Read DD_Behav.pdf text version


Vol. 29, No. 12 December 2005

Impulsive Responding in Alcoholics

Jennifer M. Mitchell, Howard L. Fields, Mark D'Esposito, and Charlotte A. Boettiger

Background: Impaired decision-making is one diagnostic characteristic of alcoholism. Quantifying decision-making with rapid and robust laboratory-based measures is thus desirable for the testing of novel treatments for alcoholism. Previous research has demonstrated the utility of delay discounting (DD) tasks for quantifying differences in decision-making in substance abusers and normal controls. In DD paradigms subjects choose between a small, immediate reward and a larger, delayed reward. Methods: We used a novel computerized DD task to demonstrate that abstinent alcoholics (AA, n 14) choose the larger, delayed option significantly less often than control subjects (n 14; p 0.02). This difference in choice tendency was independent of subject age, gender, years of education, or socioeconomic status. Results: All subjects discounted as a function of reward delay and amount, with alcoholics demonstrating steeper discounting curves for both variables. This tendency to discount delayed rewards was positively correlated with subjective reports of both alcohol addiction severity (Drug Use Screening InventoryRevised, Domain 1, p 0.01), and impulsivity (Barratt Impulsivity Scale-11, p 0.004). Novel aspects of this new paradigm include an element of time pressure, an additional experimental condition that evaluated motor impulsivity by assessing the ability to inhibit a prepotent response, and another control condition to requiring nonsubjective choice. Conclusions: Non-alcoholic controls and alcoholics did not differ on motor impulsivity or nonsubjective choice, suggesting that the differing choice behavior of the two groups was due mainly to differences in cognitive impulsivity. Key Words: Decision-Making, Choice, Inhibitory Control, Human.


diagnostic characteristic of alcoholism is the inability to refrain from drinking even in the face of severe consequences. This impairment may reflect faulty decisionmaking, in which consequences are not effectively taken into account before choosing a course of action. This would represent a form of cognitive impulsivity. Alternatively, such a failure of "willpower" may reflect poor inhibitory control, in which one is unable to suppress an undesired action. This would suggest a contribution of motor impulsivity to maladaptive drinking behavior. This latter possibility is supported by anecdotal reports from alcoholics that they are unable to resist drinking alcohol, despite being aware of adverse consequences. Alcohol abuse is often accompanied by extreme financial, social, and psychological repercussions. Such problems may occur as an indirect result of alcohol abuse, but the possibility remains that alcoholism and other life problems may result, in part, from

From the Ernest Gallo Clinic and Research Center (JMM, HLF, CAB), University of California, San Francisco, CA; Wheeler Center for the Neurobiology of Addiction (HLF), University of California, San Francisco, CA; Helen Wills Neuroscience Institute (MD), University of California, Berkeley, CA. Received for publication July 18, 2005; accepted September 15, 2005. Reprint requests: Charlotte A. Boettiger, Ernest Gallo Clinic and Research Center, University of California San Francisco, 5858 Horton St., Suite 200, Emeryville, CA 94608; FAX: (510) 985-3101; E-mail: [email protected] Copyright © 2005 by the Research Society on Alcoholism. DOI: 10.1097/01.alc.0000191755.63639.4a


common decision-making and/or impulse-control impairments. Whether decision-making deficits are a cause or a result of alcoholism, understanding the neural basis of such impairment is critical for the development of new treatments. One approach toward this goal is to develop behavioral paradigms that are compatible with cognitive neuroscience techniques for measuring physiological correlates of behavior. To this end, the implementation of a delay-discounting (DD) task may prove especially useful, as DD is the only decision-making paradigm demonstrated to be independent of IQ within the normal range (Monterosso et al., 2001; Kirby and Petry, 2004). DD tasks can be briefly summarized as follows: subjects are given choices between a small, sooner reward and a larger, delayed reward. Traditionally, the outcomes of a series of such choices are used to estimate the present subjective value of a delayed reward as a function of delay time, yielding hyperbolic temporal discount curves (Mazur, 1987; Rachlin, 2000). Following the influential rational theory of addiction proposed by Becker and Murphy (1988), a number of studies have supported their hypothesis that preference for the present and underweighting of delayed consequences contributes to addictive behaviors. For example, steeper discounting functions have been associated with alcohol abuse (Vuchinich and Simpson, 1998; Petry, 2001a), opiate abuse (Madden et al., 1997; Bretteville-Jensen, 1999; Kirby et al., 1999; Madden et al., 1999; Odum et al., 2000; Kirby and Petry, 2004),

Alcohol Clin Exp Res, Vol 29, No 12, 2005: pp 2158­2169



stimulant abuse (Coffey et al., 2003; Kiby and Petry, 2004), pathological, gambling (Petry and Casarella, 1999; Petry, 2001b; Dixon et al., 2003; although cf Holt et al., 2003), and cigarette smoking (Bickel et al., 1999; Mitchell, 1999; Odum et al., 2002; Baker et al., 2003; Reynolds et al., 2004). However, to date, task designs used in previous studies of delay discounting in substance abusers are not compatible with cognitive neuroscience techniques, such as functional MRI (fMRI). Our goal in the present study was to design a DD task that would be compatible with fMRI or other neurophysiological techniques. In addition, our task design incorporates two novel control conditions. First, in previous designs, motor impulsivity, or a tendency to generate an unintended response, could have accounted for the tendency to make impulsive choices. An alternative explanation is that "impulsive" choices are the result of cognitive impulsivity within the context of controlled motor responding. While resulting in the same outcome, the two possibilities are likely to be implemented quite differently in terms of underlying neural circuitry and have distinct implications for treatment strategies. To address this confound, our task incorporates a condition in which subjects are required to select the outcome they don't want instead of the one they desire. In addition, we include a condition in which subjects are required to make a selection based on objective, rather than subjective criteria. By randomly varying the delay, amount, discount, and trial type, and incorporating a time limit for responding, subjects maintain attention, and are less likely to use a strategy for consistency based on recall of previous choices. In this way, each choice is made with less interference from previous trials. Moreover, to be considered successful, it was essential that the task be able to detect significant differences in the choice behavior of alcoholics and control subjects. We report here the results of using a novel computerized DD task to assess differences in both cognitive impulsivity and motor impulsivity between abstinent alcoholics (AA) and control subjects (CS). Finally, while DD procedures are often used as a quantitative measure of impulsivity (Kirby and Marakovic, 1996; Rachlin, 2000; Ainslie, 2001; Barkley et al., 2001; Bickel and Marsh, 2001; Critchfield and Kollins, 2001; Reynolds and Schiffbauer, 2005), few studies have explicitly confirmed a correlation between delay discounting behavior and subjective measures of impulsivity (Kirby et al., 1999; Crean et al., 2000; Kirby and Petry, 2004). As impulsivity may be an important risk factor for substance abuse (Jentch and Taylor, 1999; Moeller et al., 2001; Gerald and Higley, 2002), we employed an independent measure of impulsivity, the Barratt Impulsivity Scale-11 (BIS), to confirm that behavior in this paradigm correlates with subjective impulsivity scores. We hypothesized that alcoholics would demonstrate greater cognitive impulsivity, as measured by greater discounting of delayed rewards, and greater motor impulsivity, manifest as poorer inhibitory control, than control subjects. In addition, we hypothesized that these behavioral measures would correlate with BIS scores.


