Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Alcohol. Author manuscript; available in PMC 2007 October 10.
Published in final edited form as:
PMCID: PMC2013928

Influence of Reinforcement Schedule on Ethanol Consumption Patterns in Non-Food Restricted Male C57BL/6J Mice


Ethanol reinforcement should ideally be evaluated in animals that are not food deprived to ensure that the motivation behind its consumption is pharmacological, and not caloric, in nature. The objective of this work was to assess the influence of reinforcement schedule on ethanol intake in non-deprived mice. Male C57BL/6J mice were trained to respond on an ethanol-reinforced lever on a fixed ratio (FR) 4 reinforcement schedule for 10% ethanol (10E). The appetitive and consummatory phases were then procedurally separated by changing the response requirement (RR), so that mice were permitted 30-min continuous 10E access after completion of either four (RR4) or eight (RR8) responses. Phase separation yielded a heightened appetitive drive to acquire 10E access (as indexed by a significant decrease in the latency to first active lever and a trend towards a decrease in the latency to 1st sipper contact) and an augmented level of drinking (2-fold elevation in the ethanol dose consumed). Robust extinction responding on the ethanol-appropriate lever indicated that ethanol was effective as a behavioral reinforcer. These results suggest that the separation of appetitive and consummatory phases of ethanol self-administration may prove useful in future evaluations of the pharmacological and genetic bases of ethanol reinforcement in mice.

Keywords: operant conditioning, alcohol reinforcement, appetitive, consummatory, animal model, drinking patterns

1. Introduction

Manipulations of environmental conditions have been used to enhance the acquisition of self-administration in animal models of drug reinforcement (see Campbell and Carroll, 2000). Chief among these manipulations has been the restriction of food access. De Vry and colleagues demonstrated a relationship between food restriction (and subsequent weight loss) and increased acquisition of intravenous cocaine self-administration in rats, with a 15-25% loss being optimal for acquisition (De Vry et al., 1989). Food restriction continues to be a commonly used tool to promote acquisition of self-administration of psychostimulants, opioids, cannabinoids, ethanol and other drugs in many animal models. In the case of ethanol, extensive characterization of pre- and post-prandial drug presentation on ethanol-reinforced responding and consumption in mice has been conducted (e.g., Middaugh and Kelley, 1999; Middaugh et al., 1999). Although food restriction provides a convenient methodology for drug self-administration studies, the influence of this mnipulation on neurobiological substrates that underlie behavioral reinforcement has been largely overlooked. Pothos and colleagues have hypothesized that food restricted animals escalate drug self-administration as a means to artificially restore neurochemical deficits associated with food restriction (Pothos et al., 1995). Thus, there is a potential for food restriction to alter reward-related neurochemical pathways, which could confound the interpretation of a drug’s rewarding or behaviorally reinforcing effects.

In studies involving ethanol reinforcement and consumption, the use of food restriction harbors an additional complication for the interpretation of experimental findings. Ethanol, unlike other drugs of abuse, possesses a caloric value as well as a pharmacological one. The seminal observation by Richter (1926) that ethanol attenuated food consumption in proportion to the caloric value of self-administered ethanol suggests that animals both defend a total caloric intake set point and treat ethanol as an energy-yielding nutrient (Forsander, 1998; Richter, 1926). It has been argued that the motivation for consuming ethanol under conditions of food restriction is rooted primarily in the drug’s nutrient value rather than its pharmacological properties (see Cunningham et al., 2000). This proposition makes the use of food restriction particularly problematic when measuring the behavior reinforcing effects of ethanol. In contrast, a procedure to evaluate ethanol reinforcement under a non-food restricted condition would have interpretative advantages over procedures that incorporate food restriction with regard to our understanding of ethanol’s salience as a behavioral reinforcer in mice.

Historically, the development of an ethanol self-administration procedure using operant conditioning without food restriction has proven difficult in mice. Although the reinforcing effects of ethanol have been successfully demonstrated when the drug is continuously available (Besheer et al., 2004; Hodge et al., 2006; Olive et al., 2000; Risinger et al., 1998, 1999), data evaluating ethanol reinforcement in mice under a limited access condition is less well studied. Studies with mice responding on a fixed ratio 1 (FR1) (Middaugh et al. 2000; Roberts et al. 2000) or FR3 (Tsiang and Janak, 2006) reinforcement schedule reported ethanol intake approximating 0.5 g/kg, which correspond to values in rats responding on a FR schedule (e.g., Samson, 2000).

