The therapeutic effect of mindfulness interventions on problematic drinking is thought to be driven by increased resilience to the impact of stress on negative mood and alcohol-seeking behaviour, but this claim needs empirical support. To address this hypothesis, the current study tested whether brief training of one component of mindfulness – breath counting – would reduce drinkers’ sensitivity to the effect of noise stress on subjective mood and alcohol-seeking behaviour. Baseline alcohol-seeking was measured by choice to view alcohol versus food thumbnail pictures in 192 student drinkers. Participants then received a 6-minute audio file which either trained breath counting or recited a popular science extract, in separate groups. All participants were then stressed by a loud industrial noise and alcohol-seeking was measured again simultaneously to quantify the change from baseline. Subjective mood was measured after all three stages (baseline, post intervention, post stress test). The breath counting group were instructed to deploy this technique during the stress test. Results showed that the breath counting versus control intervention improved subjective mood relative to baseline, attenuated the worsening of subjective mood produced by stress induction, and accelerated recovery from a stress induced increase in alcohol-seeking behaviour. Exploratory moderation analysis showed that this accelerated recovery from stress induced alcohol-seeking by breath counting was weaker in more alcohol dependent participants. Mindfulness therapies may improve problematic drinking by increasing resilience to stress induced negative mood and alcohol-seeking, as observed in this study. The weaker therapeutic effect of breath counting in more dependent drinkers may reveal limitations to this intervention strategy. Questionnaires Participants completed the following questionnaires. The adult Patient-Reported Outcomes Measurement Information System Alcohol Use Short Form (PROMIS; Pilkonis et al., 2016) which contains 7 items assessing loss of control over drinking in the past 30 days (e.g., “I drank more than I planned”), endorsed on a 1–5 scale ranging from “Never” to “Always” (we report the average scale scores). The Alcohol Use Disorder Identification Test (AUDIT; Babor, Higgins-Biddle, Saunders, & Monteiro, 2001) which contains 10 items assessing the frequency of alcohol use and alcohol-related problems experienced in the past 12 months. Total scores can range from 0 to 40 split into categories: low-risk (0–7), hazardous (8–15), harmful (16–19) and possibly dependent (20–40). The modified five factor Drinking Motives Questionnaire Revised (DMQR; Grant, Stewart, O'Connor, Blackwell, & Conrod, 2007), which measures how frequently drinking is motived by each listed reason, on a 1–10 scale ranging from “Never” to “Almost always”. It has five subscales: drinking to cope with anxiety and depression, conformity, enhancement and socialising (the two coping subscales were collapsed). The Generalised Anxiety Disorder (GAD; Spitzer, Kroenke, Williams, & Lowe, 2006) scale which contains 7 items assessing generalised anxiety disorder in the past two weeks (e.g., “feeling nervous, anxious or on edge”). The score on each item ranges from 0 (“Not at all”) to 3 (“Nearly every day”). The total score can range from 0 to 21, with a score of 5, 10, and 15 as the cut-off points for mild, moderate and severe anxiety, respectively. The Patient Health Questionnaire depression scale (PHQ; Kroenke et al., 2009) which contains 8 items assessing depressive symptoms in the past two weeks (e.g., “little interest or pleasure in doing things”). The score on each item ranges from 0 (“Not at all”) to 3 (“Nearly every day”). The total score can range from 0 to 24, with a score of 5, 10, 15 and 20 as the cut-off points for mild, moderately severe and severe depression, respectively. Baseline alcohol choice Alcohol pictorial choice was measured at baseline by participants completing 24 two-alternative forced-choice trials in which they freely chose to enlarge thumbnail pictures of either alcohol or food by pressing a left or right arrow key (Hardy & Hogarth, 2017). Instructions were: ‘In this task, you can view alcohol and food pictures by pressing the left or right arrow key’. In each trial, the alcohol and food thumbnail stimuli presented were each sampled from a set of 28 pictures, and presented randomly in the left or right screen position. The dependent variable was the percentage choice of alcohol across all choice trials. Following baseline alcohol choice, subjective mood was measured, at the baseline timepoint, by asking participants to what extent they currently felt happy and annoyed, in random order, on a 5-point scale