Subjects Due to the impact of alcohol intoxication on decision-making (Steele & Josephs, 1990), subjects were required to be abstinent from alcohol. In addition, to minimize systematic differences in physiological state between groups, we elected not to recruit active alcoholics, as this would have required those subjects to be in a state of acute withdrawal. Alcoholic subjects were recruited on the basis of a minimum of two weeks of abstinence from alcohol. Three subjects reported relapse to drinking alcohol after the initial screening, but prior to the experiment. Abstinence duration for these three subjects was 1, 5, and 5 days . For the remaining subjects abstinence duration ranged between two weeks and 17 years (by self-report); for the entire group, the median reported abstinence time was approximately five months. We chose to test these subjects, as 1) none reported or demonstrated any detectable withdrawal symptoms at the time of the experiment, and 2) our abstinence duration measure relied on self report rather than an objective measure, such as liver enzyme levels, and therefore cannot be taken as an absolute measure. Abstinent alcoholic subjects (AA) were recruited through flyers posted by local recovery organizations and through referral by addiction treatment professionals. Control subjects (CS) were recruited by flyers posted in the local community. Subjects were screened by telephone prior to participation. All subjects were between 19 and 38 years of age and were screened for neurological disease, psychoactive medications, current treatment for other psychological disorders, and addiction to substances other than alcohol. Control subjects were also screened for alcohol abuse. Due to the high incidence of chronic tobacco use among alcoholics, nicotine addiction was not considered grounds for exclusion for either group. Smokers comprised approximately 33% of both groups, and smokers were allowed smoking breaks if needed, thus, attentional deficits due to nicotine withdrawal were unlikely to differentiate the two groups. Subjects who met our inclusion criteria in the phone screening were invited to participate in the task, which took place in a laboratory setting on the University of California, Berkeley campus. Subjects provided written, informed consent, as approved by the U.C.B. Committee for the Protection of Human Subjects. Subjects were compensated approximately $60 for their participation. Subjects completed the behavioral inventories in approximately 1 hr. The delay discounting (DD) task lasted about one additional hour. A total of 16 alcoholic and 15 control subjects participated in the task. Two alcoholics failed to meet our performance criterion on the DD task (see below), and are excluded from the behavioral and demographic data. One control subject did not identify themselves as an alcoholic, but had a score 8 on the Alcohol Use Disorders Identification Test (AUDIT), indicating a possible drinking problem, and was therefore excluded from all analyses. Behavioral Inventories Immediately prior to participating in the behavioral task, subjects filled out a series of questionnaires. The AUDIT (Saunders et al., 1993) was used to verify alcohol abuse severity, as it is the most sensitive alcohol screening instrument with strong correlation to DSM-III-R diagnosis for alcoholism (Bradley et al., 1998) and demonstrated cross-cultural validity (Cherpital, 1998). Domain I of the Drug Use Screening Inventory-Revised (DUSI; Tarter, 1990) provided additional information regarding the severity of each subject's alcohol abuse behaviors. DUSI scores are reported in terms of the percent of affirmative answers from Domain 1, part B. For the AUDIT & DUSI questionnaires only, subjects were asked to answer according to the year prior to achieving abstinence if they had ceased consuming alcohol. Due to our interest in whether affective and/or behavioral difference between groups could impact choice behavior, we administered the Beck Depression Inventory (BDI; Beck et al., 1996), the Depression Anxiety and Stress Scales (DASS; Lovibond and Lovibond, 1993), the Barratt Impulsivity Scale-11 (BIS; Patton et al., 1995), Rotter's Locus of Control Scale (LOC; Rotter, 1966), the South Oaks Gambling Screen (SOGS; Lesieur and Blume, 1987), and the Future Time Perspective Inventory (FTPI; Wallace, 1956). We also collected information about



Fig. 1. Illustration of behavioral paradigm. A) The temporal sequence of events are shown for one example Want (W) trial. Illumination of a fixation point ("Ready") indicated the initiation of each trial. The instruction cue was then displayed for two seconds, alerting the subject to the upcoming trial type. The four instruction cues were color coded by type: Want green, Don't want red, Sooner yellow, Larger magenta. The two options ("earlier" and "later"; see Materials and Methods for values) were then presented while the instruction cue remained on the screen. The choices remained on the screen for 2 sec, however subjects had a total of 6 sec to indicate their choice following the appearance of the two options. B) Depiction of the four trial types. The four trial types included W, Don't Want (DW), and two controls (CON): Sooner and Larger. Trial ratio was 1/2 for the W condition and 1/6 each for the other three trial types.

familial alcohol abuse using the Family Tree Questionnaire (FTQ; Mann et al., 1985). Occupation and education information were collected to calculate the Hollingshead Socio-economic Status (SES) score (Hollingshead, 1975). Delay Discounting Task Subjects were positioned in front of a color computer monitor and instructed in the use of a keypad for response selection. Subjects were given a brief practice session immediately prior to beginning the experiment. Each behavioral session consisted of eight blocks of 47 or 48 trials with rest periods between blocks as needed (total duration 1 hr ). Trials were one of four conditions: "Want" (W), "Don't want" (DW), and two types of control decision conditions, "Sooner" and "Larger," which are considered together as "Control" (CON). Trial types were randomly intermixed, with weighted ratios of 1/2 for the want W condition and 1/6 for the other three trial types. Trial type was indicated by a cue word that instructed the subject how to choose between the two options that appeared following the cue (Fig. 1). On each trial, two options were presented, each consisting of a dollar amount and a point in time. On every trial, the delayed option was one of six "full" amounts ($1, $2, $5, $10, $20, or $100) at one of five future delays (one week, two weeks, one month, three months, or six months). The earlier option was always a lesser (or "discounted") amount available at an earlier time. In most cases, "Today" was the earlier option, but three months was the earlier alternative for some six-month-delay trials. The discounting rate randomly varied among the following four percentages: 70, 85, 90, or 95%. This range of discounts was selected based on pilot studies, which demonstrated that subjects needed to "think over" their selection with options in this range (data not shown). Interestingly, this discount range corresponds to "difficult" decisions in a recent report showing that difficult and "easy" decisions appear to recruit distinct brain circuits (McClure et al., 2004). We refer to these two alternatives as the "earlier" versus the "later" option. The earlier and later options randomly appeared on the right or the left side of the computer monitor. In the W condition, subjects were asked to choose the

option they preferred, as though they would actually receive the money at the time specified. Hypothetical rewards were used based on results from numerous studies comparing choices for real versus hypothetical monetary rewards in discounting paradigms (Critchfield and Kollins, 2001; Johnson and Bickel (2002); Madden et al., 2003; Madden et al., 2004; Lagorio and Madden, 2005). Subjects then indicated their choice by pressing one of two buttons on a keypad. In the DW condition, subjects were asked to make the same evaluation, but to press the button corresponding to the opposite choice. The Sooner & Larger (CON) conditions served as controls to ensure that subjects comprehended and were compliant with the task instructions. In these trials, no subjective evaluation was required; subjects simply pressed the button corresponding to the side with the sooner time point or larger amount of money, respectively. The order of trial types did not vary across subjects; however, the delayed amount, delay time, and discount rate were randomly selected on each trial. Two subjects were excluded from behavioral data analysis due to chance performance levels on CON trials. Data Analysis While DD paradigms have traditionally focused on the "indifference point" (Kirby and Petry 2004; Vuchinich and Simpson, 1998; Petry 2001; Crean et al., 2000), we chose to develop a novel design, compatible with functional magnetic resonance imaging. Instead of indifference pointbased discount rates, we used two indices of discounting: the proportion of earlier choices made, a measure previously used by Ainslie and Monterosso (2003), and the ratio of the cumulative dollar amount chosen to the maximum dollar amount available. These values were calculated across all W trials, as well as divided according to delay time and delayed choice amount. As the results for six months versus either Today or three months did not differ, these trial types were collapsed for analysis purposes. In contrast to breakpoint analyses, which result in hyperbolic discount rates (k), our design does not allow robust determination of k. This is due to the fact that our maximum discount was fixed at 70%. The range of 70-95% was chosen on the basis of pilot experiments (data not