The current understanding of conditioned responding for ethanol under a non-food restricted condition suggests that ethanol is a weak behavioral reinforcer in mice. Oral self-administration of ethanol and other drugs supports a much lower response rate than that observed for intravenous infusions of psychostimulants such as cocaine (e.g., Little, 2000). One major difficulty in assessing the reinforcing properties of orally administered drugs (such as ethanol) in animal models is the delayed onset of pharmacological effect (Meisch, 2001), which likely interferes with operant conditioning processes associated with a FR schedule of reinforcement. One modification that may hold promise in the evaluation of ethanol reinforcement in mice is the procedural separation of the appetitive and consummatory phases of ethanol self-administration, as documented by Samson and colleagues in a rat “sipper” model (Samson et al., 1998, 2000). By providing continuous access to ethanol following the completion of a single response requirement (RR), this experimental procedure permits the animal to regulate its consumption (and hence the onset of pharmacology) rather than having intake dictated by intermittent access following repeated response demands associated with a FR reinforcement schedule.

The goal of the present work was to examine the influence of reinforcement schedule manipulations (i.e., FR versus RR) on ethanol-reinforced responding and consumption patterns in non-food restricted mice. A second purpose was to document that the “sipper” model, in which the appetitive and consummatory phases of ethanol self-administration were procedurally separated (Samson et al., 1998, 2000), could be modified for use in mice.

2. Materials and Methods

2.1. Subjects

Twenty-four male C57BL/6J mice (20.6 ± 0.2 grams at 6 weeks of age) were purchased from The Jackson Laboratory (Bar Harbor, ME). Each mouse was individually housed and acclimated to a 12hr/12hr light/dark cycle (lights on at 0600 hrs) for 7 days prior to experimental manipulation. Mice were provided with ad libitum access to food in their home cages. They also had unrestricted water access except during initial sipper training as noted below. Animals were weighed and handled daily. All instrumental training and test sessions were conducted between 1400-1700 hrs. The local Institutional Animal Care and Use Committee approved all procedures in accordance with the guidelines described in the Guide for the Care and Use of Mammals in Neuroscience and Behavioral Research (National Research Council of the National Academies, 2003).

2.2. Apparatus

Daily sessions were carried out in eight modular chambers (21.6 × 17.8 × 12.7 cm) with stainless steel grid floors (Med-Associates Inc., St Albans, VT). Each chamber was outfitted with a house light, two ultra-sensitive retractable levers, two stimulus lights (light-emitting diodes), and a retractable sipper apparatus for extending/retracting a 10 ml graduated pipette with a double ball-bearing metal sipper tube. A lickometer circuit was connected to each metal sipper to monitor lick patterns. Circuits were interfaced to an IBM compatible computer operated by MED-PC software (Med-Associates Inc.). One wall of the chamber contained the retractable levers 11 cm apart (with a stimulus light positioned directly above each lever). The opposing wall included the access portal for the sipper tube. Each chamber was positioned within a sound-attenuating cabinet (61 × 38 × 33 cm; Fisher Custom Woodworking, Portland, OR). An exhaust fan was installed in each cabinet to facilitate ventilation throughout the daily sessions.

2.3. Sipper and lever acquisition training phases

Mice were initially trained to consume a 0.2% w/v sodium saccharin solution (0.2S in tap water; Sigma-Aldrich Company; St. Louis, MO) from a retractable sipper during two, 20 min trials per day for 7 days. Continuous sipper access was permitted during the first two days. The remaining 5 days of sipper training involved a gradual increase in the frequency of limited access opportunities as follows: multiple 4 min access periods interspersed by 1 min time outs resulting in 4 access opportunities per trial (4’-1’ × 4) on day three, 3’-2’ × 4 on day four, 2’-2’ × 5 on day five, 1’-2’ × 7 on day six, and 0.5’-2’ × 8 on day seven. This shaping procedure was designed to habituate the mice to the sound of the retracting/extending sipper and to encourage the mice to consume fluid immediately following sipper presentation.

The lever acquisition phase involved establishing a relationship between the depression of a lever, the activation of a stimulus light (5-sec duration), and the presentation of a sipper containing 0.2S. Mice were trained to press the active lever to obtain 30-sec access to 0.2S on a FR1 reinforcement schedule throughout 30-min sessions. Responding on the inactive lever had no scheduled consequence. To account for side preference, the position of the active lever was counterbalanced between the left and right sides across operant chambers. Over successive training sessions the sipper access time was reduced from 30- to 15-sec to encourage the mice to respond for more sipper presentations. Twelve sessions were required to achieve a minimum criterion of ten 15-sec sipper presentations per session. To elevate motivational state, mice were water restricted for 16 hrs per day prior to sessions throughout the sipper-training phase and during the first 9 sessions of the lever acquisition-training phase. The restriction of water was eliminated during the last two lever training sessions. Water was provided ad libitum in the home cage during all subsequent maintenance and testing phases.

