Individual differences in voluntary alcohol consumption are associated with conditioned fear in the fear incubation model

Abstract

Previous research in male Long Evans rats has shown a relationship between low voluntary alcohol consumption and high conditioned fear after a single training session. Here, we determined whether chronic intermittent access (CIA) to alcohol during adolescence/early adulthood or during adulthood would alter or be associated with auditory-cued conditioned fear levels using an extended training fear incubation procedure. This training procedure leads to low fear soon after training that grows over one month. Rats received 6 weeks of CIA to 20% alcohol or water from PND 26–66. Ten or eleven days later, the rats began behavioral testing that included 10 sessions of tone-shock pairings. Rats then received 4 weeks of CIA exposure during the 1-month fear incubation period and were tested for conditioned fear 6 days after the end of alcohol access. We found no evidence that voluntary alcohol consumption during adolescence/early adulthood or adulthood altered fear expression. However, we found that rats that consumed more alcohol during early adulthood (PND 54-66) had lower fear than low-consumption rats on day 1 of conditioned fear training and in the day 2 and 1-month tests. This extends associations we previously found between individual differences in alcohol consumption and conditioned fear to a different fear conditioning procedure. Combined with our previous data that show that the rate of instrumental extinction is associated with both alcohol consumption and conditioned fear, these data provide further support for the generality and reliability of a pair of phenotypes that encompass a wide variety of learning traits.

Introduction

Alcohol use is related to fear and anxiety in humans [1–3], but the nature of this relationship is unclear. Prospective analyses have provided evidence that previous alcohol abuse can increase the likelihood of developing an anxiety disorder (alcohol abuse preceding fear/anxiety) and pre-existing anxiety disorders can increase the likelihood of later alcohol abuse (fear/anxiety preceding alcohol abuse) [4, 5]. Human research, however, lacks the experimental control to exclude the possibility that some third variable (such as genetics) is responsible for the relationship.

Laboratory fear conditioning in non-human animals, in which a previously neutral cue is paired with footshock and animals acquire a fear reaction to the cue, is often used as a model of fear and anxiety disorders [6, 7]. In laboratory fear conditioning, pre-training alcohol exposure (using alcohol intraperitoneal [i.p.] injections, increased consumption in water-restricted adolescent rats, or an alcohol-containing food source in adult rats) has been shown to alter conditioned fear to discrete tone and light cues, with evidence for increases [8] or decreases [9–11]. However, pre-training alcohol exposure (specifically using voluntary consumption with free food and water, intragastric alcohol injections, or increased consumption in water-restricted adult rats) does not always affect subsequent fear conditioning to these discrete cues [11–14]. Post-training i.p. alcohol injections can also lead to an increase in fear expression if testing occurs 3 days after the final injection [8], although this pattern is not observed with longer alcohol-test intervals of 11 days. Conversely, high levels of alcohol consumption (due to alcohol in the sole food source) after fear conditioning slowed fear extinction, but had no effect on fear expression on the first test trial [9]. As a result, the literature on effects of pre- or post-training alcohol exposure on conditioned fear is mixed.

We have recently shown that male Long-Evans rats given 6 weeks of alcohol access during adolescence and early adulthood (PND26-66) are unaltered (compared to a group given water-only access) in a number of behavioral tasks, including operant devaluation, operant extinction, go/no-go discrimination and reversal learning, and fear conditioning [12, 15]. We also found that, although alcohol access does not alter behavior in these tasks, individual differences in task performance are related to the alcohol consumption level in the last 2 weeks of alcohol access (early adulthood) in the group given alcohol access. In particular, rats that consumed higher levels of alcohol exhibited low conditioned fear, faster instrumental extinction, and fewer errors in reversal learning (with stronger correlations with discontinuation errors [inability to maintain the reversal for consecutive days]) [12, 15]. We hypothesized the existence of two separate phenotypes of rat: a HALF-FIELDER rat (with high alcohol (consumption), low fear, fast instrumental extinction and low discontinuation errors in reversal), and a non-HALF-FIELDER rat (which has none of these traits).

