Pavlovian-instrumental transfer in alcohol dependenceSchad et al.
Supplementary Information
Group details
Table S1. Group details (alcohol-dependent patients and healthy controls at baseline assessment)
Alcohol-dependent Patients / Healthy Controls / p-valueGender / ♀:4, ♂:27 / ♀:3, ♂:21 / .96 a
Mean / SD / Mean / SD / p
Age in years / 45.29 / 11.43 / 42.17 / 11.16 / .31 b
Craving (OCDS-G total score) / 13.03 / 9.42 / 3.21 / 3.23 / <.01c
Socio-economic status (SES) / -0.36 / 1.92 / 0.38 / 1.84 / .23 b
Verbal intelligence (MWT-B) / 105.87 / 10.79 / 102.27 / 10.02 / .22 b
Smokers / 87.1 % / 66.7 % / .069a
Severity of AD (ADS) / 14.23 / 7.5 / 2.83 / 3.87 / <.001c
Lifelong alcohol consumption in kg (pure alcohol)d / 2017.48 / 1005.21 / 178.89 / 141.40 / .001c
Note. OCDS-G – Obsessive Compulsive Drinking Scale, German version [1]; SES – Socioeconomic status (computed as the sum of z-transformed self-ratings of social status, household income, and inverse personal debt scores). MWT-B – Mehrfachwahl-Wortschatz-Intelligenztest[10] (German multiple-choice vocabulary intelligence test, where subjects are repeatedly asked to select the correct German word among a list of nonsense words); Smoking – smoking status (Are you currently smoking?) assessed with the Fagerström Test for Nicotine Dependence [FTND; 5]; ADS – Alcohol Dependence Scale [12]. Note that the table reports results for the same subjects reported before [4], with a few changes in the reported values. Compared to the previous report, we were able to reassess some of the missing values (one for ADS and three for SES) and estimated lifetime drinking slightly differently in accordance with recent developments in converting CIDI drinking assessments[8, 9].aχ2-test. bt-test. cWilcoxon rank-sum test.dprior to detoxification in alcohol-dependent patients, measured by the CIDI [7, 13].
Instrumental responding during training
Figure S1. Instrumental responding (number of button presses) during instrumental conditioning scaled in six blocks. As the length of the task was individually different dependent on task performance, we show the relative block number for each subject, respectively. Both, patients and controls, learned to enhance number of button presses for good shells (“Collect”) and to reduce number of button presses to leave bad shells (“Leave”).
Instrumental responding during Pavlovian-instrumental transfer
Figure S2.Instrumental responding during PIT for drink‐related stimuli and a neutral CS (0 EUR)for healthy controls (HC) vs. AD (cf. Fig. 2a), low vs. high severity of dependence (ADS, cf. Fig. 3a), and abstainers vs. relapsers (relapse vs. abstinence assessed six weeks after detoxification; cf. Fig. 4a). Number of button presses are displayed for water and alcohol pictures, as well as for CSs conditioned to predict 0 EUR reward (neutral CSs).
Figure S3. Instrumental responding (number of button presses) during PIT task independent of background stimuli. Both groups show stable instrumental behavior during the task despite nominal extinction.
Covariance analyses
To further elucidate our results regarding possible confounding variables, we conducted analyses of covariance (ANCOVA).Firstly, we tested whether the observed differences in behavioral and neural PIT effects between patients and controls, low versus high severity dependence, and relapsers versus abstainers (week 6) were still reliable after controlling for participant’s sex. For five out of these six comparisons, the observed differences were stable (significant or marginal difference, p < .1) when controlling for sex differences. Only the difference in the neural PIT effect between relapsers and abstainers (week 6) was not significant any more (p .1). Secondly, wetested the effects of severity of dependence and relapse (abstainers versus relapsers; week 6) simultaneously, controlling for both influences in a single ANCOVA model. We found that PIT differed between low versus high severity of dependence after controlling for relapse (behavioural PIT: p= .03; marginal effect for neural PIT: p = .095), but that the relapse effect was not significant any more (p > .1).
Testing an interference account
It is one possibility that alcohol pictures interfered with instrumental task execution (rather than reducing approach via general PIT). Indeed, alcohol interference studies have found[2, 3, 6, 11]that alcohol pictures enhance response latencies to the target and can reduce accuracies. Based on these previous findings, we predicted that an interference effect of alcohol pictures should be visible in a reduced instrumental go/nogo effect, a reduced instrumental accuracy, and/or in a reduced latency of the first response. We tested these effects in the present data. We found that the difference in response rates between instrumental collect versus leave conditions (i.e., the instrumental go/nogo effect;p=.97) and the percentage of correct instrumental responses (i.e., collecting “good” shells, and leaving “bad” shells; reflecting instrumental accuracy;p= .71) did not significantly differ between alcohol versus water pictures. Likewise, the latency of the first response per trial did not differ between alcohol versus water pictures (p= .14). These key results dismiss the assumption that alcohol pictures interfered with instrumental task performance, and instead suggests that the aversive features of alcohol cues reduce instrumental approach via general PIT.
