03, Cohen’s d = .36; elasticity, p = .03, rank-biserial correlation (rrb) = .21) and higher proportionate alcohol-free reinforcement (p < .001, Cohen’s d = .75). However, contrary to hypotheses, there were no robust between-group differences in VBDM parameters. Conclusions: Self-report behavioral economic measures demonstrate that alcohol moderation without treatment is characterized by lowered alcohol demand, and greater behavioral allocation to alcohol-free reinforcement, in line with behavioral economic theory. However, a computerized VBDM measure yielded inconclusive findings.

Objective: A substantial number of people reduce their consumption of alcohol in the absence of formal treatment; however, less is known about the mechanisms of change. The aim of this study is to explore whether constructs derived from behavioral economics and computational decision-modeling characterize the moderation of alcohol consumption that many heavy drinkers experience without treatment. Method: Case-control, pre-registered design. People who reside in the United Kingdom and who drink heavily (n = 60) or used to drink heavily but now consume alcohol in moderation (n = 60) were recruited. Participants completed self-report behavioral economic measures (alcohol demand and alcohol-related and alcohol-free reinforcement), and a two-alternative forced choice task in which they chose between two alcoholic (in one block) or two soft-drink images (in a different block). A drift-diffusion model was fitted to responses from this task to yield the underlying parameters of value-based choice. Results: Compared to heavy drinkers, moderated drinkers had significantly lower alcohol demand (Omax, p = .03, Cohen’s d = .36; elasticity, p = .03, rank-biserial correlation (rrb) = .21) and higher proportionate alcohol-free reinforcement (p < .001, Cohen’s d = .75). However, contrary to hypotheses, there were no robust between-group differences in VBDM parameters. Conclusions: Self-report behavioral economic measures demonstrate that alcohol moderation without treatment is characterized by lowered alcohol demand, and greater behavioral allocation to alcohol-free reinforcement, in line with behavioral economic theory. However, a computerized VBDM measure yielded inconclusive findings.

Background: Individuals who report greater meaning in life tend to consume less alcohol. However, research elucidating pathways through which meaning in life influences consumption is lacking. Behavioural economic theories posit that distortions in valuation processes, whilst negative reinforcement models posit that avoidance or regulation of negative internal states, are central in decisions to consume alcohol. Method: Pre-registered, cross-sectional design. Five hundred forty-six participants (≥ 18 years old) completed an online questionnaire which asked about their alcohol use and related problems, meaning in life, alcohol-free reinforcement, alcohol value, depressive symptoms, and drinking to cope motives. Results: Presence of meaning in life had a significant negative association with AUDIT scores (β = -.26, p < .001), but neither search for meaning in life nor total alcohol-free reinforcement were significant predictors (ps > .53). Subsequent path analyses using structural equation modelling revealed a significant indirect effect of presence of meaning in life on AUDIT scores through lower alcohol value (95% CI = -.17 to -.08) and drinking to cope (95% CI = -.07 to -.00). Although search for meaning in life was not a direct predictor of AUDIT scores, there was a significant indirect effect through greater drinking to cope (95% CI = .01 to .06). Conclusions: Presence of meaning in life predicts lower alcohol consumption indirectly via individual differences in alcohol value and drinking to cope. These findings provide insight into mechanisms that underpin the relationships between meaning in life and alcohol consumption.

Background: Individuals with greater meaning in life tend to consume less alcohol. However, research elucidating pathways through which meaning in life influences consumption is lacking. Behavioural economic theories posit that distortions in valuation processes, whilst negative reinforcement models posit that avoidance or regulation of negative internal states, are central in decisions to consume alcohol. Method: Pre-registered, cross-sectional design. Five hundred forty-six regular alcohol consumers (≥ 18 years old) completed an online questionnaire which asked about alcohol use, meaning in life, alcohol-free reinforcement, alcohol value, depressive symptoms, and drinking to cope motives. Results: Presence of meaning had a significant negative association with AUDIT scores (β = -.26, p < .001), but neither search for meaning nor total alcohol-free reinforcement were significant predictors (ps > .53). Subsequent path analyses revealed a significant indirect effect of presence of meaning on AUDIT scores through lower alcohol value (95% CI = -.17 to -.08) and drinking to cope (95% CI = -.07 to -.00), and a serial mediation effect through both lower depressive symptoms and drinking to cope (95% CI = -.09 to -.04). Although search for meaning was not a direct predictor of AUDIT scores, there was a significant indirect effect through greater drinking to cope (95% CI = .01 to .06) and a serial mediation effect through both greater depressive symptoms and drinking to cope (95% CI = .01 to .04). Conclusions: Meaning in life subscales predict alcohol consumption indirectly via individual differences in alcohol value, depressive symptoms, and drinking to cope.