2.4. Saccharin fading procedure and reinforcement schedule manipulation

A modified saccharin fading procedure was utilized to initiate ethanol self-administration (Czachowski et al., 2001; Middaugh et al., 2000; Samson, 1986). Throughout saccharin fading, mice were maintained on a FR1 reinforcement schedule for 15-sec sipper access periods during 30-min sessions. Sessions were conducted 5-6 days per week. Over the ensuing 6-week period, the ethanol concentration was increased in the 0.2S solution [for two sessions each at 0.2% saccharin/3% ethanol (0.2S/3E), 0.2S/6E, 0.2S/9E, and 0.2S/12E], and then saccharin was subsequently faded out in a step-wise manner to eventually yield 10E alone [for 5 sessions each at 0.1S/12E, 0.05S/12E, 0.025S/12E, 0.01S/12E, and 10E]. Ethanol solutions (v/v) were made with ethyl alcohol (200 proof; Pharmco Products, Inc.; Brookfield, CT) in tap water.

Upon completion of saccharin fading, the lever press requirement was increased from FR1 to FR4 across 5-7 daily sessions. Ethanol-reinforced responding on a FR4 schedule was then maintained for an additional 4 weeks, during which period sipper access time was incrementally extended in the following fashion: 9 sessions at 15-sec, 13 sessions at 30-sec and the remainder of sessions at 60-sec. At this point, the appetitive and consummatory phases of operant self-administration were procedurally separated such that the completion of a response requirement of 4 presses (RR4) resulted in 30-min of continuous access to the 10E solution. A 20-min time limit to complete the RR4 was imposed. Throughout a subsequent 2-week period, the response requirement was incrementally increased from RR4 to RR8. All twenty-four mice acquired and consistently completed the FR4, RR4, and RR8 reinforcement schedules.

2.5. Blood ethanol concentration (BEC)

During the last session in which mice were responding on the RR8 schedule, a 20 μl retro-orbital sinus blood sample was collected from a subgroup of the mice (n = 10) to assess BEC at the conclusion of 30-min continuous 10E access. The blood samples were processed, and supernatants were assayed for BEC by gas chromatography as previously described (Gallaher et al., 1996). Seven pairs of external standards with known ethanol concentrations (ranging from 0.25-4.00 mg/ml) were analyzed to construct a standard curve from which unknown concentrations of samples were interpolated.

2.6. Extinction of ethanol-reinforced responding

Extinction of ethanol-reinforced responding was performed as previously described for rats (Bachteler et al., 2005) and mice (Tsiang and Janak, 2006). After mice responded on the RR8 schedule for 16 weeks, the animals were exposed to 30-min extinction sessions during which both levers were extended and the house light remained illuminated. However, responding did not result in sipper presentation or stimulus diode illumination (that previously signaled sipper access). Due to the demand on chamber equipment in the laboratory, a subgroup of mice (n = 10) was tested in this experimental phase.

2.7. Data analysis

Ethanol intake (grams of ethanol per kilogram of body weight; g/kg) was determined from the volume of 10E depleted (to the nearest 1/20 ml) from graduated pipettes and pre-session body weights. Cumulative records of responding and sipper contacts were recorded via MED-PC software. Appetitive measures (latency to first active lever response, response rate, and latency to lick), consummatory parameters (total licks, lick run frequency, lick run size, inter-run interval, lick run duration, lick run rate, and latency to first lick run), and extinction values (responses on previously assigned active and inactive levers) were derived from cumulative records. Patterns of responding and lick run patterns were analyzed by a custom program written for the R Project for Statistical Computing software (version 2.1.1; available at A lick run was experimentally defined as ≥ 5 licks (sipper contacts) with less than a 60-sec pause between successive licks.

All statistical analyses were performed using the SigmaStat version 2.0 software package (Jandel Scientific; San Rafael, CA). One-way repeated measures analysis of variance (ANOVA) was used to evaluate the within-subject effect of reinforcement schedule on appetitive and consummatory measures and to assess the influence of extinction session on responding. The Fisher’s least significant difference multiple comparisons procedure was conducted following a significant main effect of reinforcement schedule or extinction session. Correlation coefficients for the relationship between variables were determined by the Pearson Product Moment correlation test (SigmaStat version 2.0). For all analyses, the threshold for statistical significance was set at p ≤ 0.05. In limited cases, statistical outliers (identified outside the range of the mean ± 2 standard deviations) were excluded from further analyses. In particular, one mouse was eliminated from mean and first lick run measures.