ranging from 1 (“not at all”) to 5 (“extremely”). Breath counting versus control intervention The half of participants who were assigned to the breath counting group listened to a 6-minute audio file (inspired by Ramsburg & Youmans, 2014) in which they were instructed (via a female voice) to relax and concentrate on their breath sensations, then count each outbreath, at normal pace, from one to ten, and then start again from one (see Supplementary material for full transcript and the audio file). The half of participants who were assigned to the control group received a 6-minute audio file in which was recited (by the same female voice as the breath counting audio) an extract from the popular science book A Short History of Nearly Everything by Bill Bryson. For both groups, after the audio file, participants were asked how much attention they had paid to the recording on a scale ranging from 1 (‘a little’) to 5 (‘a lot’), and how pleasant they had found to the experience on a scale ranging from 1 (‘Unpleasant’) to 5 (‘Pleasant’). Finally, all participants had their subjective happiness and annoyance measured at this post-intervention timepoint (identical to the baseline timepoint). The stress-induced alcohol choice test All participants then completed an alcohol pictorial choice task identical to baseline, except a loud and unpleasant industrial noise (70 dB; file: airsander.mp3 from www.freesfx.co.uk) was played continuously through headphones over 36 trials, to induce mild stress and augment alcohol choice (Cherek, 1985). The 36 trials of the test phase were broken into three time bins of 12 trials each to examine changes over time. The breath counting group were instructed to deploy the breath counting technique during the stress test, whereas the control group received no comparable instruction. All participants reported their subjective happiness and annoyance identical to the baseline and post intervention timepoints. Finally, the breath counting group reported their attention to and pleasantness of the breath technique deployed during the stress test, identical to the post-intervention timepoint. In the end, all participants completed a mood repair procedure (Hardy & Hogarth, 2017) to normalise mood prior to departure (for ethical reasons).The first aim of the fellowship was to build on my PhD research to develop a novel brief intervention for hazardous drinking young people. My proposed intervention combines a number of elements with prior evidence of efficacy in hazardous alcohol use. High-risk individuals will be provided with personalised feedback regarding the specific negative emotions which trigger their drinking (for example, anger, sadness, boredom) (Blevins and Stephens 2016) and encouraged to generate individualised alternative coping strategies (Conrod et al. 2013). Crucially, individuals will also be instructed in functional imagery training, a promising technique used to encourage adoption of adaptive behaviours in high-risk scenarios. The two pilot studies proposed in clinical populations in this project are on hold due to COVID-19 restrictions. However, development of this intervention has continued and online pilot testing in student populations is ongoing. Data will be uploaded as and when these projects are complete. Additional aims of the fellowship included: 1) To publish completed research demonstrating that a brief, 6 minute mindfulness training procedure can reduce drinking under stress in students. This research has been published in the journal Addictive Behaviours and a full dataset is uploaded here. 2) To publish research validating a novel measure of negative coping motives (the Coping Motives Checklist). This measure will be used to identify specific negative triggers to alcohol use - and to provide personalised feedback on these motives - in my proposed novel intervention. Initial data linkage and analysis is ongoing and data will be uploaded when this is finalised. 3) Validation of a measure of alcohol valuation - the Brief Assessment of Alcohol Demand (BAAD) - to be used as a brief screening tool to identify those at risk of dependence. This research has been published in Addictive Behaviours and a full dataset is uploaded here.

192 participants, who had drunk at least once in the past month and were therefore not teetotal, were recruited from the University of Exeter student population (age range: 18–52 years) and were randomly assigned to either the breath counting group or control group. This was therefore a convenience method of sampling. Participants provided informed consent, were debriefed and received a chocolate bar as the reimbursement for participation. The study was approved by the School of Psychology Research Ethics Committee.