Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

DIAZEPAM DOSE-DEPENDENTLY INCREASES OR DECREASES IMPLICIT PRIMING OF ALCOHOL ASSOCIATIONS IN PROBLEM DRINKERS

DIAZEPAM DOSE-DEPENDENTLY INCREASES OR DECREASES IMPLICIT PRIMING OF ALCOHOL ASSOCIATIONS IN... Abstract Aims: Words denoting negative affect (NEG) have been found to prime alcohol-related words (ALC) on semantic priming tasks, and this effect is tied to severity of addiction. Previous research suggested that high doses of benzodiazepines may dampen NEG-ALC priming. The present study tested this possibility and the role of motivation for alcohol in this process. Methods: A placebo-controlled, double blind, between-within, counterbalanced design was employed. Two groups of male problem drinkers ( n = 6/group) received a high (15-mg) or low (5-mg) dose of diazepam versus placebo on two identical test sessions. A lexical decision task assessed priming. Results: Under placebo, significant NEG→ALC priming emerged in each group. High-dose diazepam selectively reversed this effect, while low-dose selectively enhanced it. Correlations between NEG→ALC priming and desire for alcohol provided further support that semantic priming of ALC concepts reflects a motivational process. The bi-directional effects found here parallel the effects of high- versus low-dose benzodiazepines on alcohol self-administration in animals. Conclusions: High-dose diazepam reduces prime-induced activation of ALC concepts in problem drinkers. Low-dose diazepam facilitates this process, and cross-priming of motivation for alcohol appears to explain this effect. Neurochemical modulation of the alcohol memory network may contribute to the motivational effects of benzodiazepines in problem drinkers. ( Received 9 December 2005; first review notified 6 March 2006; in revised form 17 August 2006; accepted 17 August 2006) INTRODUCTION In problem drinkers, negative affect (NEG) words (e.g. TENSE) have been found to prime or facilitate response time (RT) to alcohol target words (e.g. WHISKY) on a lexical decision task, and this effect coincided with greater alcohol dependence ( Zack et al ., 1999a ). Such semantic priming effects are ‘implicit’: involuntary and not mediated by conscious intention ( Nelson et al ., 1992 ). In contrast to this initial study, a subsequent naturalistic study found that problem drinkers medicated with benzodiazepines as part of their treatment exhibited a complete absence of negative affect→alcohol related (NEG→ALC) priming ( Zack et al ., 1999b ). Priming of ALC targets by other alcohol words, which typifies problem drinkers ( Hill and Paynter, 1992 ), was also absent. Importantly, these medicated drinkers displayed significant priming in a control condition of neutral, categorically-related words (e.g. HORSE-COW). Thus, benzodiazepines' generic amnestic effects ( Curran, 1986 ; Boucart et al ., 2002 ), cannot account for the selective dampening ALC semantic priming. One possible explanation is that the benzodiazepines were partially substituting for alcohol and inducing a degree of satiety (cf. Jackson et al ., 2003a , b ). If so, the lack of priming of alcohol associations in these drinkers reflected a reduction in their motivation for alcohol. In animals, benzodiazepines have dose-dependent bi-directional effects on alcohol self-administration: high doses decrease whereas low doses increase this response ( Petry, 1997 ). However, both low and high dose benzodiazepines decrease reinstatement of alcohol seeking in rats with a history of alcohol self-administration ( Jackson et al ., 2003b ). The differential effects of the low dose may reflect differences in the experimental models and the processes that underlie reinstatement of an extinguished response versus ongoing self-administration of alcohol (e.g. Liu and Weiss, 2004 ). Importantly, high-doses reduced motivation for alcohol in both models. Low-dose diazepam (5-mg) has been found to increase consumption of placebo (de-alcoholized) beer in problem drinkers ( Poulos and Zack, 2004 ). This result is consistent with Petry's (1997) finding, which may reflect the close correspondence between the experimental models in the two studies. Taken together, the results of the semantic and drink priming studies strongly imply that activation and de-activation of the alcohol memory network by benzodiazepines involves a motivational process. The present study tested this motivational hypothesis by administering a high (15-mg) or low dose (5-mg) of diazepam to community-recruited, non-abstinent problem drinkers. (Testing the effects of low doses on ‘reinstatement’ of activation of the alcohol memory network would require use of abstinent problem drinkers, which although not impossible, has important ethical ramifications.) It was predicted that, relative to a placebo capsule, the high dose would dampen whereas the low dose would augment NEG→ALC priming. Neither dose should alter motivationally irrelevant Neutral→Neutral (NEU→NEU) priming. Subjective effects of treatment were assessed by self-report. METHODS The study was approved by The Research Ethics Board of the Centre for Addiction and Mental Health, and conformed to principles of the Helsinki Declaration (1975). All participants provided written informed consent prior to participation and were fully debriefed upon completion. Participants Participants were 12 non-treatment-seeking male problem drinkers, ages 21–55 years (M = 37.3; SD = 13.9), recruited through newspaper advertisements. All were drug- and medication-free, as verified by urinalysis. A score ≥9 on the Alcohol Dependence Scale (ADS), which corresponds to DSM-III-R alcohol abuse ( Ross et al ., 1990 ), was required for inclusion. Detailed participant characteristics are provided in the Results section. No participant had a current psychiatric disorder, apart from problem drinking, or a history of benzodiazepine abuse, as determined by self-report and lack of record for addictions treatment. Participants were advised that they may receive a dose of a central nervous system medication, but were not aware that the medication was diazepam. Based on pre-screening, pairs of participants were formed who were similar in age, drinks per week, alcohol dependence and trait anxiety. One member from each pair was then randomly assigned to receive the 15-mg dose + placebo (Group High), and one member assigned to receive the 5-mg dose + placebo (Group Low). Participants received $300 compensation at the end of the study. Apparatus and materials A Pentium PC administered the lexical decision task and recorded the data with milliseconds accuracy. Participants sat facing the screen at a distance of 60 cm. Responses to target letter strings were made by pressing the ‘z’ key (non-word) or the ‘?’ key (word) with the index finger of the left or right hand, respectively. A handheld breathalyser (J-4-X ALERT, Alcohol Countermeasures Inc., Mississauga, Ontario, Canada) confirmed that all blood-alcohol levels were zero at the start of each session. Physiological measures were taken to ensure subject safety and detect any differences in cardiovascular response that may relate to task performance. An automated blood pressure cuff (HEM-601, OMRON Corp., Vernon Hills, IL), positioned over the participant's non-dominant hand, assessed heart rate and blood pressure at points during each test session. The ADS assessed severity of alcohol dependence. The Timeline Followback (TLFB; Sobell and Sobell, 1992 ) assessed drinks/week in the 90 days prior to testing. The trait portion of the State-Trait Anxiety Inventory (STAI; Spielberger et al ., 1970 ) assessed trait anxiety. During test sessions, the Positive and Negative Affect Scales (PANAS; Watson et al ., 1988 ) assessed participants' emotional state at pre-capsule baseline and again 70 min post-capsule, when cognitive-behavioural effects of oral diazepam would be expected to be maximal (de Wit et al ., 1993). A modified visual analogue scale (m-VAS) assessed desire for Alcohol (0–10: ‘not at all’ to ‘extreme’) at these same times under drug and placebo. Post-capsule ratings were taken immediately after the lexical decision task. Procedure Participants attended four sessions in total: an interview where the trait scales were completed, a physician's exam, and two test sessions, held 1 week apart to ensure a complete washout of diazepam for participants who received the drug on session 1. Participants were instructed to abstain from alcohol for 12 h prior to the start of each test session. Upon arrival, they received a breathalyzer and a standard breakfast. They then completed their baseline PANAS and m-VAS ratings and received their assigned dose. Each participant performed the lexical decision task alone in a quiet well-lit room. Trials began at 50 min post-capsule and ended at ∼70 min post-capsule. Stimuli (The stimulus list is available from M.Z. upon request.) and task parameters were the same as those used in the original studies ( Zack et al ., 1999a , b ; see below). After performing 15 practice trials with NEU prime–target pairs, participants proceeded to the test trials. They performed 20 trials in each of the five word target conditions and an equal number of trials for non-word targets. Prime Type and Target Type were randomized over the course of trials, and a 30-s rest break was provided after each set of 60 trials. The sequence of events on each trial was identical: fixation stimulus (++++; 250 ms)/blank screen (250 ms)/prime stimulus (750 ms)/target stimulus (until