ALCOHOLISM AND IMPULSIVITY Table 1. Personal and Family History of Alcoholism Group AA (n CS (n 14) 14) FTQ (p 0.008) 2.9 0.9 2.6 0.9 AUDIT (p 0.001) 20.9 4.9 8.4 2.0 DUSI (I) (p 0.001) 71 15 28% 14%


quantification of the degree of difference between individual subjects as well as between groups. Psychometric Comparisons of AA and CS Groups To detect whether any observed group differences in either choice behavior or inhibitory control were due to behavioral trait differences between the AA and CS groups, we collected a series of subjective measures. In cases where the AA and CS groups differed significantly (Table 2), we tested the correlation between that behavioral measure and the task behavior parameter of interest. Alcoholics, as a population, suffer from depression at a higher rate than nonalcoholics (Merikangas and Gelernter, 1990). Since depressed mood could theoretically increase the discounting of future rewards, we collected depression data via the BDI. Consistent with previous findings, the AA group reported significantly more depression than the CS group, however the mean scores for both groups were below that considered diagnostic for clinical depression (12; Table 2). Alcoholics also suffer disproportionately from anxiety disorders (Weiss and Rosenberg, 1985). Since anxiety or stress could affect performance on this task, particularly in terms of inhibitory control, we collected generalized "emotional distress" data via the Depression Anxiety and Stress Scales (DASS; Table 2). Consistent with our BDI data, the groups differed significantly on the depression subscale (6.7 6.8 vs. 1.4 2.4, p 0.01). The DASS anxiety subscale also demonstrated a higher level of anxiety in the AA group compared to the CS group (5.5 7.6 vs. 0.5 0.7, respectively, p 0.022). Another factor with the potential to impact explicit choice behavior is Locus of Control, a trait also previously reported to differ between alcoholics and controls (Mills and Taricone, 1991). Our findings are consistent with previous findings, in that the AA group demonstrated a significantly more external Locus of Control than the CS group (Table 2). However, for the alcoholics in our sample, both the mean (10.9) and the median (10) fell into the middle range, reflecting neither internal nor external dominance in attribution style (Rotter, 1966). Another factor that might impact delay discounting behavior, potentially resulting in a difference between groups is orientation toward the future. Substance abusers have been reported to operate on a foreshortened time horizon, although some studies have not found systematic differences between alcoholics and controls (Smart, 1968; Hulbert and Lens, 1988; Petry et al., 1998). The AA and CS groups did not significantly differ in the maximum event extension on the Future Time Perspective Inventory (FTPI, part I; Table 2). There was also no detectable difference between groups in mean event extension or Pearson correlation (p 0.24, and p 0.89, respectively). Nonparametric tests (Mann-Whitney U test) also failed to distinguish the two groups on any FTPI measures (data not shown),

Data from the Family Tree Questionnaire (FTQ), Drug Use Screening Inventory, Domain I (DUSI-I) and Alcohol Use Disorders Identification Test (AUDIT). Values are reported as mean standard deviation. Reported p-values reflect the results of unpaired two-tailed comparisons between groups. Exact p-values reported unless p 0.001. AA, Abstinent alcoholic; CS, Control subject.

shown) that determined most control subjects would discount at least some of the time within this range. This range was selected as one in which subjects in both AA and CS groups would both evaluate the choices, rather than taking a strategy of always picking the sooner or later option. However, limiting our discounts to this range likely results in an underestimation of k. This suspicion was borne out by empirical test; we derived k from the cumulative dollar ratio (CDR) for each delay time (D) according to the following equation (Mazur, 1987): CDR 1/(1 kD) (1)

and taking the mean k across delay times. Consistent with expectation, the AA group demonstrated significantly larger values of k, and thus significantly faster discounting rates than the CS group (0.005 0.001, and 0.001 3 10 4, respectively (mean SEM), p 0.02, t 3.78). However, the values for both are much lower than expected for true values of k. Rather, delaydiscounting frequency as a function of delay time and delayed amount were well fit by logarithmic functions. Both the slope and intercept of these fits were measured using the built in curve-fitting function in Microsoft Excel. Thus as our task appears not to accurately estimate k, this is not an appropriate dependent measure. For single factor statistical comparisons between groups we used unpaired two-tailed t-tests. For multi-factorial comparisons, we used mixed ANOVAs, using group as a between subjects factor. Where sphericity assumptions were violated, a Greenhouse-Geisser correction for inhomogeneity of variance was applied. Moreover, to ensure the validity of parametric statistical tests, when data were not normally distributed, appropriate arcsine-root transformation was applied prior to statistical comparison, (instances specified in Results), nonparametric tests were applied using an Excel add-in (Analyze-It Software, Ltd.). Regression analysis and macro-based analysis of covariance (ANCOVA) were also performed using commercially available software (Excel). Reward preference in DW trials was inferred as the rejected option. In other words, when asked to choose the reward they did not want, a subject choosing the left option was assumed to want the right option. The absolute difference between the actual and inferred earlier:later choice ratios for each delay was used as a measure of inhibitory control.


Familial Alcohol Abuse The AA group in our study reported a significantly greater number of family members with possible or definite drinking problems than did the control subjects (CS; Table 1). While this represents an unsubstantiated subjective report, the FTQ has been validated as a reliable instrument for assessing familial alcohol abuse (Mann et al., 1985). Table 1 also includes summary data regarding the individual alcohol abuse histories of our subjects. The fact that the AA group scored significantly higher than the CS group on measures of alcohol abuse (AUDIT & DUSI) is not unexpected. These data support the presence of a personal history of alcohol abuse and provided

2162 Table 2. Psychometric Comparison of Subject Groups Group p-value AA (n 14) CS (n 14) BDI p 0.004 10.5 10.0 1.9 2.1 BIS p 0.001 75.1 6.6 57.0 9.9 DASS p 0.028 21.4 23.8 6.1 5.7 FTPI p 0.195 23.9 17.1 34.1 23.1 LOC p 0.009 10.9 3.4 7.4 2.5


SOGS p 0.044 1.4 2.2 0.1 0.5

Psychometric data gathered via inventories. Values are reported as mean standard deviation. Reported p -values reflect the results of unpaired two-tailed comparison between groups. Exact p -values reported unless p 0.001. AA, Abstinent alcoholic; CS, Control subject; BIS, Barratt Impulsivity Scale-11; BDI, Beck Depression Inventory; DASS, Depression, Anxiety, and Stress Scale; FTPI, Future Time Perspective Inventory; LOC, Rotter's Locus of Control Scale; SOGS, South Oaks Gambling Screen. Table 3. Reaction Times and Accuracy Data Group AA (n CS (n 14) 14) CON 1510 1343 118 83 W 1796 1898 124 110 DW 2026 2112 123 127 ACC 94 97 2% 2%

Group comparison of reaction times for each trial type and CON accuracy. There was a significant effect of trial type on reaction time (RT) based on a mixed effects ANOVA (p 0.001); however, there was not a significant effect of group on RT (p 0.962). Accuracy in CON trials did not show a statistically significant difference between groups (unpaired 2-tailed t-test; p 0.249).