3. Results

As saccharin was incrementally faded from the administered solutions, mean ± SEM ethanol intakes (g/kg dose; total licks) declined as follows: 0.2S/12E (1.20 ± 0.11; 325 ± 30), 0.05S/12E (0.69 ± 0.08; 173 ± 20), and 10E (0.26 ± 0.05; 72 ± 13). Consumption levels of 10E during FR1 and FR2 reinforcement schedules were comparable to those described below for FR4.

3.1. FR4 schedule of reinforcement

The training history during manipulations in sipper access time following the completion of a FR4 schedule of reinforcement is depicted in sessions 1-25 of Figure 1. Increasing the sipper access time from 15- to 60-sec during each 10E presentation did not alter the total licks recorded during the 30-min sessions. The mean ± SEM lick values during the last 3 sessions of each access time were 59 ± 14 (15-sec), 70 ± 16 (30-sec) and 61 ± 12 contacts (60-sec). Similarly, increasing sipper access time did not significantly alter the frequency of sipper presentations, with mean ± SEM frequencies of 8.0 ± 0.8 (15-sec), 7.0 ± 0.8 (30-sec) and 8.1 ± 0.6 (60-sec). A significant positive correlation between total session sipper contacts and sipper presentation frequency was found with sipper access periods of 30-sec (r = 0.382; p = 0.001) and 60-sec (r = 0.278; p = 0.022), but not with 15-sec access.

Figure 1
History of reinforcement schedule manipulations during operant conditioning. Depicted are the mean ± SEM for 24 mice per session. FR refers to a ‘fixed ratio’ schedule of reinforcement whereas RR refers to a single ‘response ...

3.2. Comparison of appetitive response measures across reinforcement schedules

In order to estimate the strength of the 10E solution as a behavioral reinforcer across reinforcement schedules, measures corresponding to the initial 4 responses completed during the FR4 schedule contingency were compared with values from the RR4 and RR8 schedules. All mice completed the response requirement under each reinforcement schedule. Manipulation of the reinforcement schedule significantly influenced the response latency [F(2,46) = 4.40; p < 0.05], with the transition from FR4 to RR4 significantly decreasing this latency by 34% (p < 0.05; Table 1). The RR8 schedule was associated with an even more pronounced reduction in the response latency, when compared to the FR4 schedule (p < 0.01). A significant main effect of reinforcement schedule on response rate also was detected [F(2,46) = 4.44; p < 0.05], with a significant decrease in response rate with the RR8 schedule, when compared to the FR4 schedule (p < 0.01). Latency to 1st lick values (time elapsed from the onset of sipper access) did not to follow a normal distribution. Consequently, these values were log10-transformed, at which point a normal distribution was observed. A trend toward significance for reinforcement schedule on the latency to first lick was noted [F(2,46) = 2.81; p = 0.07], with the transition to RR4, and then to RR8 resulting in latency reductions of 54% and 65%, respectively, when compared to the FR schedule. Although alteration of the reinforcement schedule also significantly affected inactive lever responding [F(2,46) = 11.17; p < 0.001], the slight but significant increase was found only with the RR8 schedule.

Table 1
Appetitive measures during FR and RR reinforcement schedules. Values are the mean ± SEM of the last 3 sessions on each reinforcement schedule. Data for the first completed response requirement in mice on the FR4 schedule is depicted (except when ...

3.3. Comparison of ethanol intake and consumption patterns across reinforcement schedules

Transition from a FR to RR schedule of reinforcement favored a significant elevation in 10E self-administration (see Table 2), as evidenced by the significant increase in ethanol dose [F(2,46) = 22.79; p < 0.001], licks [F(2,46) = 28.10; p < 0.001] and 10E volume [F(2,46) = 23.60; p < 0.001]. Ethanol dose, licks and 10E volume were significantly augmented following the transition from FR4 to RR4 by 93% (p < 0.05), 174% (p < 0.001) and 95% (p < 0.05), respectively. Furthermore, the RR8 schedule was associated with significantly greater total licks than those observed with the FR4 (p < 0.001) and RR4 (p < 0.05) schedules. Likewise, the RR8 schedule significantly increased the 10E dose and 10E volume consumed versus the FR4 and RR4 schedules (p < 0.001 for each comparison). Importantly, significant positive correlations between total licks and 10E dose were observed for the FR4 (r = 0.73, p < 0.001, n = 24), RR4 (r = 0.63, p < 0.001, n = 24) and RR8 (r = 0.80, p < 0.001, n = 24) reinforcement schedules, thereby indicating that licks accurately reflected actual ethanol consumption throughout the limited access sessions.

Table 2
Consummatory measures during FR and RR reinforcement schedules. Ethanol intake and lick run dynamics are depicted for mice permitted either intermittent sipper tube access on a FR4 schedule or 30-min continuous access following the completion of a RR4 ...