response)/blank screen (1000 ms). Stimuli were white upper-case letter strings (1 cm in height) displayed against a black background. Primes appeared in the centre of the screen, and targets appeared 1 cm below the centre of the screen. To prevent potential biasing of performance by prior exposure to ALC or NEG words, post-capsule self-report scales were administered after the lexical decision task. Participants received comprehensive instructions prior to the task. They were told to read the first string (i.e. the prime) of each pair silently, and respond as quickly and carefully as possible, by pressing the appropriate key, when the second string (i.e. the target) appeared. Upon task completion, participants filled out the questionnaires. Stimulus conditions for the lexical decision task The stimulus conditions for the lexical decision task are shown in Table 1 . As indicated, the two test conditions were NEG-ALC and ALC-ALC. These were designed to assess activation of Alcohol concepts by putatively related (i.e. NEG) or explicitly related (i.e. ALC) prime words. NEU-ALC was the baseline condition for both of these test conditions. In the baseline condition, primes and targets were explicitly unrelated. For NEU-ALC baseline trials, prime words denoted parts of a building. RT in this condition reflected un-primed accessibility of ALC concepts. Table 1. Stimulus conditions in the lexical decision task Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl NEG = negative affect word; ALC = alcohol word; NEU = neutral word; NEU (R) = neutral control word, explicitly related to its target (same category); NEU (U) = neutral control word, explicitly unrelated to its target (different category); NON = non-word target. Open in new tab Table 1. Stimulus conditions in the lexical decision task Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl NEG = negative affect word; ALC = alcohol word; NEU = neutral word; NEU (R) = neutral control word, explicitly related to its target (same category); NEU (U) = neutral control word, explicitly unrelated to its target (different category); NON = non-word target. Open in new tab The control condition was NEU-NEU [related, (R)]. This Condition assessed activation of NEU concepts by explicitly related primes, and determined effects of diazepam on semantic priming of motivationally irrelevant concepts. The corresponding baseline condition was NEU-NEU (U). Words for both conditions were drawn from two NEU categories: Four-legged mammals and articles of clothing. In the NEU-NEU (R) condition, primes and targets came from the same category; in the NEU-NEU (U) condition primes and targets came from opposite categories. The latter condition assessed un-primed accessibility of NEU concepts. For each class of targets, the difference in RT on trials containing putatively or explicitly related primes as opposed to explicitly unrelated primes measured priming. For example, Δ RT = (NEU-ALC) minus (NEG-ALC). Positive RT difference scores indicate priming, and imply an association between the prime and target concepts in memory. Data analytic plan Data reduction Individual RT scores >2500 or <250 ms (outliers) were excluded from analyses of treatment effects (cf. Neely, 1991 ). Similarly, errors (reporting a word as a non-word and vice versa) were excluded from analyses of treatment effects. The remaining RT scores (out of a possible 20 trials) for each participant were averaged to compute a mean RT for each prime–target condition. RT to non-words was not subjected to statistical analysis (cf. Neely, 1991 ). Analyses of mean effects NEG→ALC priming was assessed by the difference in RT for the NEU-ALC minus the NEG-ALC conditions. ALC→ALC priming was assessed by the difference in RT for the NEU-ALC minus ALC-ALC condition. Priming of motivationally irrelevant concepts was assessed by the difference in RT for NEU-NEU (U) minus NEU-NEU (R) conditions. A 2 (Group) × 2 (Treatment) × 3 (Condition) analysis of variance (ANOVA) assessed RT difference scores in the three test conditions. Simple effects analyses employed the MS error term from the ANOVA ( Winer, 1971 ). The total number of observations that contributed to the ANOVA was 2400 (200 trial scores/participant). Thus, despite the relatively small sample, the large number of observations per participant supported the reliability of the ANOVA. Desire for alcohol was assessed in a 2 (Group) × 2 (Treatment) ANOVA. Positive and negative affect subscales of the PANAS were assessed in separate 2 × 2 ANOVA's. Background characteristics for Groups High and Low were compared with independent samples t -tests. Correlational analyses A Pearson correlation assessed the relationship between NEG→ALC priming (RT difference) and motivation for alcohol (desire for alcohol) under placebo in the full sample. To assess the effect of diazepam, the same two indices were correlated using scores from the drug session, while variation under placebo was controlled by partial correlation. Separate analyses of the drug effect were performed for Group High and Group Low. RESULTS Participant characteristics Table 2 reports the mean (SD) background characteristics for Groups High and Low. There were no group differences on any variable, P -values > 0.55. This confirms the effectiveness of the matching procedure. Therefore, group differences in response to the Treatment are not attributable to background factors. Table 2. Mean (SD) participant characteristics for problem drinkers assigned to receive high- or low-dose diazepam Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) ADS = score on Alcohol Dependence Scale ( Skinner and Allen, 1982 ); Drinks per Week = average for 90 days preceding testing from TLFB ( Sobell and Sobell, 1992 ); Trait anxiety = score on trait portion of the State-Trait Anxiety Inventory (Speilberger et al ., 1970). Open in new tab Table 2. Mean (SD) participant characteristics for problem drinkers assigned to receive high- or low-dose diazepam Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) ADS = score on Alcohol Dependence Scale ( Skinner and Allen, 1982 ); Drinks per Week = average for 90 days preceding testing from TLFB ( Sobell and Sobell, 1992 ); Trait anxiety = score on trait portion of the State-Trait Anxiety Inventory (Speilberger et al ., 1970). Open in new tab ADS scores indicated ‘moderate’ dependence ( Skinner and Allen, 1982 ). Drinks/week were moderate-heavy for outpatient problem drinkers ( Wilkinson and LeBreton, 1986 ). Trait anxiety scores corresponded to the 50th percentile, based on general population norms ( Vagg et al ., 1980 ). Lexical decision task performance Procedural checks Errors. An ANOVA of errors for each Group, prime–target condition and Treatment yielded no significant effects, P -values > 0.17. Thus, error rates did not bias effects of the experimental factors. The frequency of errors ranged from 3.1 to 7.5% (0.6–1.5 of 20 trials) across conditions. Outliers. An ANOVA of outliers yielded a significant Group × Treatment interaction, F (1, 8) = 7.36, P < 0.05, a main effect of Treatment, F (1, 8) = 7.36, P < 0.05, and no other significant effects, P -values > 0.10. In Group High, more outliers occurred under drug (M = 1.7%) than placebo (M = 0.1%), whereas in Group Low, the number of outliers did not differ under drug versus placebo (M = 0.4%). The main effect of Treatment reflected more outliers overall under drug (M = 1.0%) versus placebo (M = 0.3%). The lack of effects involving prime condition, along with the very small absolute frequency of outliers (<0.4 of 20 trials) indicates that outliers did not bias effects of the experimental factors. Baseline performance. To ensure that differences in baseline RT did not influence priming effects, a preliminary 2 (Group) × 2 (Treatment) × 2 (Target Type) ANOVA assessed RT to ALC and NEU targets paired with explicitly unrelated, NEU primes. The analysis yielded a marginal Group × Treatment interaction, P = 0.058, a marginal main effect of Treatment, P = 0.094, and no other marginal or significant effects, P -values > 0.23. In Group High, mean (SD) RT (ms), averaged across ALC and NEU targets, tended to be slower under drug, 889 (148) than placebo, 776 (100), whereas in Group Low, mean RT was slightly faster under drug, 755 (128) than placebo, 763 (115). More importantly, the lack of significant effects involving Target Type ensures that any significant effects for RT difference scores (i.e. priming) as a function of Target Type cannot be attributed to variation in baseline RT under either Treatment. Parametric requirements. Mauchly's test was non-significant for all factors, P -values > 0.16. Therefore, the variance–covariance matrix did not deviate from the expected pattern under the null hypothesis (i.e. sphericity). The mean inter-correlation for RT difference scores across levels of Prime Type under placebo was r = 0.52, P < 0.05, one-tailed. Skewness statistics (SPSS v. 11.5; Explore Function, SPSS Inc., Chicago, IL) for each prime–target condition found no significant effects, P -values > 0.07. Thus, the RT distributions were symmetrical. These results confirm that the data meet the requirements of repeated-measures ANOVA. Primary analysis of RT difference scores A 2 (Group) × 2 (Treatment) × 3 (Prime Type) ANOVA of RT difference scores for the NEG-ALC, ALC-ALC and NEU-NEU (R) conditions yielded a significant three-way interaction, F (2, 20) = 3.51, P < 0.05, and no other significant effects, P -values > 0.39. Figure 1A and B shows the mean (SEM) RT scores to ALC and NEU targets under placebo in Groups High and Low, respectively. For each class of targets, significant priming effects are indicated by RT difference (Δ RT * ) from baseline. Comparing panels A and B reveals that RT values for each baseline condition under placebo were very similar in the two groups. Inspection of the RT difference from baseline for ALC targets reveals that NEG primes induced significant and comparable priming in each group, P -values < 0.05. ALC primes induced significant priming of ALC targets in Group Low, P < 0.05, and a marginally significant trend in priming ( P < 0.05, one-tailed) in Group High. Control NEU targets were also primed significantly by NEU (R) primes relative to NEU (U) baseline in each group, P -values < 0.05. The similar pattern of NEG-ALC priming effects in each group under placebo ensures that any group differences in this condition under diazepam cannot be attributed to drug-free patterns of association. Fig. 1. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under placebo. (A) Left panel shows scores for participants who received 15-mg diazepam on the drug session ( n = 6). (B) Right panel shows scores for participants who received 5-mg diazepam on the drug session ( n = 6). Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related prime conditions. Larger positive Δ RT scores indicate greater priming. * Δ RT: Simple effect of prime relative to baseline, P < 0.05 Fig. 1. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under placebo. (A) Left panel shows scores for participants who received 15-mg diazepam on the drug session ( n = 6). (B) Right panel shows scores for participants who received 5-mg diazepam on the drug session ( n = 6). Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related prime conditions. Larger positive Δ RT scores indicate greater priming. * Δ RT: Simple effect of prime relative to baseline, P < 0.05 Figure 2A and B shows the mean (SEM) RT difference scores for ALC and NEU targets in Groups High and Low under diazepam. Comparing panels A and B reveals marked differences in the pattern of effects for the two groups in the case of ALC targets. In addition, priming of NEU targets by NEU (R) primes appears to differ in Groups High and Low, although the pattern of effects is the opposite to that which emerged for ALC targets. Simple effects analyses determined that, in Group High, NEG primes and ALC primes each led to a significant slowing of RT to ALC targets relative to baseline (i.e. an interference effect), P -values < 0.05. In contrast, both NEG and ALC primes significantly primed RT to ALC targets in Group Low, P -values < 0.05. Fig. 2. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under diazepam. (A) Left panel shows scores for participants who received 15-mg dose ( n = 6). (B) Right panel shows scores for participants who received 5-mg dose. Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related Prime conditions. Larger positive Δ RT scores indicate greater priming. Negative Δ RT scores indicate interference (slower RT to a class of Targets in response to putatively/explicitly related Primes as opposed to explicitly unrelated Primes). * Δ RT: simple effect of Prime relative to baseline, P < 0.05 Fig. 2. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under diazepam. (A) Left panel shows scores for participants who received 15-mg dose ( n = 6). (B) Right panel shows scores for participants who received 5-mg dose. Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related Prime conditions. Larger positive Δ RT scores indicate greater priming. Negative Δ RT scores indicate interference (slower RT to a class of Targets in response to putatively/explicitly related Primes as opposed to explicitly unrelated Primes). * Δ RT: simple effect of Prime relative to baseline, P < 0.05 Of particular relevance to the hypothesis, was the comparison of RT difference scores in the NEG-ALC and ALC-ALC conditions in Group Low under diazepam versus placebo (see Figs 1B versus 2B ). Simple effects analyses revealed that, in these participants, diazepam significantly increased the magnitude of the difference score relative to placebo, reflecting a drug-induced augmentation of priming in the NEG-ALC condition, P < 0.05. Although the mean RT difference score on ALC-ALC trials also increased under drug versus placebo in Group Low, the degree of increase (30%) was not statistically significant, P > 0.25. Comparing Fig. 2A versus B reveals that Group High displayed significant NEU→NEU (R) priming under drug, P < 0.05, whereas Group Low did not. Comparing the degree of priming under drug versus placebo in each group reveals that NEU-NEU priming increased significantly under drug in Group High, P < 0.01, but decreased significantly under drug in Group Low, P < 0.01. Self-report ratings Desire for alcohol There were no differences in pre-capsule ‘desire’ ratings as a function of Group or Treatment, P -values > 0.43. A 2 × 2 ANOVA of desire ratings yielded no significant effects, P -values > 0.28. The mean (SD) scores for placebo versus drug, respectively, were 2.7 (3.3) versus 2.0 (1.1) in Group High and 0.8 (1.3) versus 1.3 (2.4) in GroupLow. This pattern was consistent with the expected decrease under the high dose and increase under the low dose, although simple effects did not achieve significance, P -values > 0.07. PANAS positive and negative affect scores There were no significant differences in pre-capsule baseline scores on either the positive or negative subscale, P -values > 0.14. An ANOVA of PANAS positive subscale scores yielded a Group × Treatment interaction, F (1, 10) = 6.50, P < 0.05, and no other significant effects, P -values > 0.48. Simple effects analyses revealed that, in Group High, mean (SD) positive affect ratings were lower under diazepam, 24.3 (9.7) than placebo 28.8 (10.7), P < 0.01, whereas in Group Low, positive affect was higher under diazepam, 24.0 (10.0) than placebo 21.2 (8.7), P < 0.05. The ANOVA of PANAS negative subscale scores yielded no significant effects, P -values > 0.30. The overall mean (SD) negative subscale rating was 12.8 (5.9). Physiological effects An ANOVA of heart rate at 1-h post-capsule yielded no significant effects, P -values > 0.37. Mean (SD) overall heart rate was 77.8 (6.8 beats/min). An ANOVA of systolic/diastolic blood pressure yielded no significant effects involving Treatment or Group, P -values > 0.06. Mean (SD) overall scores were 129/86 (13/10) mm Hg. Thus, neither dose of diazepam significantly altered cardiovascular activity in these problem drinkers. Correlational analyses The correlation between NEG→ALC priming (RT difference score) and desire for alcohol for all participants under placebo was significant, r = 0.61, P < 0.05. Thus, in the absence of a drug, greater priming of ALC targets by NEG words was associated with stronger motivation for alcohol. The partial correlation between NEG→ALC priming and desire for alcohol under diazepam, controlling for variation under placebo was non-significant in Group High, r = 0.34, P > 0.25, but significant in Group Low, r = 0.79, P = 0.03, one-tailed. Thus, under the low dose, diazepam-induced changes in NEG→ALC priming directly predicted drug-induced changes in desire for alcohol. DISCUSSION The results of this study confirm that a single high dose of diazepam can abolish NEG→ALC priming and ALC→ALC priming in problem drinkers. Indeed, this dose led to significant interference of these associations (slowing of RT) relative to the NEU-ALC baseline, and relative to placebo. The interference effect was selective for ALC associations; the high dose led to significantly increased NEU→NEU (Related) priming relative to placebo. These findings provide a controlled experimental demonstration that high-dose diazepam selectively dampens implicit priming of ALC associations in problem drinkers. The finding that high-dose diazepam slowed RT in the NEG-ALC and ALC-ALC conditions relative to NEU-ALC baseline indicates that, under diazepam, motivationally relevant cues impeded access to ALC targets. Such interference effects imply that, under drug, the NEG and ALC primes were activating an alternative set of associations rather than simply failing to activate ALC-related ones ( Dagenbach et al ., 1990 ). High doses of benzodiazepines have been found to induce unusual (i.e. categorically unrelated, statistically rare) associations to common words in problem drinkers ( Weingartner et al ., 1998 ). This suggests that the interference observed under diazepam in Group High reflected activation by NEG and ALC primes of unusual associates that were incompatible with alcohol. The near doubling of variability in the ALC-ALC condition and 50% increase in variability in the NEG-ALC condition relative to NEU-ALC baseline (see error bars in Fig. 2 ) is consistent with this explanation. NEU words significantly primed RT to NEU, categorically-related targets under the high dose. Thus, NEU cues can still facilitate access to their associates under high dose diazepam. This implies that the motivational relevance of a stimulus influences the degree of idiosyncratic associations exhibited by problem drinkers under high-dose benzodiazepines. The increase in NEU→NEU priming coupled with the interference to ALC targets under drug in Group High further implies that high-dose diazepam renders NEU concepts relatively more accessible than ALC targets in problem drinkers. Given the potential therapeutic value of such a shift, the generality of this effect deserves