perhaps because abstinence reflects healthy future orientation in the AA group. As we measured subjects' valuation of hypothetical monetary rewards, gambling tendencies could factor into explicit choices. Moreover, pathologic gambling has been found to co-occur with substance abuse (Crockford and el-Guebaly, 1998). Indeed, South Oaks Gambling Screen scores indicated significantly greater gambling tendencies among the AA group (Table 2). Finally, as our paradigm was designed to objectively measure impulsive responding, we wished to determine whether choice behavior and inhibitory control in this task correlated with a subjective measure of impulsiveness. Consistent with previous findings (Nagoshi et al., 1991; Chalmers et al., 1993; Petry 2001a), the AA group was significantly more impulsive (Table 2), based on responses on the BIS. Reaction Times and Accuracy Both subject groups showed slower reaction times (RT) for conditions requiring greater cognitive processing (Table 3). This supports the inference that subjects are engaging additional processes for the subjective decision trials (W) compared to the objective decision trials (CON). Furthermore, these data also indicate that DW trials, in which subjects were instructed to decide which option they preferred and select the opposite one, engaged further processes beyond those in the W condition. A one-way ANOVA with group (AA, CS) as a between subjects factor demonstrated a significant main effect of trial type on RT (F(1.2,32.2) 61.05, p 0.001), no main effect of group (F(1,26) 0.002, p 0.962), but a trend toward a group trial type interaction (F(1.2,32.2) 3.252, p 0.073). Post hoc tests demonstrated no significant differences between groups for any trial type (minimum p 0.24) suggesting that the AA and CS groups showed a different pattern of increasing RT with trial type. Both subject groups demonstrated high accuracy in the CON trials (Table 3), and performance on these trials was not significantly different

Fig. 2. Comparison between groups of cumulative discounting behavior in W trials. A) Over the course of the experiment, the AA group chose the earlier option significantly more frequently than the CS group (t(26) 2.92, p 0.007). B) The AA group accumulated significantly less total theoretical gain over the course of the experiment (t(26) 3.10, p 0.005). Gain is expressed as a ratio of selected options to the maximum money available (i.e., if subjects had chosen the later option on every trial).

between groups (t(26) 1.18, p 0.249). Taken together, these results indicate that both groups understood and were compliant with the task instructions. The AA Group Chose the Earlier Option Significantly More Often Than the CS Group Over the course of the entire experiment, the AA group more frequently chose the earlier option in the W trials (Fig. 2A). To assess the cumulative result of each subject's decision-making, we calculated a cumulative gain ratio, based on the sum of all W trial selections divided by the total possible monetary gain (corresponding to a "later" choice in every W trial). As shown in Fig. 2B, the decisions



Fig. 4. Comparison of reward magnitude discounting functions. Semilog plot of earlier:later choice ratio as a function of the later reward amount. Data reflect mean SEM. Curves represent logarithmic fit the group averaged data, with regression terms shown for each group.

Fig. 3. Comparison of temporal discounting functions. A) Ratio of earlier:later choices as a function of later choice delay time. Data reflect mean SEM. Curves represent logarithmic fit of the group averaged data, regression terms are shown for each group. B) Comparison by subject group of individual curve fit parameters. Left panel: Group comparison of discount curve intercepts. The AA group had significantly higher intercept terms (t (26) 3.94, p 0.001). Right panel: In contrast, the slope of the discount curves did not show a significant difference between groups (t (26) 1.60, p 0.123).

of the AA group resulted in significantly less average cumulative gain than for the CS group. Effect of Delay Time and Reward Amount on Reward Choice The frequency with which subjects chose the earlier amount in the W trials varied as a function of the delay time (Fig. 3A). A mixed effects 2 5 ANOVA found significant main effects of both group (F(1,26) 8.809; p 0.006), and delay time (F(2.4,61.7) 14.83; p 0.001). A subsequent 2 2 mixed effects ANOVA considering only the 7 days and 180 days delay times found a highly significant effect of delay time (F(1,26) 9.461; p 0.005). Neither the full (F(1,26) 1.211; p 0.309) nor the reduced (F(1,26) 1.769; p 0.195) ANOVA revealed a significant Group Delay interaction, suggesting that the slope of the temporal discounting function did not differ between groups, but that the intercept of these functions was different. Direct comparison of the slope and intercept terms of the logarithmic fit to each subject's temporal discounting curve supported this conclusion (Fig. 3B). The intercept differences suggest that within this range of discount rates, the delay time which initiates discounting in the AA group is shorter than in the CS group. Additionally, the lack of difference between slopes indicates that once temporal discounting is

initiated, the rate of further discounting with increasing delay is similar between groups. The tendency to choose the earlier amount in the W trials also varied as a function of the delayed reward amount (Fig. 4). A 2 6 mixed effects ANOVA found significant main effects of group (F(1,26) 6.715; p 0.015) and amount (F(2.4,62.2) 17.41; p 0.001). This analysis did not detect a significant Group Amount interaction (F(2.4,62.2) 2.395; p 0.090), suggesting that the slope of the reward magnitude discounting function did not differ between groups, but that the intercept of these functions was different. Direct comparison of the slope and intercept terms of the logarithmic fit to each subject's reward magnitude discounting curve partially supported this conclusion. The average slopes did not differ between groups. Means were 0.05 0.02 and 0.08 0.02 for the AA and CS groups, respectively (t(26) 0.93, p 0.361). However, the intercepts showed only a trend toward a difference between groups. The mean intercepts were 0.84 0.06 and 0.61 0.11 for the AA and CS groups, respectively (t(26) 1.86, p 0.074). The simplest explanation for these findings is that the AA group differs from the CS group in their evaluation of relative reward magnitudes. In other words, AA perceive the subjective values of $20 and $14 to be in a larger ratio than do CS, thereby increasing the likelihood of AA selecting the earlier, lesser amount. AA Impulsive Choices Are Not Due to Impaired Inhibition While the AA group responded differently than the CS group in the W condition, it is unclear whether this behavior was "intentional," i.e., whether the AA group actually preferred the earlier to the later options. Intentional preference differences would be consistent with recent literature regarding temporal discounting of alcoholics (Vuchinich and Simpson, 1998; Petry, 2001a). However, our task includes an element of mild time pressure that those tasks did not, which could have caused unintended responses. As a result, group differences in the W condition may be due to increased unintended responding or inhibitory failure in the AA group. To

2164 Table 4. Demographic Data Group p-value AA (n 14) CS (n 14) Age (yrs) 0.178 29 6 26 6 Gender (% male) 0.246 71% 46% Ethnicity (% white) 0.115 50% 77%


Education (yrs) 0.010 15 3 18 4

SES 0.036 38 14 49 12

Demographic data gathered via self-report. Values are reported as mean standard deviation. Reported p-values reflect the results of unpaired two-tailed comparison between groups, or 2 test (gender and ethnicity). AA, Abstinent alcoholic; SES, Hollingshead Socio-economic status score.