Debriefing
To further investigate the interference account we studied debriefing questions.
Assessment: Immediately after the PIT paradigmwas conducted, we asked subjects whether they agreed with the following statements: “I felt distracted by the alcohol/water stimuli in the background”, “The alcohol/water stimuli in the background worried me”, and “The alcohol/water background stimuli indicated to collect the corresponding shells”.
Analysis: We analyzed the questions regarding possible distractions by the drink-related background stimuli using χ2-tests comparing the frequency of “yes” and “no” answers between patients and controls, relapsers and abstainers and high and low severe patients. We aimed to exclude possible distraction effects that could explain our behavioral results.
Results: Two data sets (1 AUD, 1 HC) were not collected, resulting in a sample of 53 subjects for these analyses. Drink stimuli resulted in minimal self-reported distraction(water: 2 subjects; alcohol: 15 subjects) or worry (water: 3 subjects; alcohol: 6 subjects). The majority thought that the background stimuli were irrelevant to the choice of whether to collect shells (41/53 for water background and 40/53 for alcohol background). Importantly, we did not find any group differences regarding these questions (AD vs. HC: ps > .21; relapsers vs. abstainers: ps > .15; low ADS vs. high ADS patients: ps.33). These results further speak against systematic biases arising from task disengagement or complex strategies.
References
1.Anton RF, Moak DH, Latham P (1995) The obsessive compulsive drinking scale: A self-rated instrument for the quantification of thoughts about alcohol and drinking behavior. Alcoholism, clinical and experimental research 19:92-99
2.Baxter BW, Hinson RE (2001) Is smoking automatic? Demands of smoking behavior on attentional resources. Journal of abnormal psychology 110:59-66
3.Cox WM, Hogan LM, Kristian MR, Race JH (2002) Alcohol attentional bias as a predictor of alcohol abusers' treatment outcome. Drug and alcohol dependence 68:237-243
4.Garbusow M, Schad DJ, Sebold M, Friedel E, Bernhardt N, Koch SP, Steinacher B, Kathmann N, Geurts DE, Sommer C, Muller DK, Nebe S, Paul S, Wittchen HU, Zimmermann US, Walter H, Smolka MN, Sterzer P, Rapp MA, Huys QJ, Schlagenhauf F, Heinz A (2016) Pavlovian-to-instrumental transfer effects in the nucleus accumbens relate to relapse in alcohol dependence. Addiction biology
5.Heatherton TF, Kozlowski LT, Frecker RC, Fagerstrom KO (1991) The fagerstrom test for nicotine dependence: A revision of the fagerstrom tolerance questionnaire. British journal of addiction 86:1119-1127
6.Hogarth LC, Mogg K, Bradley BP, Duka T, Dickinson A (2003) Attentional orienting towards smoking-related stimuli. Behavioural pharmacology 14:153-160
7.Jacobi F, Mack S, Gerschler A, Scholl L, Hofler M, Siegert J, Burkner A, Preiss S, Spitzer K, Busch M, Hapke U, Gaebel W, Maier W, Wagner M, Zielasek J, Wittchen HU (2013) The design and methods of the mental health module in the german health interview and examination survey for adults (degs1-mh). International journal of methods in psychiatric research 22:83-99
8.Kuitunen-Paul S, Bühringer G, Rehm J, Manthey J, Lachenmeier DW, & , Wittchen H (2017) Validation of alcohol consumption self-reports in the munich composite international diagnostic interview (m-cidi). In: Lisbon Addictions 2017. Lissabon, Portugal
9.Kuitunen-Paul S, Rehm J, Lachenmeier DW, Kadric F, Kuitunen PT, Wittchen HU, Manthey J (2017) Assessment of alcoholic standard drinks using the munich composite international diagnostic interview (m-cidi): An evaluation and subsequent revision. International journal of methods in psychiatric research 26
10.Lehrl S, Triebig G, Fischer B (1995) Multiple choice vocabulary test mwt as a valid and short test to estimate premorbid intelligence. Acta neurologica Scandinavica 91:335-345
11.Mogg K, Field M, Bradley BP (2005) Attentional and approach biases for smoking cues in smokers: An investigation of competing theoretical views of addiction. Psychopharmacology 180:333-341
12.Skinner HA, Horn JL (1984) Alcohol dependence scale (ads): Users guide. Addiction Research Foundation, Toronto
13.Wittchen H-U, Pfister H (1997) Dia-x-interviews: Manual für screening-verfahren und interview; interviewheft längsschnittuntersuchung (dia-x-lifetime); ergänzungsheft (dia-x-lifetime); interviewheft querschnittuntersuchung (dia-x-12 monate); ergänzungsheft (dia-x-12 monate); pc-programm zur durchführung des interviews (längs- und querschnittuntersuchung); auswertungsprogramm. Swets & Zeitlinger, Frankfurt
1