further investigation. The low-dose manipulation permitted a test of the motivational hypothesis. The results of Group Low clearly support this hypothesis. The 5-mg dose led to increased NEG→ALC priming and a parallel, though non-significant, increase in ALC→ALC priming. Conversely, this dose decreased NEU→NEU priming under diazepam relative to placebo. Thus, low-dose diazepam selectively enhanced the ability of motivationally relevant cues to activate NEG→ALC associations. The intermediate effects on ALC→ALC priming may reflect the involvement of both categorical and motivational associations in this condition. The reason for the decline in NEU→NEU priming under the low dose is unclear. It is possible that motivationally irrelevant words lose priming potential during states of heightened motivation. This is in line with previous research, which found that RT to NEU word stimuli was inhibited following administration of a drug prime ( Zack and Poulos, 2004 ). Whatever the mechanism, the decline in NEU→NEU priming under the low dose makes the increase in priming of ALC concepts all the more remarkable. Importantly, the findings for Group Low preclude an associative strength explanation for the dampening effects seen in Group High. According to this explanation, high-dose diazepam negated ALC because they are weaker than NEU, categorical associations. However, by this account, low-dose diazepam should also have reduced alcohol associations, albeit to a lesser extent. In fact, the opposite occurred. The correlational results further support the motivational hypothesis. Under placebo, NEG→ALC priming significantly predicted the desire for alcohol in the full sample. Thus, under drug-free conditions, NEG→ALC priming is directly linked to motivation for alcohol. Under low-dose diazepam, NEG-ALC priming continued to predict the desire for alcohol when co-variation between these indices under placebo was partialled out. Thus, low-dose diazepam exerts parallel effects on subjective motivation for alcohol and implicit priming of ALC associations in problem drinkers. Under the high dose, the correlation between NEG→ALC priming and ‘desire’ was non-significant. This may reflect the idiosyncratic associations alluded to earlier, which would have obscured systematic variation in NEG→ALC priming and weakened the correlation with ‘desire’. Future research that assesses problem drinkers' spontaneous associations to NEG and ALC words under high-dose diazepam could test this possibility. The correlational and mean effects strongly support the assertion that diazepam was modulating ALC-related motivational systems. The mean effects also mirror the bi-directional dose–response effect of benzodiazepines on alcohol self-administration in animals ( Petry, 1997 ). Taken together, the findings indicate that drug stimuli with similar neurochemical properties exert parallel effects on semantic and motivational priming. From a clinical standpoint, the results for Group High imply that a high dose of a benzodiazepine may reduce the (primed) bias to think about or seek alcohol during NEG states. Because priming effects are involuntary, a medication that can inhibit them could be valuable to someone who is motivated to avoid drinking in response to NEG (e.g. an anxious problem drinker). At the same time, the inability to achieve such cognitive dampening effects without benzodiazepines, could lead to dependence on these medications to remain abstinent from alcohol. In light of this, it is noteworthy that a GABA-ergic medication with low abuse potential (the GABA-B agonist, baclofen) has shown promise for treating alcohol dependence ( Addolorato et al ., 2002 ; Ameisen, 2005 ). Interestingly, baclofen has also been found to exert amnestic effects in animals and humans ( Sandyk and Gillman, 1985 ; Pitsikas et al ., 2003 ). Given this profile of effects, the possibility that baclofen can modulate the alcohol memory network in problem drinkers is an important issue for future investigation. REFERENCES Addolorato, G., Caputo, F., Capristo, E. et al. ( 2002 ) Baclofen efficacy in reducing alcohol craving and intake: a preliminary randomized controlled study. Alcohol and Alcoholism 37 , 504 –508. Ameisen, O. ( 2005 ) Complete and prolonged suppression of symptoms and consequences of alcohol-dependence using high-dose baclofen: a self-case report of a physician. Alcohol and Alcoholism 40 , 147 –150. Boucart, M., Biederman, I., Cuervo, C. et al . ( 2002 ) Effect of benzodiazepines on structural and conceptual/lexical priming. Psychopharmacology 165 , 43 –50. Curran, H. V. ( 1986 ) Tranquillising memories: a review of the effects of benzodiazepines on human memory. Biological Psychology 23 , 179 –213. Dagenbach, D., Carr, T. H. and Barnhardt, T. M. ( 1990 ) Inhibitory semantic priming of lexical decisions due to failure to retrieve weakly activated codes. Journal of Experimental Psychology: Learning, Memory, and Cognition 16 , 328 –340. de Wit, H., Dudish, S. and Ambre, J. ( 1993 ) Subjective and behavioural effects of diazepam depend on its rate of onset. Psychopharmacology 112 , 324 –330. Hill, A. B. and Paynter, S. ( 1992 ) Alcohol dependence and semantic priming of alcohol related words. Personality and Individual Differences 13 , 745 –750. Jackson, A., Stephens, D. N., and Duka, T. ( 2003 ) Lorazepam substitutes for the alcohol stimulus in social drinkers. Psychopharmacology 166 , 181 –187. Jackson, A., Duka, T., and Stephens D. N. ( 2003 ) Effects of alcohol and lorazepam during extinction of alcohol self-administration in rats. Journal of Psychopharmacology 17 , 293 –299. Liu, X. and Weiss, F. ( 2004 ) Nitric oxide synthesis inhibition attenuates conditioned reinstatement of ethanol-seeking, but not the primary reinforcing effects of ethanol. Alcoholism: Clinical and Experimental Research 28 , 1194 –1199. Neely, J. H. ( 1991 ) Semantic priming effects in visual word recognition: A selective review of current findings and theories. In, Basic processes in reading: Visual Word Recognition , Besner, D. and Humphreys, G. W. eds, pp. 264–337. Erlbaum, Hillsdale, NJ. Nelson, D. L., Schreiber, T. A. and McEvoy, C. L. ( 1992 ) Processing implicit and explicit representations. Psychological Review 99 , 322 –348. Petry, N. M. ( 1997 ) Benzodiazepine-GABA modulation of concurrent ethanol and sucrose reinforcement in the rat. Experimental and Clinical Psychopharmacology 5 , 183 –194. Pitsikas, N., Rigamonti, A. E., Cella, S. G. et al. ( 2003 ) The GABA-B receptor and recognition memory: possible modulation of its behavioral effects by the nitrergic system. Neuroscience 118 , 1121 –1127. Poulos, C. X. and Zack, M. ( 2004 ) Low-dose diazepam primes motivation for alcohol and alcohol-related semantic networks in problem drinkers. Behavioural Pharmacology 15 , 503 –512. Ross, H. E., Gavin, D. R. and Skinner, H. A. ( 1990 ) Diagnostic validity of the MAST and the alcohol dependence scale in the assessment of DSM-III alcohol disorders. Journal of Studies on Alcohol 51 , 506 –513. Sandyk, R. and Gillman, M. A. ( 1985 ) Baclofen-induced memory impairment. Clinical Neuropharmacology 8 , 294 –295. Skinner, H. A. and Allen, B. A. ( 1982 ) Alcohol dependence syndrome: measurement and validation. Journal of Abnormal Psychology 91 , 199 –209. Sobell, L. C. and Sobell, M. B. ( 1992 ) Timeline Followback: a technique for assessing self-reported alcohol consumption. In Measuring Alcohol Consumption, Psychosocial And Biological Methods , Litten, R. Z. and Allen, J. eds, pp 41–72. Humana, Totowa, NJ. Spielberger, C. D., Gorsuch, R. L. and Lushene, R. E. ( 1970 ) Manual for the State-Trait Anxiety Inventory . Palo Alto, CA: Consulting Psychologists Press. Vagg, P. R., Spielberger, C. D. and O'Hearn, T. P. J. ( 1980 ) Is the State-Trait Anxiety Inventory multidimensional? Personality and Individual Differences 1 , 207 –214. Watson, D., Clark, L. A. and Tellegen, A. ( 1988 ) Development and validation of brief measures of positive and negative affect, the PANAS scales. Journal of Personality and Social Psychology 54 , 1063 –1070. Weingartner, H. J., Rawlings, R., George, D. T. et al. ( 1998 ) Triazolam-induced changes in alcoholic thought processes. Psychopharmacology 138 , 311 –317. Wilkinson, D. A. and LeBreton, S. ( 1986 ) Early indicators of treatment outcome in multiple drug users. In Treating Addictive Behaviours , Miller, W. R. and Heather, N. eds, pp. 239–261. New York: Plenum. Winer, B. J. ( 1971 ) Statistical Principles in Experimental Design , 2nd edn. McGraw-Hill, New York. Zack, M. and Poulos, C. X. ( 2004 ) Amphetamine primes motivation to gamble and gambling-related semantic networks in problem gamblers. Neuropsychopharmacology 29 , 195 –207. Zack, M., Toneatto, T. and MacLeod, C. M. ( 1999 ) Implicit activation of alcohol concepts by negative affective cues distinguishes between problem drinkers with high and low psychiatric distress. Journal of Abnormal Psychology 108 , 518 –531. Zack, M., Toneatto, T. and MacLeod, C. M. ( 1999 ) Clinical use of benzodiazepines and decreased memory activation in anxious problem drinkers. Alcoholism: Clinical and Experimental Research 23 , 174 –182. © The Author 2006. Published by Oxford University Press on behalf of the Medical Council on Alcohol. All rights reserved http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Alcohol and Alcoholism Oxford University Press

DIAZEPAM DOSE-DEPENDENTLY INCREASES OR DECREASES IMPLICIT PRIMING OF ALCOHOL ASSOCIATIONS IN PROBLEM DRINKERS

Download PDF
7 pages

Loading next page...