is evident between the AA and CS groups (Fig. 5B). This was statistically verified with a mixed-design repeated measures ANOVA, which found no main effect of group on choice mismatch (F(1,25) 0.654; p 0.426). There was also no significant main effect of delay time on choice mismatch (F(4,100) 1.244; p 0.297), and no significant interaction between group and delay time F(4,100) 1.199; p 0.316). A second repeated measures group amount ANOVA found no significant effects (all F's 1.0). Thus, it appears that the AA group does not show increased motor impulsivity relative to the CS group, and that differences in explicit choice behavior are not attributable to inhibitory failure. Discounting Behavior Is Not Dependent on Demographic Factors While every attempt was made to match subject groups for demographic variables, in the end, the two groups differed somewhat in several domains (Table 4). To test whether demographic differences between groups might be driving the group difference in choice behavior, we 1) stratified the groups according to categorical variables (gender and ethnicity) and tested for differences between groups, 2) tested the correlation between continuous variables (education and SES) and discounting behavior, and 3) used an ANCOVA to determine whether controlling for demographic factors eliminated the difference between groups in discounting behavior. Direct comparison of male and female subjects revealed no significant difference in earlier:later choice ratio (62 9, and 56 10, respectively (mean SEM), p 0.62). Likewise, white and nonwhite subjects did not demonstrate significantly different choice behavior (55 7, and 62 9, respectively (mean SEM), p 0.33). These results strongly suggest that gender and ethnic differences between the AA and CS subject groups are not driving the differences in choice behavior between these groups. As our experimental groups differed in terms of average years of education and SES, we tested the correlation between these values and earlier:later choice ratio. We found that the tendency to choose a smaller immediate reward did not correlate with years of education (r 0.07, t 0.35, p 0.731). If we used general education level scores instead of actual years, we again saw no correlation (r 0.06, t 0.29, p 0.771). Similarly, SES was not significantly correlated with earlier:later choice ratio (r 0.06, t

Fig. 5. Motor impulsivity comparison between groups. Data shown reflects mean SEM A) Average absolute value of the difference between each subject's W earlier:later choice ratio and DW inferred earlier:later choice ratio. The AA and CS groups showed statistically equivalent discrepancies between their inferred choice ratio and actual choice ratio, a measure of inhibitory failure (t (26) 1.49, p 0.148). B) Plot of inferred discounted choice ratio (from DW trials) as a function of later choice delay time. Solid lines represent logarithmic fit to the DW data, while the dashed lines depict fits to W data (Fig. 3A). Data demonstrates that members of the AA group are no more likely to make mistakes in the DW condition than members of the CS group.

test this possibility, we included the DW condition, which, due to its relatively lower frequency, requires a heightened level of control to avoid making erroneous responses. In addition, in the DW condition, subjects were instructed to select one option and then to execute a response indicating the opposite of that selection, which should require a greater level of response inhibition relative to the W condition. If alcoholics are making more discounted choices in the W condition due to inhibitory failure, we would expect to see even greater inhibitory failure in the DW condition for that group. However, when we compared the absolute value of the difference between each subject's discounted choice ratio (W condition) and the inferred discounted choice ratio (DW condition), we found that the AA and CS groups were not significantly different (t(26) 1.49, p 0.148; Fig. 5A). Indeed, there was a tendency for greater mismatch between W and inferred W choices in the CS group. When the inferred choice ratio is plotted as a function of delay time, and the data are compared to the fits derived from the W trials, no systematic difference



Fig. 6. Severity of alcohol abuse history positively correlates with impulsive choice frequency. The linear fit is derived from the data of all subjects considered together. The two subjects groups are plotted separately for clarity.

group with regard to abstinence duration, we examined the correlation between time abstinent (in days) vs. earlier: later choice ratio within the AA group. We found that time abstinent was not a reliable predictor of delay discounting in the AA group (r 0.22, t 0.8, p 0.451). We also found that a higher frequency of earlier choices was correlated with a greater number of family members who abuse alcohol (r 0.39, t 2.1, p 0.043). This correlation appeared to be mainly driven by the AA group (r 0.33, vs. r 0.02 for the CS group), although neither within group correlation was significant (p 0.24, and p 0.95, for the AA and CS groups, respectively). Discounting Is Positively Correlated with Subjective Impulsivity As shown in Table 2, the AA and CS groups differed on a number of psychometric measures including depression, impulsivity, emotional distress, locus of control, and gambling behavior. Therefore, we tested the correlation between these values and earlier:later choice ratio. We found that the tendency to prefer a smaller, earlier reward did not correlate significantly with a subject's level of emotional distress, as measured by the DASS (r 0.27, t 1.42, p 0.167). We also determined that a subject's time horizon, as measured by either the maximum or mean extension in the FTPI, part I, did not correlate significantly with earlier: later choice ratio (r -0.08, t 0.41, p 0.687; r 0.17, t 0.11, p 0.563, respectively). Similarly, a subject's locus of control rating did not significantly correlate with earlier:later choice ratio (r 0.26, t 1.42, p 0.166). Finally, a subject's tendency to gamble, as measured by the SOGS did not correlate with their choice behavior in this DD task (r 0.03, t 0.15, p 0.882). We did detect a trend toward a significant correlation between depression (BDI) scores and temporal discounting (r 0.37, t 2.04, p 0.051). However, when we used an ANCOVA to test for a difference in the earlier:later choice ratios of the AA and CS groups after controlling for BDI scores, we found that the groups remained significantly different in their discounted choice frequencies (F (1,25) 4.29, p 0.049). This suggests that the difference between groups in choice behavior is not determined solely by differences in depression levels between the two groups. Taken together, these results indicate that the differences between the impulsive choice tendencies of the AA and CS groups were largely independent of differences between the two groups on these psychometric measures. In contrast, we did find significant correlations between discounting behavior and the psychometric parameter of subjective impulsivity (Fig. 7). When we tested the correlation between discounted choice ratio and BIS score, we found a significant positive correlation (r 0.53, t 3.18, p 0.004; Fig. 7A). Significant positive correlations were also found with each of the attention, motor, and nonplanning BIS subscales (Attention: r 0.58, t 3.65, p 0.001;

0.3, p 0.768). In addition, there was no significant correlation within groups between SES and discounting frequency (p 0.55, and p 0.08, for the AA and CS groups, respectively). These results indicate that the differences between the choice tendencies of the AA and CS groups were not due to education or to socioeconomic differences between the two groups. Alcohol Addiction Severity Is Positively Correlated with Discounted Choice Frequency While our AA and CS groups clearly differed in terms of their choice behavior, suggesting that alcoholism and increased discounting of delayed rewards are coexisting traits, a stronger link would be made by testing for a direct correlation between alcohol addiction severity and discounting behavior. When we tested the correlation between earlier:later choice ratio and DUSI scores, not surprisingly, we indeed found a significant correlation (r 0.48, t 2.8, p 0.01; Fig. 6). Moreover, when we used an ANCOVA to test for a difference in the discounted choice ratios of the AA and CS groups after controlling for DUSI scores, we found that the two groups were no longer significantly different (p 0.308). This suggests that the difference in choice behavior between groups is tightly correlated with the severity of alcohol abuse history. However, we did not find significant within-group correlation of DUSI and impulsive choice (p 0.26, and p 0.84, for the AA and CS groups, respectively), perhaps reflecting our small sample size. The correlation between impulsive choice and AUDIT scores showed a non-significant trend (r 0.36, t 2.0, p 0.060). However, the AUDIT has three subscales: consumption, alcohol-dependence, and alcohol-related harm. When we excluded the consumption subscale and only considered the dependence and harm subscales, as these better reflect addiction severity, we again found a significant positive correlation between addiction severity and impulsive choices (r 0.41, t 2.3, p 0.031). However, again this correlation did not remain within groups (p 0.52, and p 0.76, for the AA and CS groups, respectively). As there was heterogeneity within our AA