Here, we examined whether this pattern might be observed in a different fear conditioning procedure, the extended training fear incubation procedure. The experiment reported below (and all of the previous alcohol-fear conditioning experiments reviewed above) examines fear rather than anxiety. Fear can be described as the aversive emotion produced when a threat is imminent (likely to occur soon and localizable to source that is present) and anxiety can be described as the emotion produced when a threat is not imminent (either not occurring soon, not localizable to a present source, or uncertain) [16, 17]. As our conditioned fear cues are discrete auditory cues that are presented for a relatively short duration (30-sec) with 100% of the cues leading to footshock, our procedure studies fear. After a single fear conditioning session, conditioned fear is high a few days after training, and remains high weeks or months later [18–21]. However, extended conditioned fear training (10 tone-shock pairings/day for 10 days) leads to low fear 2-3 days after training that increases 1 month later, a phenomenon termed fear incubation [20, 22–25]. Notably, conditioned fear after both a single training session of fear conditioning and the extended training fear incubation procedure exhibits individual differences that have an orderly association with another pair of phenotypes, the sign-tracking and goal-tracking rats (which differ in the form of their response to a cue that predicts food) [22, 26]. With a single session of fear conditioning, sign-tracker rats (who predominately focus responding towards a cue predicting food) have been shown to exhibit higher fear to auditory cues paired with shocks than goal-tracker rats (who predominately ignore a food-predictive cue and approach the food delivery site during cue presentation) [26]. This higher conditioned fear in sign-tracker rats could parallel the higher conditioned fear observed during and after a single fear session in the non-HALF-FIELDER phenotype [12]. In the fear incubation procedure, goal-tracker rats (who approach the food delivery site during a cue predicting food) have not been shown to exhibit increased fear over time, while sign-tracker rats (who focus responding towards a cue predicting food) and intermediate responding rats (who exhibit a mix of responding to the site of food delivery and responding to the food-predictive cue) do exhibit increases in fear over time [22]. Notably, there is also evidence that links within-strain differences in alcohol consumption to sign-tracking/goal-tracking. When Lister rats were given CIA during adulthood and divided into the top 12.5% (high drinkers) and bottom 12.5% (low drinkers), the high drinkers exhibited more sign-tracking behaviors (lever-presses) and fewer goal-tracking responses (head-entries) than the low drinking rats [27]. Thus, there is evidence for higher drinkers exhibiting greater sign-tracking [27] and sign-trackers exhibiting greater fear incubation [22], so it is possible that high drinkers might exhibit greater fear incubation. In the current report, we determined whether fear in the extended training fear incubation procedure would also show associations with alcohol consumption in the HALF-FIELDER/non-HALF-FIELDER phenotypes, with a focus on 2 possible types of association. The first type of association would be represented by either higher or lower fear in the alcohol consumption group (on average, including both HALF-FIELDER and non-HALF-FIELDER rats in the alcohol-access group) than in the water consumption group (on average), suggesting a neurotoxic effect of alcohol exposure that altered the neuronal substrates of fear conditioning. The second type of association would be represented by the same average level of fear in the water group and the alcohol group (as a whole), but the high alcohol drinkers differing from the low alcohol drinkers. This pattern would suggest that there is no observable neurotoxic effect of alcohol on conditioned fear at these relatively low levels of alcohol consumption, but pre-existing individual differences in the propensity to consume alcohol are associated with levels of fear. This second pattern would suggest that the behavioral differences in high vs. low drinkers (the observed HALF-FIELDER and non-HALF-FIELDER phenotypes) extend to fear conditioning with extended fear training in the fear incubation model.

In the current experiment, we determined whether alcohol access during adolescence and early adulthood would alter conditioned fear during the early day 2 test by comparing rats with or without prior alcohol access. In our previous experiments [12, 15], we chose to examine potential effects of adolescent alcohol because adolescence is a common time for the initiation of alcohol drinking in humans [28], and previous literature has demonstrated that there may be more significant long-term effects of alcohol exposure if this exposure occurs during adolescence, rather than adulthood (as reviewed in [29]). The current report is a continuation of this line of research, so we examined alcohol consumption during the exact time period used for alcohol access in these previous experiments (PND26-66) [12, 15]. Our alcohol access method was the chronic intermittent access (CIA) model, in which animals receive access to alcohol for 24-hour periods that alternate with alcohol-free periods, which leads to higher 24-h consumption levels than continuous access [30]. We also tested whether individual differences in alcohol consumption (within the alcohol-access group) would be associated with fear during the day 2 test, which was the test in which sign-tracking/goal-tracking was associated with fear levels.

Because the one-month incubation period represents a post-shock time-point that could simulate a period of alcohol-use to cope with prior stress, we then tested whether alcohol consumption during the 1-month incubation period would affect the level of fear in the 1-month test. The rats remained on food-restriction during this period, so alcohol consumption was higher than in free-feeding rats. This allowed for assessment of the effects of higher levels of alcohol consumption on fear, but prevented a determination of whether motivation for alcohol’s pharmacological properties (rather than hunger) during adulthood was associated with subsequent fear expression.