Consistent with the augmented intakes noted above during the transition from a FR to a RR schedule, the lick run patterns were altered in a manner commensurate with an enhanced onset and maintenance of ethanol drinking. Table 2 shows that the frequency and size of lick runs were significantly influenced [run frequency: F(2,46) = 46.05; p < 0.001; run size: F(2,46) = 13.29; p < 0.001] by manipulations in reinforcement schedule. Mice responding on a RR4 and RR8 schedule showed elevated lick run frequencies of 168% (p < 0.001) and 195% (p < 0.001), respectively, and increases in lick run sizes of 37% (p < 0.01) and 68% (p < 0.001), respectively, when compared to the FR4 schedule. Notably, the RR8 schedule was affiliated with significantly greater run sizes than those calculated for the RR4 schedule (p < 0.05). Table 2 also shows that there were significant main effects of reinforcement schedule on the size of the largest lick run [F(2,46) = 20.69; p < 0.001], inter-run interval [F(2,40) = 13.20; p < 0.001], lick run duration [F(2,46) = 21.19; p < 0.001], and lick run rate [F(2,44) = 6.72; p < 0.01].

A separate analysis of the initial lick run of the self-administration session was conducted to more specifically delineate the impact of reinforcement schedule manipulations on drinking onset (Table 2). A significant main effect of schedule was determined for the size [F(2,46) = 15.16; p < 0.001], duration [F(2,46) = 7.09; p < 0.01] and lick rate [F(2,43) = 4.35; p < 0.05] of the first run. First lick run durations were significantly increased by approximately 1.9-fold for both the RR4 and RR8 schedules (p< 0.01 for each), whereas first lick run rates fell by 52% (p < 0.01) and 32% (p < 0.05) for the RR4 and RR8 schedules, respectively, when compared to the FR4 schedule. The latency to first lick run also was significantly decreased by the transition from FR4 to RR4 and RR8 schedules [F(2,46) = 14.04; p < 0.001].

3.4. Comparison of cumulative records across reinforcement schedules

Representative cumulative records of responding and sipper contacts for FR4, RR4, and RR8 reinforcement schedules are depicted in Figure 2. An analysis of the FR4 schedule with 60-sec sipper access revealed that only 53 ± 4% of the sipper presentations were accompanied by a sipper contact (see panel A of Figure 2). Furthermore, only 28 ± 4% of the sipper tube presentations yielded a lick run of ≥ 5 licks. Although mice made contact with the sipper tube 66 ± 7% of the time during the first sipper presentation of the session, a point in which the salience of ethanol as a reinforcer would presumably be the greatest, the mean lick run size was only 7 ± 2 contacts (comparable to representative record shown in panel A of Figure 2). These observations indicate that when mice were performing on a FR4 schedule, they failed to reliably consume ethanol following each sipper presentation and the initial access to the 10E solution corresponded to negligible levels of drinking.

Figure 2
Representative cumulative records of operant ethanol self-administration during FR4 (panel A), RR4 (panel B) and RR8 (panel C) reinforcement schedules. The top portion of each panel depicts the pattern of cumulative licks. The gray boxes along the abscissa ...

Consistent with the appetitive responding results detailed above (Table 1), transition to the RR schedules corresponded to decreases in the response latency but an attenuated response rate, as shown in the representative records for the RR4 (Figure 2, panel B) and RR8 (Figure 2, panel C) schedules. Pronounced differences in the onset and maintenance of 10E self-administration between the reinforcement schedules also were readily apparent from the cumulative records, consistent with the first and mean lick run measures reported in Table 2, respectively.

3.5. Extinction of ethanol-reinforced responding

Ten mice were selected for an evaluation of extinction. Ethanol dose and BECs were 0.64 ± 0.16 g/kg and 37 ± 11 mg%, respectively, in mice on the RR8 schedule prior to extinction onset. There was a statistically significant positive correlation between 10E dose and BEC for these mice (r = 0.92; p < 0.001; n = 10). A significant influence of extinction session was determined for responding on the active (ethanol-appropriate) lever [F(10,90) = 16.00; p < 0.001]. Mice robustly responded on the active lever during the first day of extinction when compared to their pre-extinction baseline (p < 0.001; Figure 3). Responding on the active lever remained significantly elevated over the pre-extinction baseline during sessions 2-6 (p < 0.05 for all), but dropped to levels that were not significantly different from baseline during sessions 7-10. A significant influence of extinction session also was observed for responding on the inactive lever [F(10,90) = 2.03; p < 0.05], with small, but significant increases in responding during extinction sessions 1, 5-6 and 10 when compared to baseline values (p < 0.05 for all).