 
/lp/oxford-university-press/diazepam-dose-dependently-increases-or-decreases-implicit-priming-of-s8ObBCtQcC

References (28)

Publisher
Oxford University Press
Copyright
© The Author 2006. Published by Oxford University Press on behalf of the Medical Council on Alcohol. All rights reserved
ISSN
0735-0414
eISSN
1464-3502
DOI
10.1093/alcalc/agl076
pmid
17020910
Publisher site
See Article on Publisher Site

Abstract

Abstract Aims: Words denoting negative affect (NEG) have been found to prime alcohol-related words (ALC) on semantic priming tasks, and this effect is tied to severity of addiction. Previous research suggested that high doses of benzodiazepines may dampen NEG-ALC priming. The present study tested this possibility and the role of motivation for alcohol in this process. Methods: A placebo-controlled, double blind, between-within, counterbalanced design was employed. Two groups of male problem drinkers ( n = 6/group) received a high (15-mg) or low (5-mg) dose of diazepam versus placebo on two identical test sessions. A lexical decision task assessed priming. Results: Under placebo, significant NEG→ALC priming emerged in each group. High-dose diazepam selectively reversed this effect, while low-dose selectively enhanced it. Correlations between NEG→ALC priming and desire for alcohol provided further support that semantic priming of ALC concepts reflects a motivational process. The bi-directional effects found here parallel the effects of high- versus low-dose benzodiazepines on alcohol self-administration in animals. Conclusions: High-dose diazepam reduces prime-induced activation of ALC concepts in problem drinkers. Low-dose diazepam facilitates this process, and cross-priming of motivation for alcohol appears to explain this effect. Neurochemical modulation of the alcohol memory network may contribute to the motivational effects of benzodiazepines in problem drinkers. ( Received 9 December 2005; first review notified 6 March 2006; in revised form 17 August 2006; accepted 17 August 2006) INTRODUCTION In problem drinkers, negative affect (NEG) words (e.g. TENSE) have been found to prime or facilitate response time (RT) to alcohol target words (e.g. WHISKY) on a lexical decision task, and this effect coincided with greater alcohol dependence ( Zack et al ., 1999a ). Such semantic priming effects are ‘implicit’: involuntary and not mediated by conscious intention ( Nelson et al ., 1992 ). In contrast to this initial study, a subsequent naturalistic study found that problem drinkers medicated with benzodiazepines as part of their treatment exhibited a complete absence of negative affect→alcohol related (NEG→ALC) priming ( Zack et al ., 1999b ). Priming of ALC targets by other alcohol words, which typifies problem drinkers ( Hill and Paynter, 1992 ), was also absent. Importantly, these medicated drinkers displayed significant priming in a control condition of neutral, categorically-related words (e.g. HORSE-COW). Thus, benzodiazepines' generic amnestic effects ( Curran, 1986 ; Boucart et al ., 2002 ), cannot account for the selective dampening ALC semantic priming. One possible explanation is that the benzodiazepines were partially substituting for alcohol and inducing a degree of satiety (cf. Jackson et al ., 2003a , b ). If so, the lack of priming of alcohol associations in these drinkers reflected a reduction in their motivation for alcohol. In animals, benzodiazepines have dose-dependent bi-directional effects on alcohol self-administration: high doses decrease whereas low doses increase this response ( Petry, 1997 ). However, both low and high dose benzodiazepines decrease reinstatement of alcohol seeking in rats with a history of alcohol self-administration ( Jackson et al ., 2003b ). The differential effects of the low dose may reflect differences in the experimental models and the processes that underlie reinstatement of an extinguished response versus ongoing self-administration of alcohol (e.g. Liu and Weiss, 2004 ). Importantly, high-doses reduced motivation for alcohol in both models. Low-dose diazepam (5-mg) has been found to increase consumption of placebo (de-alcoholized) beer in problem drinkers ( Poulos and Zack, 2004 ). This result is consistent with Petry's (1997) finding, which may reflect the close correspondence between the experimental models in the two studies. Taken together, the results of the semantic and drink priming studies strongly imply that activation and de-activation of the alcohol memory network by benzodiazepines involves a motivational process. The present study tested this motivational hypothesis by administering a high (15-mg) or low dose (5-mg) of diazepam to community-recruited, non-abstinent problem drinkers. (Testing the effects of low doses on ‘reinstatement’ of activation of the alcohol memory network would require use of abstinent problem drinkers, which although not impossible, has important ethical ramifications.) It was predicted that, relative to a placebo capsule, the high dose would dampen whereas the low dose would augment NEG→ALC priming. Neither dose should alter motivationally irrelevant Neutral→Neutral (NEU→NEU) priming. Subjective effects of treatment were assessed by self-report. METHODS The study was approved by The Research Ethics Board of the Centre for Addiction and Mental Health, and conformed to principles of the Helsinki Declaration (1975). All participants provided written informed consent prior to participation and were fully debriefed upon completion. Participants Participants were 12 non-treatment-seeking male problem drinkers, ages 21–55 years (M = 37.3; SD = 13.9), recruited through newspaper advertisements. All were drug- and medication-free, as verified by urinalysis. A score ≥9 on the Alcohol Dependence Scale (ADS), which corresponds to DSM-III-R alcohol abuse ( Ross et al ., 1990 ), was required for inclusion. Detailed participant characteristics are provided in the Results section. No participant had a current psychiatric disorder, apart from problem drinking, or a history of benzodiazepine abuse, as determined by self-report and lack of record for addictions treatment. Participants were advised that they may receive a dose of a central nervous system medication, but were not aware that the medication was diazepam. Based on pre-screening, pairs of participants were formed who were similar in age, drinks per week, alcohol dependence and trait anxiety. One member from each pair was then randomly assigned to receive the 15-mg dose + placebo (Group High), and one member assigned to receive the 5-mg dose + placebo (Group Low). Participants received $300 compensation at the end of the study. Apparatus and materials A Pentium PC administered the lexical decision task and recorded the data with milliseconds accuracy. Participants sat facing the screen at a distance of 60 cm. Responses to target letter strings were made by pressing the ‘z’ key (non-word) or the ‘?’ key (word) with the index finger of the left or right hand, respectively. A handheld breathalyser (J-4-X ALERT, Alcohol Countermeasures Inc., Mississauga, Ontario, Canada) confirmed that all blood-alcohol levels were zero at the start of each session. Physiological measures were taken to ensure subject safety and detect any differences in cardiovascular response that may relate to task performance. An automated blood pressure cuff (HEM-601, OMRON Corp., Vernon Hills, IL), positioned over the participant's non-dominant hand, assessed heart rate and blood pressure at points during each test session. The ADS assessed severity of alcohol dependence. The Timeline Followback (TLFB; Sobell and Sobell, 1992 ) assessed drinks/week in the 90 days prior to testing. The trait portion of the State-Trait Anxiety Inventory (STAI; Spielberger et al ., 1970 ) assessed trait anxiety. During test sessions, the Positive and Negative Affect Scales (PANAS; Watson et al ., 1988 ) assessed participants' emotional state at pre-capsule baseline and again 70 min post-capsule, when cognitive-behavioural effects of oral diazepam would be expected to be maximal (de Wit et al ., 1993). A modified visual analogue scale (m-VAS) assessed desire for Alcohol (0–10: ‘not at all’ to ‘extreme’) at these same times under drug and placebo. Post-capsule ratings were taken immediately after the lexical decision task. Procedure Participants attended four sessions in total: an interview where the trait scales were completed, a physician's exam, and two test sessions, held 1 week apart to ensure a complete washout of diazepam for participants who received the drug on session 1. Participants were instructed to abstain from alcohol for 12 h prior to the start of each test session. Upon arrival, they received a breathalyzer and a standard breakfast. They then completed their baseline PANAS and m-VAS ratings and received their assigned dose. Each participant performed the lexical decision task alone in a quiet well-lit room. Trials began at 50 min post-capsule and ended at ∼70 min post-capsule. Stimuli (The stimulus list is available from M.Z. upon request.) and task parameters were the same as those used in the original studies ( Zack et al ., 1999a , b ; see below). After performing 15 practice trials with NEU prime–target pairs, participants proceeded to the test trials. They performed 20 trials in each of the five word target conditions and an equal number of trials for non-word targets. Prime Type and Target Type were randomized over the course of trials, and a 30-s rest break was provided after each set of 60 trials. The sequence of events on each trial was identical: fixation stimulus (++++; 250 ms)/blank screen (250 ms)/prime stimulus (750 ms)/target stimulus (until response)/blank screen (1000 ms). Stimuli