Alcoholism and Impulsivity Impulsivity is widely accepted as an essential feature of substance abuse disorders, including alcoholism (Jentsch and Taylor, 1999; Moeller et al., 2001). In fact, impulsivity has been proposed to be a key risk factor for alcoholism (Gerald and Higley, 2002). However, the relationship between alcoholism and impulsivity remains somewhat controversial. While some studies suggest that alcoholics are significantly more impulsive than control subjects (Colder & Chassen, 1997; Soloff et al., 2000; Mulder 2002; Simons et al., 2004), others have found no correlation between alcoholism and impulsivity (Lejoyeux et al., 1988). Our results support the idea that alcoholism is associated with elevated levels of both cognitive impulsivity, as measured by choices in our DD task, and impulsive trait, as measured by the BIS-11. However, our results do not support the conclusion that abstinent alcoholics experience elevated motor impulsivity (Liraud et al., 2000). This dissociation of cognitive and motor impulsivity is consistent with previous findings suggesting the relative independence of systems mediating long-term delay of gratification and short-term behavioral inhibition (Crean et al., 2002). Moreover, there is evidence that opioid, amphetamine, nicotine, and cocaine addicts also discount delayed rewards to a greater extent than do controls on DD tasks (Bickel et al., 1999; Bretteville-Jensen, 1999; Kirby et al., 1999; Bickel & Marsch, 2001; Heyman & Dunn, 2002; Coffey et al., 2003). This suggests that the form of impulsivity assayed by DD tasks is a general feature of addiction, representing an important risk factor, a long-term consequence, or both. Although subjects in the AA group appeared to have normal future orientation, perhaps reflecting the process of recovery, both impulsivity and impaired decision-making persisted in abstinence from alcohol. In fact, impulsivity was independent of abstinence duration. Thus, impulsivity may represent a potential risk factor not only for alcoholism, but also for relapse to drinking. An important caveat with regard to these results is that the sample sizes are rather small. Thus, despite the robustness and statistical significance of our results, it is uncertain that our findings represent the population as a whole. We can assert that experimental volunteers who self-identify as alcoholics who have chosen to cease drinking alcohol appear to be both more impulsive and to more readily discount delayed monetary rewards than do control volunteers. It's possible that our AA volunteers represent a subtype of alcoholic phenotype, a possibility that may merit genetic investigation.

Fig. 7. Subjective impulsivity, impulsive choices, and alcoholism. A) Subjective ratings of impulsivity positively correlate with impulsive choice frequency. B) Subjective ratings of impulsivity also positively correlate with alcohol addiction severity. For both panels, the linear fit is derived from the data of all subjects considered together. The two subjects groups are plotted separately for illustrative purposes.

Motor: r 0.44, t 2.46, p 0.021; Nonplanning: r 0.39, t 2.16, p 0.04). Of these, the only significant withingroup correlation was for CS impulsive choice versus BISAttention (r 0.539, t 2.21, p 0.047). When we applied an ANCOVA to test for a difference in the earlier:later choice ratios of the AA and CS groups after controlling for BIS scores, we found that the two groups were no longer significantly different (F(1,25) 0.77, p 0.387). This result suggests that our DD task provides a reliable measure of how impulsive a given subject is, regardless of alcohol abuse history. As we also found a strong correlation between individual alcohol addiction severity and impulsive choice frequency, we tested for a correlation between individual BIS and DUSI scores. We found that the BIS and DUSI scores were indeed highly correlated (r 0.78, t 6.36, p 0.001; Fig. 7B). Moreover, this correlation remained significant within the AA group (r 0.58, t 2.49, p 0.028), indicating that impulsive traits reliably predict alcohol addiction severity. There was a trend toward a similar correlation within the CS group (r 0.47, t 1.84, p 0.091). Similar results were obtained when we compared BIS and AUDIT scores, whether full scale (r 0.67, t 4.59, p 0.001), or dependence and harm subscales only (r 0.66, t 4.45, p 0.001) were analyzed.

Impulsivity and Delay Discounting One conception of impulsivity is an inability to delay gratification. This view is supported by reports that individuals with impulsive disorders discount future rewards more



steeply (Crean et al., 2000; Barkley et al., 2001; Petry, 2001). Moreover, numerous studies using animal models of impulsivity also demonstrate temporal discounting of rewards (Evenden, 1999; Cardinal et al., 2001; Kheramin et al., 2004; Winstanley et al., 2004). Further evidence that DD is dissociable from motor impulsivity or inhibitory control deficits comes from studies of tryptophan depletion in humans, which impairs inhibitory control without altering temporal discounting curves (Crean et al., 2002). However, the supposition that discounting is a valid indicator of impulsivity is not without controversy (Monterosso and Ainslie, 1999; Critchfield and Kollins, 2001; Frederick et al., 2002; Reynolds and Schiffbauer, 2005). While some investigators have found correlations between discounting behavior and subjective impulsivity scores (Swann et al., 2002), others have found weak or no correlation between the two measures (Mitchell, 1999; Lane et al., 2003). Some of these differences are likely due to the wide range of methodological approaches used in the study of delay discounting. The strong correlations that we report here between BIS scores and delay discounting behavior thus provide valuable reinforcement for the validity of delay discounting behavior as a measure of impulsivity. Possible Neural Mechanisms Underlying Impulsive Choice Behavior A number of possibilities exist regarding the neural basis of impulsive choice. Clinical data and lesion studies point to orbitofrontal cortex dysfunction as a possible substrate (Berlin et al., 2004; Winstanley et al., 2004). However, there is also evidence in humans that frontal lobe damage has no impact on temporal discounting of rewards (Fellows and Farah, 2005). Several subcortical structures have also been implicated in animal studies of impulsive choice, for example the basolateral amygdala (Winstanley et al., 2004), the core of the nucleus accumbens (Cardinal et al., 2001), and the hippocampus (Cheung and Cardinal, 2005). Alternatively, specific neurochemical abnormalities may underlie impulsive choice. The serotonergic system is the most commonly studied with regard to delayed reward preference (Ho et al., 1999). However, while low CSF serotonin metabolite levels have been associated with impulsive behavior (Moeller et al., 2001; Gerald and Higley, 2002), serotonin depletion does not alter DD behavior in humans (Crean et al., 2002). This latter result makes us inclined to suspect that other systems are more critically involved in impulsive decision-making. For example, the endogenous opioid system has been implicated in an animal model of delayed reward choice (Kieres et al., 2004). The possibility that endogenous opioids regulate preference for immediate versus delayed rewards remains to be tested in humans. Alcoholics are known to experience reduced endogenous opioid signaling even after prolonged abstinence, lending support to this hypothetical mechanism (del Arbol et al., 1995). Another intriguing possibility for increased reward

discounting in abstinent alcoholics is HPA axis dysregulation. This possibility is supported by reports that abstinent alcoholics have relatively lower basal cortisol levels (Anthenelli et al., 2001), a condition that may dissipate over many years of abstinence (del Arbol et al., 1995). Moreover, Takahashi (2004) recently reported that basal cortisol levels are negatively correlated with discounted reward choice in humans. There is also evidence that dopaminergic systems may play an important role in deciding between immediate and delayed rewards (Wade et al., 2000). In summary, we have found that abstinent alcoholics discount delayed rewards to a greater extent than do controls and that this discounting tendency is sensitive to both delay time and reward magnitude. This enhanced temporal discounting is associated with more severe alcohol addiction and higher subjective ratings of impulsivity but is not due to a relative impairment of inhibitory control. Future investigations of the underlying neural mechanisms of this choice behavior may prove illuminating for the development of therapeutic interventions for alcoholism and other impulsive behaviors.


We thank H. Alexander, P. Kalra, and V. Tavares for valuable technical assistance. This work was supported by DOD Center grant #W81XWH-04-1-0154 (CAB & HLF), State of California funds for medical research on alcohol and substance abuse through the University of California (HLF), the UCSF Wheeler Center for the Neurobiology of Addiction (JMM), and the Henry H. Wheeler Jr. Brain Imaging Center.