1. Methods

2.1. Subjects

Male naïve Long Evans rats (n=38) from Charles River Laboratories (Raleigh, N.C.), PND 21 upon arrival in the facility, participated in the experiment. All animals were individually housed and maintained on a 12-hour reverse light-dark cycle with lights off at 07:30 am in a temperature- and humidity-controlled room. The rats were given 5 days to acclimate to the facility, and then received CIA starting on PND 26. Water and food were available ad libitum during this 6-week period. Three days after the final alcohol access period, body weights were recorded and rats were food-restricted and subsequently allowed to grow 1.5 g/day from this initial weight. Water was available ad libitum throughout the experiment, even after the implementation of food-restriction. Training took place during the dark cycle and rats were weighed and fed after the daily sessions. All procedures and animal care were in accordance with the Kansas State University Institutional Animal Care and Use Committee guidelines, the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals, and United States federal law.

2.2. Behavioral apparatus

Experiments were conducted in 12 operant chambers (Med Associates, St Albans, VT). Each chamber had two retractable levers 9 cm above the floor, but only the right lever was extended into the chamber during behavioral sessions. Responding on the lever activated the pellet dispenser, delivering 45-mg precision pellets (#1811155, 5% fat, 66% carbohydrate, 20.3% protein; TestDiet, Richmond, IN). A red houselight was located in the center at the top of the wall opposite to the side with the lever and foodcup. This houselight remained on throughout the duration of all training and testing sessions. A tone generator that delivered a 2900 Hz tone (20 dB above background) was located directly to the right of the houselight. The chambers had grid floors connected to electric shock generators that delivered a 0.5-mA scrambled foot-shock.

2.3. Behavioral procedures

A diagram of the experimental timeline is in Figure 1. This figure contains behavioral training and testing periods in non-italicized font and alcohol/water access periods in italics.

Figure 1.

Experimental timeline. Ages from the beginning of lever-press training until the start of the incubation period represent ranges with one cohort starting and ending the phase one day soon than the other (ex: lever-press training occurred from PND 76-81 in one cohort and PND 77-82 in the other cohort). All rats had the final day of the incubation period on PND 127.

Rats received 6 weeks of alcohol (using CIA) or water-only access beginning in adolescence and extending into early adulthood (PND 26-66). A two-bottle choice procedure was used in which all rats had access to water in at least one bottle at all times. The Alcohol group (n=20) received 24-h access to 20% (v/v) ethanol mixed with tap water 3× per week in one bottle on Sunday, Tuesday, and Thursday, while the other bottle contained tap water. The other days were water-only days in which both bottles contained water. The Water group (n=18) received two water bottles during the six weeks, with no access to alcohol. Bottles were weighed and changed (for alcohol groups) starting at 1 pm and ending by 2:30 pm every day except Saturday and placement was counterbalanced to control for any side preference. As all rats had water in both bottles from Friday to Sunday, the two water bottles were removed on Saturday to weigh the rats, but the bottles were then replaced in the same positions without being weighed.

Behavioral training started 10-11 days after the end of alcohol access and was largely the same as in previous studies [20, 23]. Each rat was run in the same operant chamber throughout the experiment. Rats were given a 60-min food-cup training session, with pellet deliveries every 125 sec. The following day, the rats received 2 sessions on a fixed-interval-1 (FI-1) reinforcement schedule for lever-presses on the lever (lever-presses could earn a pellet each sec) 2–4 h apart. These sessions ended when rats received 50 pellets (with a limit of 1 h). Starting the following day, the rats were given one 90-min session in which pellets were earned on a variable-interval-30 (VI-30) reinforcement schedule (pellet availability for lever-presses ranging from 1 to 59 sec), and then 2 daily 90-min sessions on a VI-60 schedule (pellet availability ranging from 1 to 119 sec). Rats were maintained on the VI-60 schedule for the remainder of the behavioral training.

The rats then received 10 daily sessions of fear conditioning. During the fear conditioning sessions, animals earned pellets on a VI-60 schedule. Each session began with the extension of the right lever and illumination of a red houselight. The rats received 10 30-second tones (2900 Hz, 20 dB above background) ranging from 3 to 14 min apart, co-terminating with 0.5-mA, 0.5-sec foot-shocks (adjusted for inter-chamber variability) pseudo-randomly throughout each 90-min session.