Figure 3
Lever response frequencies throughout the extinction time course. The mean ± SEM active (filled squares) and inactive (open circles) lever responses recorded during 30-min extinction sessions for 10 mice are depicted.

4. Discussion

Marked differences were observed in responding for ethanol in non-food restricted C57BL/6J mice using FR versus RR schedules of reinforcement. Transition from the FR to RR schedule significantly reduced response latency and was associated with a 55-65% decrease in the latency to first lick (Table 1). These findings suggest that the appetitive drive to attain ethanol access was greater with the “sipper” method, in which mice performed a single response requirement to gain 30-min of ethanol access. When mice were on the RR4 and RR8 reinforcement schedules, they also exhibited shorter latencies to begin drinking, larger and more frequent drinking runs, and higher ethanol intake than when responding for ethanol on an FR4 schedule. These drinking patterns reflected an enhanced drinking onset as well as elevated consumption maintenance when mice self-administer ethanol via the ‘sipper’ method. Although the stability in total lick values during 25 consecutive sessions on the FR4 reinforcement schedule (see Figure 1) suggests that the transition between reinforcement schedules was the primary contributor to these alterations in ethanol drinking. However, another possibility is that the changes in appetitive and consummatory measures noted following the advancement of the response requirement from 4 to 8 responses (i.e., RR4 to RR8) were partially attributable to the additional experience of the mice to the ‘sipper’ procedure over time. Nonetheless, when ethanol was withheld in extinction sessions, mice previously responding on the RR8 schedule exhibited an initial “burst” of responding followed by a rapid extinction of this behavior (Figure 3). This pattern of responding during extinction of ethanol self-administration is comparable to recent findings in C57BL/6J mice that were responding for ethanol on an FR3 schedule (Tsiang and Janak, 2006). Collectively, these findings suggest that ethanol was effective as a behavioral reinforcer with both FR and RR schedules (based on extinction responding), whereas aspects of the appetitive and consummatory phases of ethanol self-administration were enhanced with the RR schedule.

Although ethanol intakes achieved during the FR4 schedule (0.14 g/kg) were generally lower than those previously reported, these levels of intake were consistent with the observation that less than 30% of the sipper presentations earned was accompanied by greater than 5 licks. In a review of the literature, only a small handful of studies (Middaugh et al., 2000; Roberts et al., 2001; Sanchis-Segura et al., 2006; Tsiang & Janak, 2006; Zghoul et al., 2007) were identified that demonstrated operant self-administration of ethanol under conditions of limited access sessions (i.e., 60-min or less) and in food- and water-satiated mice. In earlier work, 20-40 total responses were typically reported during 20-30 min sessions by mice performing on a FR1 schedule of reinforcement, a finding that is comparable to the current finding of 30-35 responses on an FR4 schedule. In some of these earlier reports, ethanol intakes approximating 1.00 g/kg (Zghoul et al., 2007), 0.35-0.55 g/kg (Tsiang & Janak, 2006), and 0.40 g/kg (Roberts et al., 2001) were documented. However, as far as can be gathered from the methods, these estimates were extrapolated from the total number of reinforcer presentations multiplied by the volume of ethanol solution presented in each presentation. Therefore, if the mice in these earlier studies failed to consume the ethanol solution during each presentation (as was the case for the retractable sipper mechanism employed in the current work), then the ethanol dose estimate would be inflated. Consistent with this interpretation, Middaugh and colleagues (2000) observed that during a sucrose fading procedure that began with a 16% sucrose + 12% ethanol solution and yielded a final solution of 12% ethanol, the total number of responses remained steady while the frequency of fountain contacts declined by approximately 85% (to approximately 40 contacts). Collectively, these findings would suggest poor correspondence between responding for and limited access consumption of ethanol by mice under an FR schedule.

The 2- to 3-fold elevation in ethanol intake that occurred during the transition from a FR to a RR schedule in the current study is consistent with an earlier comparison of FR versus RR schedules in rats. Samson and colleagues demonstrated that rats typically consumed 0.4-0.6 g/kg ethanol during a 30-min session when on a FR schedule of reinforcement, whereas an average of 0.8-1.0 g/kg was self-administered when a RR schedule resulted in 20-min of uninterrupted ethanol access (Samson et al., 2000). The procedural separation of the appetitive and consummatory phases of oral ethanol self-administration offers several advantages over the standard FR reinforcement schedule. First, responding under the RR schedule is not influenced by the psychopharmacological (or intoxicating) effects of the drug (Meisch, 2001). Under a FR schedule, ethanol consumption early in the session could interfere with subsequent attempts to attain access to ethanol. Second, a single response requirement permits an initial and unbiased assessment of ethanol’s stimulus strength. This contrasts with a FR schedule during which ethanol is likely to possess a momentary reward value that diminishes throughout the session with repeated fulfillment of response requirements (see Samson et al., 2000). Third, a single response requirement permits the animal to regulate its pattern of ethanol consumption and increases the likelihood that the reinforced behavior is pharmacologically motivated, due to the delay between oral intake and onset of centrally mediated effects (e.g., Meisch, 2001; Samson et al., 2000).