were white upper-case letter strings (1 cm in height) displayed against a black background. Primes appeared in the centre of the screen, and targets appeared 1 cm below the centre of the screen. To prevent potential biasing of performance by prior exposure to ALC or NEG words, post-capsule self-report scales were administered after the lexical decision task. Participants received comprehensive instructions prior to the task. They were told to read the first string (i.e. the prime) of each pair silently, and respond as quickly and carefully as possible, by pressing the appropriate key, when the second string (i.e. the target) appeared. Upon task completion, participants filled out the questionnaires. Stimulus conditions for the lexical decision task The stimulus conditions for the lexical decision task are shown in Table 1 . As indicated, the two test conditions were NEG-ALC and ALC-ALC. These were designed to assess activation of Alcohol concepts by putatively related (i.e. NEG) or explicitly related (i.e. ALC) prime words. NEU-ALC was the baseline condition for both of these test conditions. In the baseline condition, primes and targets were explicitly unrelated. For NEU-ALC baseline trials, prime words denoted parts of a building. RT in this condition reflected un-primed accessibility of ALC concepts. Table 1. Stimulus conditions in the lexical decision task Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl NEG = negative affect word; ALC = alcohol word; NEU = neutral word; NEU (R) = neutral control word, explicitly related to its target (same category); NEU (U) = neutral control word, explicitly unrelated to its target (different category); NON = non-word target. Open in new tab Table 1. Stimulus conditions in the lexical decision task Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl Condition . Prime–target relation . Prime . Target . Example . Test Putatively related NEG ALC worry-beer Test Explicitly related ALC ALC whiskey-rum Baseline Explicitly unrelated NEU ALC window-gin Control Explicitly related NEU NEU (R) horse-cow Baseline Explicitly unrelated NEU NEU (U) dress-goat Filler Explicitly unrelated NEG NON tense-slore ALC NON wine-fibbage NEU NON door-perkafe NEU NON sheep-glimp NEU NON coat-lewl NEG = negative affect word; ALC = alcohol word; NEU = neutral word; NEU (R) = neutral control word, explicitly related to its target (same category); NEU (U) = neutral control word, explicitly unrelated to its target (different category); NON = non-word target. Open in new tab The control condition was NEU-NEU [related, (R)]. This Condition assessed activation of NEU concepts by explicitly related primes, and determined effects of diazepam on semantic priming of motivationally irrelevant concepts. The corresponding baseline condition was NEU-NEU (U). Words for both conditions were drawn from two NEU categories: Four-legged mammals and articles of clothing. In the NEU-NEU (R) condition, primes and targets came from the same category; in the NEU-NEU (U) condition primes and targets came from opposite categories. The latter condition assessed un-primed accessibility of NEU concepts. For each class of targets, the difference in RT on trials containing putatively or explicitly related primes as opposed to explicitly unrelated primes measured priming. For example, Δ RT = (NEU-ALC) minus (NEG-ALC). Positive RT difference scores indicate priming, and imply an association between the prime and target concepts in memory. Data analytic plan Data reduction Individual RT scores >2500 or <250 ms (outliers) were excluded from analyses of treatment effects (cf. Neely, 1991 ). Similarly, errors (reporting a word as a non-word and vice versa) were excluded from analyses of treatment effects. The remaining RT scores (out of a possible 20 trials) for each participant were averaged to compute a mean RT for each prime–target condition. RT to non-words was not subjected to statistical analysis (cf. Neely, 1991 ). Analyses of mean effects NEG→ALC priming was assessed by the difference in RT for the NEU-ALC minus the NEG-ALC conditions. ALC→ALC priming was assessed by the difference in RT for the NEU-ALC minus ALC-ALC condition. Priming of motivationally irrelevant concepts was assessed by the difference in RT for NEU-NEU (U) minus NEU-NEU (R) conditions. A 2 (Group) × 2 (Treatment) × 3 (Condition) analysis of variance (ANOVA) assessed RT difference scores in the three test conditions. Simple effects analyses employed the MS error term from the ANOVA ( Winer, 1971 ). The total number of observations that contributed to the ANOVA was 2400 (200 trial scores/participant). Thus, despite the relatively small sample, the large number of observations per participant supported the reliability of the ANOVA. Desire for alcohol was assessed in a 2 (Group) × 2 (Treatment) ANOVA. Positive and negative affect subscales of the PANAS were assessed in separate 2 × 2 ANOVA's. Background characteristics for Groups High and Low were compared with independent samples t -tests. Correlational analyses A Pearson correlation assessed the relationship between NEG→ALC priming (RT difference) and motivation for alcohol (desire for alcohol) under placebo in the full sample. To assess the effect of diazepam, the same two indices were correlated using scores from the drug session, while variation under placebo was controlled by partial correlation. Separate analyses of the drug effect were performed for Group High and Group Low. RESULTS Participant characteristics Table 2 reports the mean (SD) background characteristics for Groups High and Low. There were no group differences on any variable, P -values > 0.55. This confirms the effectiveness of the matching procedure. Therefore, group differences in response to the Treatment are not attributable to background factors. Table 2. Mean (SD) participant characteristics for problem drinkers assigned to receive high- or low-dose diazepam Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) ADS = score on Alcohol Dependence Scale ( Skinner and Allen, 1982 ); Drinks per Week = average for 90 days preceding testing from TLFB ( Sobell and Sobell, 1992 ); Trait anxiety = score on trait portion of the State-Trait Anxiety Inventory (Speilberger et al ., 1970). Open in new tab Table 2. Mean (SD) participant characteristics for problem drinkers assigned to receive high- or low-dose diazepam Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) Group ( n = 6/group) . Age (years) . ADS . Drinks per week . Trait anxiety .     High (15 mg) 39.8 (14.9) 14.2 (9.2) 37.1 (15.7) 35.2 (12.2)     Low (5 mg) 34.8 (13.6) 12.5 (5.2) 36.2 (17.6) 38.3 (9.7) ADS = score on Alcohol Dependence Scale ( Skinner and Allen, 1982 ); Drinks per Week = average for 90 days preceding testing from TLFB ( Sobell and Sobell, 1992 ); Trait anxiety = score on trait portion of the State-Trait Anxiety Inventory (Speilberger et al ., 1970). Open in new tab ADS scores indicated ‘moderate’ dependence ( Skinner and Allen, 1982 ). Drinks/week were moderate-heavy for outpatient problem drinkers ( Wilkinson and LeBreton, 1986 ). Trait anxiety scores corresponded to the 50th percentile, based on general population norms ( Vagg et al ., 1980 ). Lexical decision task performance Procedural checks Errors. An ANOVA of errors for each Group, prime–target condition and Treatment yielded no significant effects, P -values > 0.17. Thus, error rates did not bias effects of the experimental factors. The frequency of errors ranged from 3.1 to 7.5% (0.6–1.5 of 20 trials) across conditions. Outliers. An ANOVA of outliers yielded a significant Group × Treatment interaction, F (1, 8) = 7.36, P < 0.05, a main effect of Treatment, F (1, 8) = 7.36, P < 0.05, and no other significant effects, P -values > 0.10. In Group High, more outliers occurred under drug (M = 1.7%) than placebo (M = 0.1%), whereas in Group Low, the number of outliers did not differ under drug versus placebo (M = 0.4%). The main effect of Treatment reflected more outliers overall under drug (M = 1.0%) versus placebo (M = 0.3%). The lack of effects involving prime condition, along with the very small absolute frequency of outliers (<0.4 of 20 trials) indicates that outliers did not bias effects of the experimental factors. Baseline performance. To ensure that differences in baseline RT did not influence priming effects, a preliminary 2 (Group) × 2 (Treatment) × 2 (Target Type) ANOVA assessed RT to ALC and NEU targets paired with explicitly unrelated, NEU primes. The analysis yielded a marginal Group × Treatment interaction, P = 0.058, a marginal main effect of Treatment, P = 0.094, and no other marginal or significant effects, P -values > 0.23. In Group High, mean (SD) RT (ms), averaged across ALC and NEU targets, tended to be slower under drug, 889 (148) than placebo, 776 (100), whereas in Group Low, mean RT was slightly faster under drug, 755 (128) than placebo, 763 (115). More importantly, the lack of significant effects involving Target Type ensures that any significant effects for RT difference scores (i.e. priming) as a function of Target Type cannot be attributed to variation in baseline RT under either Treatment. Parametric requirements. Mauchly's test was non-significant for all factors, P -values > 0.16. Therefore, the variance–covariance matrix did not deviate from the expected pattern under the null hypothesis (i.e. sphericity). The mean inter-correlation for RT difference scores across levels of Prime Type under placebo was r = 0.52, P < 0.05, one-tailed. Skewness statistics (SPSS v. 11.5; Explore Function, SPSS Inc., Chicago, IL) for each prime–target condition found no significant effects, P -values > 0.07. Thus, the RT distributions were symmetrical. These results confirm that the data meet the requirements of repeated-measures ANOVA. Primary analysis of RT difference scores A 2 (Group) × 2 (Treatment) × 3 (Prime Type) ANOVA of RT difference scores for the NEG-ALC, ALC-ALC and NEU-NEU (R) conditions yielded a significant three-way interaction, F (2, 20) = 3.51, P < 0.05, and no other significant effects, P -values > 0.39. Figure 1A and B shows the mean (SEM) RT scores to ALC and NEU targets under placebo in Groups High and Low, respectively. For each class of targets, significant priming effects are indicated by RT difference (Δ RT * ) from baseline. Comparing panels A and B reveals that RT values for each baseline condition under placebo were very similar in the two groups. Inspection of the RT difference from baseline for ALC targets reveals that NEG primes induced significant and comparable priming in each group, P -values < 0.05. ALC primes induced significant priming of ALC targets in Group Low, P < 0.05, and a marginally significant trend in priming ( P < 0.05, one-tailed) in Group High. Control NEU targets were also primed significantly by NEU (R) primes relative to NEU (U) baseline in each group, P -values < 0.05. The similar pattern of NEG-ALC priming effects in each group under placebo ensures that any group differences in this condition under diazepam cannot be attributed to drug-free patterns of association. Fig. 1. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under placebo. (A) Left panel shows scores for participants who received 15-mg diazepam on the drug session ( n = 6). (B) Right panel shows scores for participants who received 5-mg diazepam on the drug session ( n = 6). Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related prime conditions. Larger positive Δ RT scores indicate greater priming. * Δ RT: Simple effect of prime relative to baseline, P < 0.05 Fig. 1. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under placebo. (A) Left panel shows scores for participants who received 15-mg diazepam on the drug session ( n = 6). (B) Right panel shows scores for participants who received 5-mg diazepam on the drug session ( n = 6). Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related prime conditions. Larger positive Δ RT scores indicate greater priming. * Δ RT: Simple effect of prime relative to baseline, P < 0.05 Figure 2A and B shows the mean (SEM) RT difference scores for ALC and NEU targets in Groups High and Low under diazepam. Comparing panels A and B reveals marked differences in the pattern of effects for the two groups in the case of ALC targets. In addition, priming of NEU targets by NEU (R) primes appears to differ in Groups High and Low, although the pattern of effects is the opposite to that which emerged for ALC targets. Simple effects analyses determined that, in Group High, NEG primes and ALC primes each led to a significant slowing of RT to ALC targets relative to baseline (i.e. an interference effect), P -values < 0.05. In contrast, both NEG and ALC primes significantly primed RT to ALC targets in Group Low, P -values < 0.05. Fig. 2. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under diazepam. (A) Left panel shows scores for participants who received 15-mg dose ( n = 6). (B) Right panel shows scores for participants who received 5-mg dose. Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related Prime conditions. Larger positive Δ RT scores indicate greater priming. Negative Δ RT scores indicate interference (slower RT to a class of Targets in response to putatively/explicitly related Primes as opposed to explicitly unrelated Primes). * Δ RT: simple effect of Prime relative to baseline, P < 0.05 Fig. 2. Open in new tabDownload slide Mean ( SEM ) lexical decision response time (RT; ms) to ALC and NEU target words under diazepam. (A) Left panel shows scores for participants who received 15-mg dose ( n = 6). (B) Right panel shows scores for participants who received 5-mg dose. Priming effects are denoted by the RT difference from baseline (Δ RT) in the respective putatively/explicitly related Prime conditions. Larger positive Δ RT scores indicate greater priming. Negative Δ RT scores indicate interference (slower RT to a class of Targets in response to putatively/explicitly related Primes as opposed to explicitly unrelated Primes). * Δ RT: simple effect of Prime relative to baseline, P < 0.05 Of particular relevance to the hypothesis, was the comparison of RT difference scores in the NEG-ALC and ALC-ALC conditions in Group Low under diazepam versus placebo (see Figs 1B versus 2B ). Simple effects analyses revealed that, in these participants, diazepam significantly increased the magnitude of the difference score relative to placebo, reflecting a drug-induced augmentation of priming in the NEG-ALC condition, P < 0.05. Although the mean RT difference score on ALC-ALC trials also increased under drug versus placebo in Group Low, the degree of increase (30%) was not statistically significant, P > 0.25. Comparing Fig. 2A versus B reveals that Group High displayed significant NEU→NEU (R) priming under drug, P < 0.05, whereas Group Low did not. Comparing the degree of priming under drug versus placebo in each group reveals that NEU-NEU priming increased significantly under drug in Group High, P < 0.01, but decreased significantly under drug in Group Low, P < 0.01. Self-report ratings Desire for alcohol There were no differences in pre-capsule ‘desire’ ratings as a function of Group or Treatment, P -values > 0.43. A 2 × 2 ANOVA of desire ratings yielded no significant effects, P -values > 0.28. The mean (SD) scores for placebo versus drug, respectively, were 2.7 (3.3) versus 2.0 (1.1) in Group High and 0.8 (1.3) versus 1.3 (2.4) in GroupLow. This pattern was consistent with the expected decrease under the high dose and increase under the low dose, although simple effects did not achieve significance, P -values > 0.07. PANAS positive and negative affect scores There were no significant differences in pre-capsule baseline scores on either the positive or negative subscale, P -values > 0.14. An ANOVA of PANAS positive subscale scores yielded a Group × Treatment interaction, F (1, 10) = 6.50, P < 0.05, and no other significant effects, P -values > 0.48. Simple effects analyses revealed that, in Group High, mean (SD) positive affect ratings were lower under diazepam, 24.3 (9.7) than placebo 28.8 (10.7), P < 0.01, whereas in Group Low, positive affect was higher under diazepam, 24.0 (10.0) than placebo 21.2 (8.7), P < 0.05. The ANOVA of PANAS negative subscale scores yielded no significant effects, P -values > 0.30. The overall mean (SD) negative subscale rating was 12.8 (5.9). Physiological effects An ANOVA of heart rate at 1-h post-capsule yielded no significant effects, P -values > 0.37. Mean (SD) overall heart rate was 77.8 (6.8 beats/min). An ANOVA of systolic/diastolic blood pressure yielded no significant effects involving Treatment or Group, P -values > 0.06. Mean (SD) overall scores were 129/86 (13/10) mm Hg. Thus, neither dose of diazepam significantly altered cardiovascular activity in these problem drinkers. Correlational analyses The correlation between NEG→ALC priming (RT difference score) and desire for alcohol for all participants under placebo was significant, r = 0.61, P < 0.05. Thus, in the absence of a drug, greater priming of ALC targets by NEG words was associated with stronger motivation for alcohol. The partial correlation between NEG→ALC priming and desire for alcohol under diazepam, controlling for variation under placebo was non-significant in Group High, r = 0.34, P > 0.25, but significant in Group Low, r = 0.79, P = 0.03, one-tailed. Thus, under the low dose, diazepam-induced changes in NEG→ALC priming directly predicted drug-induced changes in desire for alcohol. DISCUSSION The results of this study confirm that a single high dose of diazepam can abolish NEG→ALC priming and ALC→ALC priming in problem drinkers. Indeed, this dose led to significant interference of these associations (slowing of RT) relative to the NEU-ALC baseline, and relative to placebo. The interference effect was selective for ALC associations; the high dose led to significantly increased NEU→NEU (Related) priming relative to placebo. These findings provide a controlled experimental demonstration that high-dose diazepam selectively dampens implicit priming of ALC associations in problem drinkers. The finding that high-dose diazepam slowed RT in the NEG-ALC and ALC-ALC conditions relative to NEU-ALC baseline indicates that, under diazepam, motivationally relevant cues impeded access to ALC targets. Such interference effects imply that, under drug, the NEG and ALC primes were activating an alternative set of associations rather than simply failing to activate ALC-related ones ( Dagenbach et al ., 1990 ). High doses of benzodiazepines have been found to induce unusual (i.e. categorically unrelated, statistically rare) associations to common words in problem drinkers ( Weingartner et al ., 1998 ). This suggests that the interference observed under diazepam in Group High reflected activation by NEG and ALC primes of unusual associates that were incompatible with alcohol. The near doubling of variability in the ALC-ALC condition and 50% increase in variability in the NEG-ALC condition relative to NEU-ALC baseline (see error bars in Fig. 2 ) is consistent with this explanation. NEU words significantly primed RT to NEU, categorically-related targets under the high dose. Thus, NEU cues can still facilitate access to their associates under high dose diazepam. This implies that the motivational relevance of a stimulus influences the degree of idiosyncratic associations exhibited by problem drinkers under high-dose