Next, the rats received a lever-press reacquisition session followed by a cued fear test on days 1 and 2 after the end of fear conditioning, respectively. The lever-press reacquisition session consisted of a 90-min VI-60 lever-press training during which no tones or shocks were presented. One main function of the lever-press reacquisition session was to extinguish any contextual fear to the training/testing context, so that lever-pressing would be robust in the precue period of the cued fear test on the following day. Although baseline lever-pressing is a crude measure of fear conditioning (due to individual differences in rats’ lever-press rates even in the absence of fear conditioning), assessment of changes in lever-press rates during this lever-reacquisition session can provide a rough measure of contextual fear and extinction of this fear. The cued fear test lasted 35 min, during which rats lever-pressed on a VI-60 schedule and 4 presentations of the tone fear cue were presented without any shock deliveries.

The rats then received 4 additional weeks of CIA to alcohol during the 1-month fear retention/incubation interval. This was identical to the adolescent alcohol access procedure, except for two differences. First, the rats remained food restricted throughout the 1-month period of access. Second, half of the Adolescent Alcohol group received alcohol access during this fear retention/incubation interval and the other half received water (n=10 Adolescent Alcohol-Adult Alcohol; 10 = Adolescent Alcohol-Adult Water; groups balanced so adolescent/early adult drinking was equivalent). Likewise, half of the Adolescent Water group received alcohol during the retention/incubation interval and the other half received water (n= 9 Adolescent Water-Adult Alcohol; 9 = Adolescent Water-Adult Water).

Finally, there was a lever-press reacquisition session followed by a cued fear test five and six days after the final alcohol access period (35 and 36 days after the final day of fear conditioning), respectively. The lever-press reacquisition session and cued fear tests were identical to the training/tests on days 1 and 2 after the end of fear conditioning. The day 36 test is referred to as the 1-month test throughout the text.

2.4. Behavioral measures

For alcohol consumption, the alcohol consumed was calculated as the bottle weight difference between the start and end of the access periods, minus 2 g for spillage/evaporation, and then multiplied by 0.162 for the weight of alcohol in 1 g of a 20% v/v alcohol solution. The 2 g subtracted for spillage was determined based on pilot research conducted in our laboratory during which bottles were placed on empty cages and weighed daily to determine general spillage/evaporation.

We used conditioned suppression of lever-pressing as our fear measure [31]. Procedures in which conditioned freezing and conditioned suppression are concurrently measured (including experiments using the same fear conditioning parameters as in the current experiment) have shown that higher conditioned suppression is associated with higher conditioned freezing [12, 25, 32, 33], Lever-presses were recorded during the 30 sec prior to tone presentation (Precue) and during the 30-sec tone presentation (Cue), and were used to calculate a suppression ratio using the formula: Suppression ratio = ((Precue−Cue)/(Precue+Cue)). The suppression ratio normalizes lever-pressing during the tone based on baseline responding [34, 35], A value of 1 indicates total conditioned suppression of lever-pressing during tone presentation (high fear), while a value of 0 reflects no lever-press suppression during tone presentation (low fear). When both precue and cue values were 0 on a trial or trial block, the rats’ responding in the trials immediately before and after that trial or trial block were averaged to determine the suppression ratio value.

2.5. Statistical analyses

Data were analyzed by Statistica 5.1 software (Tulsa, OK). The factors used in the statistical analyses are described in the Results section for each ANOVA and significant effects (p<0.05) in the different ANOVAs were followed by post-hoc Tukey’s HSD tests.

2. Results

3.1. Adolescent/early adult alcohol consumption

If all Alcohol group rats were analyzed together, there was relatively steady alcohol consumption across the 6-week access period. However, our goal was to examine individual differences in alcohol consumption. For individual differences analyses, we categorized the Alcohol group based on their consumption in the last 2 weeks of alcohol access, to be consistent with our previous work [12, 15], We did not divide the rats based on individual differences in alcohol consumption during the incubation interval, as the rats were food-restricted during this period and consumption would not represent a measure of alcohol drinking motivated by the pharmacological properties of alcohol alone. In our previous alcohol-fear conditioning experiment, we divided the rats into those that consumed more than or less than 1.5 g/kg/24-h, termed HALF-FIELDER and non-HALF-FIELDER rats, respectively, which aligned with cut-offs of 1.5—1.6 g/kg/24-h used to divide rats into low drinkers in other studies and to select for drinking levels in-bred rat lines [36, 37], In the current experiment, we found that there were 6 rats with consumption below the 1.5 g/kg/24-h level. However, in the current sample, there were no rats with consumption between 2 and 2.5 g/kg/24-h and there were 3 rats with consumption between 1.5 and 2 g/kg/24-h (Figure 2A). In light of this gap in the range of drinking in our sample, it seemed that grouping this “intermediate group” of 3 rats as high drinkers made less sense than categorizing them as part of the low drinking group or excluding them from the analysis entirely so that there would be a separation in the cut-offs for the high and low drinking groups. We proceeded to analyze the individual differences in fear conditioning behavioral data in two different ways. First, in our primary analyses, these 3 rats were categorized as “intermediate drinkers” and excluded from the individual differences analysis (with 6 low drinkers and 11 high drinkers in the analysis), as is often done in other research using an “extreme-groups approach” [27, 38, 39], Second, we ensured that our results were not dependent on excluding these 3 rats by performing secondary analyses in which these 3 rats were included in the low drinking group (with 9 low drinkers and 11 high drinkers).