One important procedural consideration regarding the usefulness of animal models to study ethanol reinforcement is that the motivation behind ethanol consumption should be tied to acquisition of a psychopharmacological effect rather than the pursuit of a nutritional (caloric) incentive (Cicero, 1980). Many operant ethanol self-administration procedures use food restriction to 80-85% of free-feeding body weight to get mice to consume ethanol to the point of intoxication (BECs ~100 mg%), which represents a potential confound in the interpretation of the motivational state underlying reinforcement processes. Notably, a 50% decline in basal extracellular dopamine levels in the rat nucleus accumbens was reported following food restriction to 80% of free-feeding body weight (Pothos et al., 1995). Thus, there is evidence that manipulations such as food restriction can alter activity in reward-related neurochemical pathways. In the present study, we avoided this confound by using non-food restricted mice that achieved BECs averaging 37 mg% at the conclusion of a 30-min drinking session. We believe that the significant positive correlation between BEC and ethanol dose consumed favors the interpretation that the motivation underlying ethanol consumption in the present study was a pharmacological one (see Cunningham et al., 2000).

The current study is the first to compare consumption patterns in mice associated with FR and RR schedules as well as to document the application of the “sipper” model in C57BL/6J mice. The cumulative lick records exhibited during the RR8 sessions in this study (Figure 2C) closely resembled the drinking patterns reported by Samson and colleagues (2000) using a rat ‘sipper’ model. Importantly, the current findings with a FR4 schedule indicate that the frequency of sipper presentations did not reliably predict consumption licks, as only 50-55% of the presentations throughout the session were accompanied by contacts with the sipper (Figure 2, panel A). This observation raises the concern that the frequency of ethanol presentations, which are commonly reported in studies employing FR schedules of reinforcement, may not accurately reflect the dose of ethanol consumed (as mentioned above).

In summary, the current study demonstrated a procedural separation of appetitive and consummatory phases of ethanol self-administration and the salience of ethanol as a behavioral reinforcer in non-food restricted mice. Ethanol’s salience with the ‘sipper’ procedure was indicated by a shift in appetitive measures that corresponded with a reduced latency to consumption onset, an elevation in 10E dose consumed, and a robust and rapid extinction response. The employment of this “sipper” model in mice should prove useful in assessing the influence of pharmacological and genetic manipulations on ethanol reinforcement without the need for imposing food restriction.