benzodiazepines. The increase in NEU→NEU priming coupled with the interference to ALC targets under drug in Group High further implies that high-dose diazepam renders NEU concepts relatively more accessible than ALC targets in problem drinkers. Given the potential therapeutic value of such a shift, the generality of this effect deserves further investigation. The low-dose manipulation permitted a test of the motivational hypothesis. The results of Group Low clearly support this hypothesis. The 5-mg dose led to increased NEG→ALC priming and a parallel, though non-significant, increase in ALC→ALC priming. Conversely, this dose decreased NEU→NEU priming under diazepam relative to placebo. Thus, low-dose diazepam selectively enhanced the ability of motivationally relevant cues to activate NEG→ALC associations. The intermediate effects on ALC→ALC priming may reflect the involvement of both categorical and motivational associations in this condition. The reason for the decline in NEU→NEU priming under the low dose is unclear. It is possible that motivationally irrelevant words lose priming potential during states of heightened motivation. This is in line with previous research, which found that RT to NEU word stimuli was inhibited following administration of a drug prime ( Zack and Poulos, 2004 ). Whatever the mechanism, the decline in NEU→NEU priming under the low dose makes the increase in priming of ALC concepts all the more remarkable. Importantly, the findings for Group Low preclude an associative strength explanation for the dampening effects seen in Group High. According to this explanation, high-dose diazepam negated ALC because they are weaker than NEU, categorical associations. However, by this account, low-dose diazepam should also have reduced alcohol associations, albeit to a lesser extent. In fact, the opposite occurred. The correlational results further support the motivational hypothesis. Under placebo, NEG→ALC priming significantly predicted the desire for alcohol in the full sample. Thus, under drug-free conditions, NEG→ALC priming is directly linked to motivation for alcohol. Under low-dose diazepam, NEG-ALC priming continued to predict the desire for alcohol when co-variation between these indices under placebo was partialled out. Thus, low-dose diazepam exerts parallel effects on subjective motivation for alcohol and implicit priming of ALC associations in problem drinkers. Under the high dose, the correlation between NEG→ALC priming and ‘desire’ was non-significant. This may reflect the idiosyncratic associations alluded to earlier, which would have obscured systematic variation in NEG→ALC priming and weakened the correlation with ‘desire’. Future research that assesses problem drinkers' spontaneous associations to NEG and ALC words under high-dose diazepam could test this possibility. The correlational and mean effects strongly support the assertion that diazepam was modulating ALC-related motivational systems. The mean effects also mirror the bi-directional dose–response effect of benzodiazepines on alcohol self-administration in animals ( Petry, 1997 ). Taken together, the findings indicate that drug stimuli with similar neurochemical properties exert parallel effects on semantic and motivational priming. From a clinical standpoint, the results for Group High imply that a high dose of a benzodiazepine may reduce the (primed) bias to think about or seek alcohol during NEG states. Because priming effects are involuntary, a medication that can inhibit them could be valuable to someone who is motivated to avoid drinking in response to NEG (e.g. an anxious problem drinker). At the same time, the inability to achieve such cognitive dampening effects without benzodiazepines, could lead to dependence on these medications to remain abstinent from alcohol. In light of this, it is noteworthy that a GABA-ergic medication with low abuse potential (the GABA-B agonist, baclofen) has shown promise for treating alcohol dependence ( Addolorato et al ., 2002 ; Ameisen, 2005 ). Interestingly, baclofen has also been found to exert amnestic effects in animals and humans ( Sandyk and Gillman, 1985 ; Pitsikas et al ., 2003 ). Given this profile of effects, the possibility that baclofen can modulate the alcohol memory network in problem drinkers is an important issue for future investigation. REFERENCES Addolorato, G., Caputo, F., Capristo, E. et al. ( 2002 ) Baclofen efficacy in reducing alcohol craving and intake: a preliminary randomized controlled study. Alcohol and Alcoholism 37 , 504 –508. Ameisen, O. ( 2005 ) Complete and prolonged suppression of symptoms and consequences of alcohol-dependence using high-dose baclofen: a self-case report of a physician. Alcohol and Alcoholism 40 , 147 –150. Boucart, M., Biederman, I., Cuervo, C. et al . ( 2002 ) Effect of benzodiazepines on structural and conceptual/lexical priming. Psychopharmacology 165 , 43 –50. Curran, H. V. ( 1986 ) Tranquillising memories: a review of the effects of benzodiazepines on human memory. Biological Psychology 23 , 179 –213. Dagenbach, D., Carr, T. H. and Barnhardt, T. M. ( 1990 ) Inhibitory semantic priming of lexical decisions due to failure to retrieve weakly activated codes. Journal of Experimental Psychology: Learning, Memory, and Cognition 16 , 328 –340. de Wit, H., Dudish, S. and Ambre, J. ( 1993 ) Subjective and behavioural effects of diazepam depend on its rate of onset. Psychopharmacology 112 , 324 –330. Hill, A. B. and Paynter, S. ( 1992 ) Alcohol dependence and semantic priming of alcohol related words. Personality and Individual Differences 13 , 745 –750. Jackson, A., Stephens, D. N., and Duka, T. ( 2003 ) Lorazepam substitutes for the alcohol stimulus in social drinkers. Psychopharmacology 166 , 181 –187. Jackson, A., Duka, T., and Stephens D. N. ( 2003 ) Effects of alcohol and lorazepam during extinction of alcohol self-administration in rats. Journal of Psychopharmacology 17 , 293 –299. Liu, X. and Weiss, F. ( 2004 ) Nitric oxide synthesis inhibition attenuates conditioned reinstatement of ethanol-seeking, but not the primary reinforcing effects of ethanol. Alcoholism: Clinical and Experimental Research 28 , 1194 –1199. Neely, J. H. ( 1991 ) Semantic priming effects in visual word recognition: A selective review of current findings and theories. In, Basic processes in reading: Visual Word Recognition , Besner, D. and Humphreys, G. W. eds, pp. 264–337. Erlbaum, Hillsdale, NJ. Nelson, D. L., Schreiber, T. A. and McEvoy, C. L. ( 1992 ) Processing implicit and explicit representations. Psychological Review 99 , 322 –348. Petry, N. M. ( 1997 ) Benzodiazepine-GABA modulation of concurrent ethanol and sucrose reinforcement in the rat. Experimental and Clinical Psychopharmacology 5 , 183 –194. Pitsikas, N., Rigamonti, A. E., Cella, S. G. et al. ( 2003 ) The GABA-B receptor and recognition memory: possible modulation of its behavioral effects by the nitrergic system. Neuroscience 118 , 1121 –1127. Poulos, C. X. and Zack, M. ( 2004 ) Low-dose diazepam primes motivation for alcohol and alcohol-related semantic networks in problem drinkers. Behavioural Pharmacology 15 , 503 –512. Ross, H. E., Gavin, D. R. and Skinner, H. A. ( 1990 ) Diagnostic validity of the MAST and the alcohol dependence scale in the assessment of DSM-III alcohol disorders. Journal of Studies on Alcohol 51 , 506 –513. Sandyk, R. and Gillman, M. A. ( 1985 ) Baclofen-induced memory impairment. Clinical Neuropharmacology 8 , 294 –295. Skinner, H. A. and Allen, B. A. ( 1982 ) Alcohol dependence syndrome: measurement and validation. Journal of Abnormal Psychology 91 , 199 –209. Sobell, L. C. and Sobell, M. B. ( 1992 ) Timeline Followback: a technique for assessing self-reported alcohol consumption. In Measuring Alcohol Consumption, Psychosocial And Biological Methods , Litten, R. Z. and Allen, J. eds, pp 41–72. Humana, Totowa, NJ. Spielberger, C. D., Gorsuch, R. L. and Lushene, R. E. ( 1970 ) Manual for the State-Trait Anxiety Inventory . Palo Alto, CA: Consulting Psychologists Press. Vagg, P. R., Spielberger, C. D. and O'Hearn, T. P. J. ( 1980 ) Is the State-Trait Anxiety Inventory multidimensional? Personality and Individual Differences 1 , 207 –214. Watson, D., Clark, L. A. and Tellegen, A. ( 1988 ) Development and validation of brief measures of positive and negative affect, the PANAS scales. Journal of Personality and Social Psychology 54 , 1063 –1070. Weingartner, H. J., Rawlings, R., George, D. T. et al. ( 1998 ) Triazolam-induced changes in alcoholic thought processes. Psychopharmacology 138 , 311 –317. Wilkinson, D. A. and LeBreton, S. ( 1986 ) Early indicators of treatment outcome in multiple drug users. In Treating Addictive Behaviours , Miller, W. R. and Heather, N. eds, pp. 239–261. New York: Plenum. Winer, B. J. ( 1971 ) Statistical Principles in Experimental Design , 2nd edn. McGraw-Hill, New York. Zack, M. and Poulos, C. X. ( 2004 ) Amphetamine primes motivation to gamble and gambling-related semantic networks in problem gamblers. Neuropsychopharmacology 29 , 195 –207. Zack, M., Toneatto, T. and MacLeod, C. M. ( 1999 ) Implicit activation of alcohol concepts by negative affective cues distinguishes between problem drinkers with high and low psychiatric distress. Journal of Abnormal Psychology 108 , 518 –531. Zack, M., Toneatto, T. and MacLeod, C. M. ( 1999 ) Clinical use of benzodiazepines and decreased memory activation in anxious problem drinkers. Alcoholism: Clinical and Experimental Research 23 , 174 –182. © The Author 2006. Published by Oxford University Press on behalf of the Medical Council on Alcohol. All rights reserved

Journal

Alcohol and Alcoholism – Oxford University Press

Published: Nov 1, 2006

There are no references for this article.