Figure 2.

A. Rank ordered alcohol consumption (in g/kg/24-h) in the last 2 weeks of the adolescent/early adult access period (corresponding to PND54-66). B. Weekly alcohol consumption (mean ± SEM) across the 6 weeks of pre-training access during adolescence and early adulthood. * = p<0.05 significant difference in Alcohol consumption in the high vs. intermediate and low alcohol drinking groups (a main effect across the 6 weeks). In A and B, white circles/bars indicate low alcohol drinkers (<1.5 g/kg/24-h) gray circles/bars indicate intermediate alcohol drinkers (between 1.5 and 2 g/kg/24-h), and black circles/bars indicate high alcohol drinkers (>2.5 g/kg/24-h). C. Weekly alcohol consumption (mean ± SEM) across the 4 weeks of post-training alcohol access during the incubation interval. Adol Water-Adult EtOH = the group that received water-only access during adolescence/early adulthood and alcohol access during the incubation interval. Adol EtOH- Adult EtOH = the group that received alcohol access during adolescence/early adulthood and during the incubation interval. D. Correlation between alcohol consumption in the last 2 weeks of the pre-training adolescent/early adult access phases and the last 2 weeks of the post-training incubation interval in Adolescent Alcohol-Adult Alcohol group, which had access to alcohol in both period.

We analyzed the drinking data in the 3 groups (Figure 2B) with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High, Intermediate and Low) and the within-subjects factor of Week (the 6 weeks of consumption). We found a significant effect of Alcohol Consumption Group (F(2, 17)=9.1, p<0.01), but no significant effect of Week and no significant Alcohol Consumption Group × Week interaction (all p>0.05). Post-hoc test found that the high drinking group differed from the low drinking group (p<0.01) and the intermediate drinking group (p<0.05), but the low and intermediate drinking groups did not differ (p=0.97).

Because half of the Adolescent Alcohol group would be later assigned to have water during the incubation interval and half would receive alcohol during the incubation interval, we performed an analysis to ensure that these 2 groups had equivalent alcohol consumption during the adolescent/early adult period. We analyzed the drinking data with a mixed-factor ANOVA with the between-subjects factor of Adult Alcohol Group (whether the rats were assigned to have alcohol or water-only access during the incubation interval) and the within-subjects factor of Week (the 6 weeks of adolescent/early adult consumption). We found no significant effects or interactions (all p>0.05). Thus, there was no significant differences in adolescent/early adult drinking between the adolescent/early adult alcohol rats that were assigned to receive alcohol or water during the incubation period.

Finally, we determined whether alcohol access or drinking behavior during adolescence/early adulthood affected the developmental increase in body weight during this period in order to determine whether alcohol access had adverse health consequences that impaired growth. We first compared body weights in the two groups receiving alcohol or water only access. We analyzed these data with a mixed-factor ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and the within-subjects factor of Access Day (the 18 days in which rats in the Alcohol group received alcohol access). We found an effect of Access Day (F(17, 612)=2026.7, p<0.01), but no effects or interactions of Adolescent Access Group (all F<1). We then compared body weights in the three Alcohol subgroups. We analyzed these data with a mixed-factor ANOVA with the between-subjects factors of Alcohol Consumption Group (High, Intermediate and Low) and the within-subjects factor of Access Day. We found an effect of Access Day (F(17, 289)=504.7, p<0.01), but no effects or interactions of Alcohol Consumption Group (all F<1). These results suggest that alcohol access had no obvious adverse effects on the rats’ health or developmental growth.