Ainslie G (2001) Breakdown of Will. Cambridge, UK: Cambridge University Press. Ainslie G, Monterosso G Jr (2003) Building blocks of self-control: increased tolerance for delay with bundled rewards. J Exp Anal Behav 79:37­48. Anthenelli RM, Maxwell RA, Geracioti TD, Hauger R (2001) Stress hormone dysregulation at rest and after serotonergic stimulation among alcohol-dependent men with extended abstinence and controls. Alcohol Clin Exp Res 25:692­703. Baker F, Johnson MW, Bickel WK (2003) Delay discounting in current and never-before cigarette smokers: similarities and differences across commodity, sign, and magnitude. J Abnorm Psychol 112:382­392. Barkeley RA, Edwards G, Laneri M, Fletcher K, Metevia L (2001) Executive functioning, temporal discounting, and sense of time in adolescents with attention deficit hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD). J Abnorm Child Psychol 29:541­556. Beck AT, Steer RA (1987) Manual for the Revised Beck Depression Inventory. Psychological Corp, San Antonio, TX. Becker GS, Murphy KM (1988) A theory of rational addiction. J Political Econ 96:675­700. Berlin HA, Rolls ET, Kischka U (2004) Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions. Brain 127:1108­1126. Bickel WK, Odum AL, Madden GJ (1999) Impulsivity and cigarette smoking: delay discounting in current, never, and ex-smokers. Psychopharmacology (Berl) 146:447­454.



Bickel WK, Marsch LA (2001) Toward a behavioral economic understanding of drug dependence: delay discounting processes. Addiction 96:73­86. Bradley KA, Bush KR, McDonell MB, Malone T, Fihn SD (1998) Screening for problem drinking: comparison of CAGE and AUDIT. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. J Gen Intern Med 13:379­388. Bretteville-Jensen AL (1999) Addiction and discounting. J Health Econ 18:393­407. Cardinal RN, Pennicott DR, Sugathapala CL, Robbins TW, Everitt BJ (2001) Impulsive choice induced in rats by lesions of the nucleus accumbens core. Science 292:2499­2501. Chalmers D, Olenick NL, Stein W (1993) Dispositional traits as risk in problem drinking. J Subst Abuse 5:401­410. Cherpitel CJ (1998) Differences in performance of screening instruments for problem drinking among blacks, whites and Hispanics in an emergency room population. J Stud Alcohol 59:420­426. Cheung TH, Cardinal RN (2005) Hippocampal lesions facilitate instrumental learning with delayed reinforcement but induce impulsive choice in rats. BMC Neurosci 13:36­. Coffey SF, Gudleski GD, Saladin ME, Brady KT (2003) Impulsivity and rapid discounting of delayed hypothetical rewards in cocaine-dependent individuals. Exp Clin Psychopharmacol 11:18­25. Colder CR, Chassin L (1997) Affectivity and impulsivity: Temperamental risk for adolescent alcohol involvement. Psych Addictive Behav 11:83­97. Crean JP, deWit H, Richards JB (2000) Reward discounting as a measure of impulsive behavior in a psychiatric outpatient population. Exp Clin Psychopharmacol 8:155­162. Crean JP, Richards JB, deWit H (2002) Effect of tryptophan depletion on impulsive behavior in men with or without a family history of alcoholism. Behav Brain Res 136:349­357. Critchfield TS, Kollins SH (2001) Temporal discounting: basic research and the analysis of socially important behavior. J Appl Behav Anal 34:101­122. Crockford D, el-Guebaly N (1998) Psychiatric comorbidity in pathological gambling: A critical review. Can J Psychiatry 43:43­50. delArbol JL, Aguirre JC, Raya J, Rico J, Ruiz-Requena ME, Miranda MT (1995) Plasma concentrations of beta-endorphin, adrenocorticotropic hormone, and cortisol in drinking and abstinent chronic alcoholics. Alcohol 12:525­529. Dixon MR, Marley J, Jacobs EA (2003) Delay discounting by pathological gamblers. J Appl Behav Anal 36:449­458. Evenden JL (1999) Varieties of impulsivity. Psychopharmacology (Berl) 146:348­361. Fellows LD, Farah MJ (2005) Dissociable elements of human foresight: a role for the ventromedial frontal lobes in framing the future, but not in discounting future rewards. Neuropsychologia 43:1214­1221. Frederick S, Loewenstein G, O'Donoghue T (2002) Time discounting and time preference: A critical review. J Econ Literature 40:351­401. Gerald MS, Higley JD, (2002) Evolutionary underpinnings of excessive alcohol consumption. Addiction 97:415­425. Heyman GM, Dunn B (2002) Decision biases and persistent illicit drug use: an experimental study of distributed choice and addiction. Drug Alcohol Depend 67:193­203. Ho M-Y, Mobini S, Chiang T-J, Bradshaw CM (1999) Theory and method in the quantitative analysis of "impulsive choice" behavior: implications for psychopharmacology. Psychopharmacology (Berl) 146:362­372. Hollingshead A (1975) Hollingshead's Four Factor Index of Social Status. New Haven, CT: Yale University Press. Holt DD, Green L, Myerson J (2003) Is discounting impulsive?. Evidence from temporal and probability discounting in gambling and nongambling college students. Behav Processes 64:355­367. Hulbert RJ, Lens W (1988) Time perspective, time attitude, and time orientation in alcoholism: a review. Int J Addict 23:279­298. Jentsch JD, Taylor JR (1999) Impulsivity resulting from frontostriatal dysfunction in drug abuse: implications for the control of behavior by reward-related stimuli. Psychopharmacology (Berl) 146:373­390.