Supported in part by the Department of Veteran Affairs, the N.L. Tartar Research Fund, and grants AA10760, AA12439, AA13478, AA07468, AA015234, DA14639 and DA07262. The authors would like to acknowledge the contribution of Dr. John C. Crabbe and his laboratory for the BEC determinations.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Bachteler D, Economidou D, Danysz W, Ciccocioppo R, Spanagel R. The effects of acamprosate and neramexane on cue-induced reinstatement of ethanol-seeking behavior in rat. Neuropsychopharmacology. 2005;30:1104–1110. [PubMed]
  • Besheer J, Lepoutre V, Hodge CW. GABAB receptor agonists reduce operant ethanol self-administration and enhance ethanol sedation in C57BL/6J mice. Psychopharmacology. 2004;174:358–366. [PubMed]
  • Campbell UC, Carroll ME. Acquisition of drug self-administration: environmental and pharmacological interventions. Exp Clin Psychopharmacol. 2000;8:312–325. [PubMed]
  • Cicero T. Alcohol self-administration, tolerance and withdrawal in humans and animals: theoretical and methodological issues. In: Rigter H, Crabbe JC, editors. Alcohol Tolerance and Dependence. Elsevier/North Holland Biomedical Press; Amsterdam: 1980. pp. 1–51.
  • Cunningham CL, Fidler TL, Hill KG. Animal models of alcohol′s motivational effects. Alcohol Res Health. 2000;24:85–92. [PubMed]
  • Czachowski CL, Legg BH, Samson HH. Effects of acamprosate on ethanol-seeking and self-administration in the rat. Alcohol Clin Exp Res. 2001;25:344–350. [PubMed]
  • De Vry J, Donselaar I, Van Ree JM. Food deprivation and acquisition of intravenous cocaine self-administration in rats: effect of naltrexone and haloperidol. J Pharmacol Exp Ther. 1989;251:735–740. [PubMed]
  • Forsander OA. Dietary influences on alcohol intake: a review. J Stud Alcohol. 1998;59:26–31. [PubMed]
  • Gallaher EJ, Jones GE, Belknap JK, Crabbe JC. Identification of genetic markers for initial sensitivity and rapid tolerance to ethanol-induced ataxia using quantitative trait locus analysis in BXD recombinant inbred mice. J Pharmacol Exp Ther. 1996;277:604–612. [PubMed]
  • Hodge CW, Miles MF, Sharko AC, Stevenson RA, Hillmann JR, Lepoutre V, Besheer J, Schroeder JP. The mGluR5 antagonist MPEP selectively inhibits the onset and maintenance of ethanol self-administration in C57BL/6J mice. Psychopharmacology. 2006;183:429–438. [PMC free article] [PubMed]
  • Little HJ. Behavioral mechanisms underlying the link between smoking and drinking. Alcohol Res Health. 2000;24:215–224. [PubMed]
  • Meisch RA. Oral drug self-administration: an overview of laboratory animal studies. Alcohol. 2001;24:117–128. [PubMed]
  • Middaugh LD, Kelley BM. Operant ethanol reward in C57BL/6 mice: influence of gender and procedural variables. Alcohol. 1999;17:185–194. [PubMed]
  • Middaugh LD, Kelley BM, Bandy AL, McGroarty KK. Ethanol consumption by C57BL/6 mice: influence of gender and procedural variables. Alcohol. 1999;17:175–183. [PubMed]
  • Middaugh LD, Lee AM, Bandy AL. Ethanol reinforcement in nondeprived mice: effects of abstinence and naltrexone. Alcohol Clin Exp Res. 2000;24:1172–1179. [PubMed]
  • National Research Council of the National Academies . Guide for the Care and Use of Mammals in Neuroscience and Behavioral Research. National Academies Press; Washington, DC.: 2003.
  • Olive MF, Mehmert KK, Messing RO, Hodge CW. Reduced operant ethanol self-administration and in vivo mesolimbic dopamine responses to ethanol in PKCepsilon-deficient mice. Eur J Neurosci. 2000;12:4131–4140. [PubMed]
  • Pothos EN, Creese I, Hoebel BG. Restricted eating with weight loss selectively decreases extracellular dopamine in the nucleus accumbens and alters dopamine response to amphetamine, morphine, and food intake. J Neurosci. 1995;15:6640–6650. [PubMed]
  • Richter CP. A study of the effect of moderate doses of alcohol on the growth and behavior of the rat. J Exp Zoo. 1926;44:397–418.
  • Risinger FO, Brown MM, Doan AM, Oakes RA. Mouse strain differences in oral operant ethanol reinforcement under continuous access conditions. Alcohol Clin Exp Res. 1998;22:677–684. [PubMed]
  • Risinger FO, Doan AM, Vickrey AC. Oral operant ethanol self-administration in 5-HT1b knockout mice. Behav Brain Res. 1999;102:211–215. [PubMed]
  • Roberts AJ, McDonald JS, Heyser CJ, Kieffer BL, Matthes HW, Koob GF, Gold LH. μ-Opioid receptor knockout mice do not self-administer alcohol. J Pharmacol Exp Ther. 2000;293:1002–1008. [PubMed]
  • Samson HH. Initiation of ethanol reinforcement using a sucrose-substitution procedure in food- and water-sated rats. Alcohol Clin Exp Res. 1986;10:436–442. [PubMed]
  • Samson HH, Czachowski CL, Slawecki CJ. A new assessment of the ability of oral ethanol to function as a reinforcing stimulus. Alcohol Clin Exp Res. 2000;24:766–773. [PubMed]
  • Samson HH, Slawecki CJ, Sharpe AL, Chappell A. Appetitive and consummatory behaviors in the control of ethanol consumption: a measure of ethanol seeking behavior. Alcohol Clin Exp Res. 1998;22:1783–1787. [PubMed]
  • Sanchis-Segura C, Borchardt T, Vengeliene V, Zghoul T, Bachteler D, Gass P, Sprengel R, Spanagel R. Involvement of the AMPA receptor GluR-C subunit in alcohol-seeking behavior and relapse. J Neurosci. 2006;26:1231–1238. [PubMed]
  • Tsiang MT, Janak PH. Alcohol seeking in C57BL/6 mice induced by conditioned cues and contexts in the extinction-reinstatement model. Alcohol. 2006;38:81–88. [PubMed]
  • Zghoul T, Abarca C, Sanchis-Segura C, Albrecht U, Schumann G, Spanagel R. Ethanol self-administration and reinstatement of ethanol-seeking behavior in Per1 (Brdm1) mutant mice. Psychopharmacology. 2007;190:13–19. [PubMed]