3.2. Alcohol consumption in the fear incubation interval in food-restricted adult rats

There was relatively steady alcohol consumption across the 4 weeks of the fear incubation interval, and there were no differences in alcohol consumption during the incubation interval between the rats that had previous adolescent/early adult alcohol or water-only access (Figure 2C). We analyzed the drinking data with a mixed-factor A NOVA with the between-subjects factor of Adolescent Alcohol Group (whether the rats had previous alcohol or water-only access during adolescence/early adulthood) and the within-subjects factor of Week (the 4 weeks of consumption). We found no significant effects or interactions (all p>0.05). To ensure that there were no differences in consumption in the last 2 weeks of access during the incubation interval, we also ran a one-way ANOVA examining the between-subjects effect of Adolescent Alcohol Group on the average consumption in the last 2 weeks and found no significant effects (p>0.05). To determine whether food-restriction during the incubation interval increased consumption, we compared the consumption of the group given alcohol during both adolescence/early adulthood and during the fear incubation interval (the Adolescent Alcohol-Adult Alcohol group; n=10). An ANOVA comparing alcohol consumption in the last two weeks of the adolescent/early adult period (corresponding to early adulthood on PND 54-66; mean = 3.8±1.1 g/kg/24-h) with consumption in the last 2 weeks of the incubation interval (mean = 8.5±1.0 g/kg/24-h) found a significant effect (F(1, 9)=17.9, p<0.01). There was no significant correlation between alcohol consumption in the two periods (r=0.39, p>0.05) (Figure 2D), suggesting that food-restriction interfered with the observation of stable individual differences in alcohol consumption.

Finally, we determined whether alcohol access or drinking behavior during adolescence/early adulthood or alcohol access during the fear incubation interval during adulthood were associated with different body weights in the different groups. Once behavioral training began, all rats were food restricted and fed to maintain them at target weights that grew by 1.5 g/day. Comparison of weights on the first day of adult alcohol access during the incubation interval (or any other day once the rats stabilized on food restriction) would provide information about any differences between the groups, as all rats were fed to maintain a steady growth from this weight without variation in the growth curves. We first compared body weights based on whether the rats received alcohol or water-only access during adolescence/early adulthood and during the incubation period during adulthood. We analyzed the body weights on the first day of alcohol access during adulthood with a between-subjects ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and Adult Alcohol Group (Water vs. Alcohol). We found no effects or interactions of either factor (all F<1). We then limited our analysis to rats that received alcohol access during adolescence/early adulthood and analyzed the body weights on the first day of alcohol access during adulthood based on the different subgroups of rats that differed in their adolescent/early adult alcohol consumption (High, Intermediate and Low groups). We performed this analysis to ensure that there were no long-term effects of alcohol consumption level on body weight that were then maintained by our food restriction regimen. We analyzed these data with a between-subjects ANOVA with the between-subjects factor of Alcohol Consumption Group (High, Intermediate and Low). We found no effect (p>0.05).

2.3. Lever-press training prior to fear conditioning

There was no effect of prior alcohol consumption during adolescence/early adulthood (alcohol vs. water access or high-drinking vs. low-drinking alcohol consumption) on lever-press rates prior to the start of fear conditioning. We first compared the rats’ lever-press rates in the 2 VI-60 lever-press sessions prior to fear conditioning based upon whether the rats had access to alcohol or water only in adolescence/early adulthood. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and the within-subjects factor of Training Session (the 2 lever-press training days). We found an effect of Training Session (F(1, 36)=5.3, p<0.05), but no effects or interactions of Adolescent Access Group (all p>0.05; data not shown). We then analyzed the lever-press rates in the alcohol consumption sub-groups, with a comparison of the high (>2.5 g/kg/24-h) and low (<1.5 or 2 g/kg/24-h) alcohol drinkers. We performed our primary analyses comparing the Low Alcohol Group and the High Alcohol Group, with the intermediate drinkers excluded. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Training Session. We found no significant effects or interactions (all p>0.05; data not shown). If we included the intermediate rats in the Low Alcohol group, we found the same pattern. An ANOVA examining the lever-press rates found no significant effects or interactions (all p>0.05; data not shown).

3.4. Conditioned fear training

There was no effect of prior alcohol consumption during adolescence/early adulthood (or of future assignment of water/alcohol group during the incubation interval) on fear during the conditioning sessions. However, the sub-groups of the Alcohol Consumption Group that were divided based on their alcohol consumption behavior during adolescence/early adulthood (Low or High consumption) significantly differed in their conditioned fear behaviors on the first conditioning day (p<0.05). Across the entire 10-day training regimen, the comparison in conditioned fear behavior between the High and Low Alcohol Consumption group trended towards significance (0.10>p>0.05), suggesting that individual differences in alcohol consumption are associated with individual differences in conditioned fear during training.