Johnson MW, Bickel WK (2002) Within-subject comparison of real and hypothical money rewards in delay discounting. J Exp Anal Behav 77:129­146. Kieres AK, Hausknecht KA, Farrar AM, Acheson A, deWit H, Richards JB (2004) Effects of morphine and naltrexone on impulsive decision making in rats. Psychopharmacology (Berl) 173:167­174. Kheramin S, Body S, Ho MY, Velazquez-Martinez DN, Bradshaw CM, Szabadi E, Deakin JFW, Anderson IM (2004) Effects of orbital prefrontal cortex dopamine depletion on inter-temporal choice: a quantitative analysis. Psychopharmacology (Berl) 175:206­214. Kirby KN, Marakovic N (1996) Modeling myopic decisions: evidence for hyperbolic delay-discounting within subjects and amounts. Organizational Behav Hum Decision Processes 64:22­30. Kirby KN, Petry NM (2004) Heroin and cocaine abusers have higher discount rates for delayed rewards than alcoholics or non-drug-using controls. Addiction 99:461­471. Kirby KN, Petry NM, Bickel WK (1999) Heroin addicts have higher discount rates for delayed rewards than non-drug-using controls. J Exp Psychol Gen 128:78­87. Lagorio CH, Madden GJ (2005) Delay discounting of real and hypothetical rewards III: steady-state assessments, forced-choice trials, and all real rewards. Behav Processes 69:173­187. Lane SD, Cherek DR, Rhoades HM, Pietras CJ, Tcheremissine OV (2003) Relationships among laboratory and psychometric measures of impulsivity: Implications in substance abuse and dependence. Addictive Disorders Treatment 2:33­40. Lesieur HR, Blume SB (1987) The South Oaks Gambling Screen (SOGS): A new instrument for the identification of pathological gamblers. Am J Psychiatry 144:1184­1188. Lejoyeux M, Feuche N, Loi S, Solomon J, Ades J (1988) Impulse-control disorders in alcoholics are related to sensation seeking and not to impulsivity. Psychiatry Res 81:149­55. Liraud F, Verdoux H (2000) Which temperamental characteristics are associated with substance use in subjects with psychotic and mood disorders? Psychiatry Res 93:63­72. Lovibond SH, Lovibond PF (1993) Manual for the Depression Anxiety Stress Scales (DASS). Psychology Foundation Monograph. New South Wales, Australia. Mann RE, Sobell LC, Sobell MB, Pavan D (1985) Reliability of a family tree questionnaire for assessing family history of alcohol problems. Drug Alcohol Depend 15:61­67. Madden GJ, Petry NM, Badger GJ, Bickel WK (1997) Impulsive and self-control choices in opioid-dependent patients and non-drug-using control participants: drug and monetary rewards. Exp Clin Psychopharmacol 5:256­62. Madden GJ, Bickel WK, Jacobs EA (1999) Discounting of delayed rewards in opioid-dependent outpatients: exponential or hyperbolic discounting functions? Exp Clin Psychopharmacol 7:284­293. Madden GJ, Begotka AM, Raiff BR, Kastern LL (2003) Delay discounting of real and hypothetical rewards. Exp Clin Psychopharmacol 11: 139­145. Madden GJ, Raiff BR, Lagorio CH, Begotka AM, Mueller AM, Hehli DJ, Wegener AA Delay discounting of potentially real and hypothetical rewards: II. Between- and within-subject comparisons. Exp Clin Psychopharmacol 12:251­261. Mazur JE (1987) An adjusting procedure for studying delayed reinforcement. In, JE Mazur, JA Nevin, H Rachlin (eds) Quantitative Analyses of Behavior: Vol 5. The Effect of Delay and of Intervening Events on Reinforcement Value pp 55­73. Hillsdale, NJ: Erlbaum. McClure SM, Laibson DI, Loewenstein G, Cohen JD (2004) Separate neural systems value immediate and delayed monetary rewards. Science 306:503­507. Merikangas KR, Gelernter CS (1990) Comorbidity for alcoholism and depression. Psychiatr Clin North Am 13:613­632. Mills JK, Taricone PF (1991) Interpersonal dependency and locus of control as personality correlates among adult male alcoholics undergoing residential treatment. Psychol Reports 68:1107­1112.



Mitchell SH (1999) Measures of impulsivity in cigarette smokers and non-smokers. Psychopharmacology (Berl) 146:455­464. Moeller FG, Barratt ES, Dougherty DM, Schmitz JM, Swann AC (2001) Psychiatric aspects of impulsivity. Am J Psychiatry 158:1783­1793. Monterosso J, Ainslie G (1999) Beyond discounting: possible experimental models of impulse control. Psychopharmacology (Berl) 146:339­347. Monterosso J, Ehrman R, Napier KL, O'Brien CP, Childress AR (2001) Three decision-making tasks in cocaine-dependent patients: do they measure the same construct? Addiction 96:1825­1837. Mulder RT (2002) Alcoholism and personality. Aust N Z J Psychiatry 36:44­52. Nagoshi CT, Wilson JR and, Rodriguez L (1991) Impulsivity, sensation seeking, and behavioral and emotional responses to alcohol. Alcohol: Clinical and Experimental Research 15:661­667. Odum AL, Madden GJ, Badger GJ, Bickel WK (2000) Needle sharing in opioid-dependent outpatients: psychological processes underlying risk. Drug Alcohol Depend 60:259­266. Odum AL, Madden GJ, Bickel WK (2002) Discounting of delayed health gains and losses by current, never- and ex-smokers of cigarettes. Nicotine Tob Res 4:295­303. Patton JH, Stanford MS, Barratt ES (1995) Factor structure of the Barratt Impulsiveness Scale. J Clin Psychol 51:768­774. Petry NM (2001a) Delay discounting of money and alcohol in actively using alcoholics, currently abstinent alcoholics, and controls. Psychopharmacology (Berl) 154:243­250. Petry NM (2001b) Pathological gamblers, with and without substance use disorders, discount delayed rewards at higher rates. J Abnorm Psychol 110:482­487. Petry NM, Bickel WK, Arnett M (1998) Shortened time horizons and insensitivity to future consequences in heroin addicts. Addiction 93: 729­738. Petry NM, Casarella T (1999) Excessive discounting of delayed rewards in substance abusers with gambling problems. Drug Alcohol Depend 56: 25­32. Rachlin H (2000) The Science of Self Control. Cambridge, MA: Harvard University Press. Reynolds B, Richards JB, Horn K, Karraker K (2004) Delay discounting and probability discounting as related to cigarette smoking status in adults. Behav Processes 65:35­42. Reynolds B, Schiffbauer R (2005) Delay of gratification and delay-discounting: a unifying feedback model of delay-related impulsive behavior. Psychological Rec 55:439­460.

Ross HE, Glaser FB, Germanson T (1988) The prevalence of psychiatric disorders in patients with alcohol and other drug problems. Arch Gen Psychiatry 45:1023­1031. Rotter JB (1966) Generalized expectancies for internal versus external control of reinforcement. Psychol Monographs 80:609­ 628. Saunders JB, Aasland OG, Babor TF, deleaFuente JR, Grant M (1993) Development of the alcohol use disorders identification test (AUDIT). WHO collaborative project on early detection of persons with harmful alcohol consumption ­ II. Addiction 88:791­804. Simons JS, Carey KB, Gaher RM (2004) Lability and impulsivity synergistically increase risk for alcohol-related problems. Am J Drug Alcohol Abuse 30:685­694. Smart RG (1968) Future time perspectives in alcoholics and social drinkers. J Abnorm Psychol 73:81­83. Soloff PH, Lynch KG, Moss HB (2000) Serotonin, impulsivity, and alcohol use disorders in the older adolescent: a psychobiological study. Alcohol Clin Exp Res 24:1609­1619. Steele CM, Josephs RA (1990) Alcohol myopia. Its prized and dangerous effects. Am Psychol 45:921­933. Swann AC, Bjork JM, Moeller FG, Dougherty DM (2002) Two models of impulsivity: relationship to personality traits and psychopathology. Biol Psychiatry 51:988­994. Takahashi T (2004) Cortisol levels and time-discounting of monetary gain in humans. Neuroreport 15:2145­2147. Tarter R (1990) Evaluation and treatment of adolescent substance abuse: A decision tree method. Am J Drug Alcohol Abuse 16:1­46. Vuchinich RE, Simpson CA (1998) Hyperbolic temporal discounting in social drinkers and problem drinkers. Exp Clin Psychopharmacol 6:292­ 305. Wade TR, deWit H, Richards JB (2000) Effects of dopaminergic drugs on delayed reward as a measure of impulsive behavior in rats. Psychopharmacology (Berl) 150:90­101. Wallace. (1956) Future time perspective in schizophrenia. J Abnorm Psychol 52:240­245. Weinstein ND (1998) Accuracy of smokers' risk perceptions. Ann Behav Med 20:135­140. Weiss KJ, Rosenberg DJ (1985) Prevalence of anxiety disorder among alcoholics. J Clin Psychiatr 46:3­5. Winstanley CA, Theobald DE, Cardinal RN, Robbins TW (2004) Contrasting roles for basolateral amygdala and orbitofrontal cortex in impulsive choice. J Neurosci 24:4718­4722. Wright MH, Obitz FW (1984) Alcoholics' and nonalcoholics' attributions of control of future life events. J Stud Alcohol 45:138­143.


12 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Eating Disorders and Alcohol Use Disorders