We first compared the rats based upon whether they had access to alcohol or water only in adolescence/early adulthood or the incubation interval (Figure 3A, left). We analyzed the conditioned suppression behavioral data with a mixed-factor ANOVA (as the ANOVA mixed a between-subjects variable with a repeated-measure) with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and Adult Access Group (Water vs. Alcohol) and the within-subjects factor of Training Session (the 10 days of conditioning). We found an effect of Training Session (F(9, 306)=5.1, p<0.01), but no effects or interactions of Adolescent Access Group or Adult Access Group (all F<1). Because we previously found that conditioned fear behaviors in a single session of fear conditioning were associated with the level of drinking in an alcohol access group [12], we also examined whether there was any effect or interaction of Adolescent Access Group or Adult Access Group on conditioned suppression on the first session of fear conditioning. An ANOVA examining fear responding only on the first day of conditioning also found no effects or interactions of Adolescent Access Group or Adult Access Group (all F<1). An analysis of precue lever-pressing across the 10 conditioning sessions found an effect of Training Session (F(9, 306)=47.9, p<0.01), but no effects or interactions of Adolescent Access Group, Adult Access Group, or Test Interval (all p>0.05).

Figure 3:

A. Average suppression ratios across 10 training fear conditioning sessions. High values indicate high fear and low values indicate low fear. * = p<0.05 difference in fear conditioning in the high vs. low alcohol drinking group when the first fear conditioning session was analyzed. B. Average lever-press rates in 5-min blocks during the 90-min Day 1 lever-reacquisition session. * = p<0.05 significant difference in lever-press rates in the high vs. low alcohol drinking groups in individual blocks. C. Average suppression ratios during the test session. * = p<0.05 significant difference in conditioned fear in the high vs. low alcohol drinking groups (a main effect across the two test sessions). # = p <0.05 significant difference between suppression ratios on day 2 vs. 1 month (a main effect across the two groups- demonstrating the incubation effect). For A-C, alcohol (gray) and water (white) groups are shown on the left and the high (black) and low (white) subgroups of the alcohol group are shown on the right. In the figure legends, Adol = Adolescent. D. Correlation between alcohol consumption in the last 2 weeks of access and conditioned suppression of lever-pressing during the first fear conditioning session. E. Correlation between alcohol consumption in the last 2 weeks of access and average conditioned suppression of lever-pressing across the 10 days of fear conditioning. F. Correlation between alcohol consumption in the last 2 weeks of access and average conditioned suppression of lever-pressing during the day 2 cued fear test. G. Correlation between alcohol consumption in the last 2 weeks of access and average conditioned suppression of lever-pressing during the 1-month cued fear test. For D-F, left panels represent correlations with the last 2 weeks of adolescent/early adult consumption and right panels represent correlations with the last 2 weeks of alcohol during the incubation period in adulthood. For D-G, values in bold with an asterisk represent significant effects.

Despite the lack of a difference between the Alcohol and Water groups, we found a difference in fear behavior between the high (>2.5 g/kg/24-h) and low (<1.5 or 2 g/kg/24-h) alcohol drinkers that trended towards significance when all 10 days of training were analyzed, and a significant difference on the first day of conditioning (Figure 3A, right). We performed our primary analyses comparing the Low Alcohol Group (n=6; consumption < 1.5 g/kg/24) and the High Alcohol Group (n=11; consumption > 2.5 g/kg/24), with the intermediate drinkers (n=3; consumption between 1.5 and 2 g/kg/24-h) excluded. We analyzed the conditioned suppression data with a mixed-factor ANOVA (as the ANOVA mixed a between-subjects variable with a repeated-measure) with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Training Session (the 10 days of conditioning). We found a significant effect of Training Session (F(9, 135)=3.8, p<0.01) and an effect of Alcohol Consumption Group that trended towards significance (F(1, 15)=3.4, p=0.09), but no interaction (F<1). However, ANOVAs do not maintain the temporal order of the training days, so we conducted a mixed-effects model (using JMP Pro 13, Cary, N.C.) of the effect of Alcohol Consumption Group (High vs. Low) and Training Day (1 through 10, treated as a continuous variable) on suppression ratios during training, while including random effects that allow for intercept and Training Day slope to vary. This analysis was unplanned and was treated as exploratory. In this analysis, no effects or interactions were significant (all p>0.05). This finding suggested that the lack of an effect or interaction of Alcohol Consumption group was not due to a lack of power in the ANOVA performed. The lack of a difference between high and low drinking rats across the 10 days of fear conditioning was surprising, as we previously found that high and low drinking rats exhibited different levels of fear during the single session of fear conditioning in rats only given 1-day training. Next, we tested whether differences in fear would be observed between the high and low drinking rats if we limited our analysis to the first day of fear conditioning (equivalent to the entire training length in our previous experiment [12]). If only the first day of fear conditioning was analyzed with a one-way ANOVA with Alcohol Consumption Group (High vs. Low) as the factor, we found a significant effect (F(1, 15)=4.6, p<0.05).

The pattern of results for the high vs. low drinkers was not dependent on excluding the intermediate drinking rats from the analysis. If we included the intermediate rats in the Low Alcohol group (as their drinking was more similar to this group and their drinking was separated from the High Alcohol group, with no rats with drinking from 2-2.5 g/kg/24-h), we found the same pattern. An ANOVA examining conditioned fear across all 10 training days found a significant effect of Training Session (F(9, 162)=2.9, p<0.01) and an effect of Alcohol Consumption Group that trended towards significance (F(1, 18)=3.2, p=0.09), but no interaction (F<1). If only the first day of fear conditioning was analyzed with a one-way ANOVA, we found a significant effect of Alcohol Consumption Group (F(1, 18)=8.5, p<0.01). Analyses of precue lever-pressing across the 10 conditioning sessions found an effect of Training Session regardless of whether the intermediate rats were excluded (F(9, 135)=23.6, p<0.01) or included (F(9, 162)=27.6, p<0.01), but no effects or interactions of Alcohol Consumption Group (all p>0.05).

2.4. Lever-press rates during the day 1 and 1-month lever-reacquisition sessions

We found no effect of adolescent alcohol access (compared with water-only access) on lever-press responding during the Day 1 and no effect alcohol access during adolescence/early adulthood or during the adult incubation period on lever-pressing responding in the 1-month lever-reacquisition session. Within the Alcohol Consumption group, the High drinkers exhibited a faster increase in lever-pressing during the early part of the Day 1 reacquisition session (possibly reflecting extinction of contextual fear to the fear-associated chamber and the loss of fear-induced suppression of lever-pressing), but no difference in lever-pressing during the Day 30 session.

We first compared the rats’ lever-press rates in the Day 1 lever-reacquisition session prior to the Day 2 cued fear test based upon whether the rats had access to alcohol or water only in adolescence/early adulthood. In our initial ANOVA, we analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and Adult Alcohol Group (Water vs. Alcohol) and the within-subjects factor of Session Block (the 18 5-min blocks within the 90-min lever-reacquisition session). We found an effect of Session Block (F(17, 578) = 9.3, p<0.01) and a significant Adolescent Alcohol Group × Adult Alcohol Group × Session Block interaction (F(17, 578) = 2.4, p<0.01), but no other effects or interactions were significant (all p>0.05; Figure 3B, left). As there were no effects or interactions of Adolescent Alcohol Group apart from this 3-way interaction, this reflects inefficiency in balancing the groups for future adult alcohol access during the incubation interval, rather than an effect of adolescent/early adult alcohol access itself. We balanced the Adult alcohol vs. water access groups (for treatment during the incubation interval) to equate for fear training and day 2 testing suppression ratios, as well as consumption in the rats given adolescent alcohol, but we did not balance the groups based on their lever-pressing during the day-1 lever reacquisition session. To ensure that this did not reflect an effect of alcohol access during adolescence/early adulthood, we re-ran the analysis without the Adult Alcohol Group factor. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and the within-subjects factor of Session Block (the 18 5-min blocks within the 90-min lever-reacquisition session). We found an effect of Session Block (F(17, 612)=9.0, p<0.01), but no effects or interactions of Adolescent Access Group (all p>0.05).

We then analyzed the lever-press rates in the Day 1 lever reacquisition session in the alcohol consumption rats, with a comparison of the high (>2.5 g/kg/24-h) and low (<1.5 or 2 g/kg/24-h) alcohol drinkers. We performed our primary analyses comparing the Low Alcohol Group and the High Alcohol Group, with the intermediate drinkers excluded. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Session Block. We found an effect of Session Block (F(17, 255)=2.6, p<0.01) and an interaction of Alcohol Consumption Group × Session Block that trended towards significance (F(17, 255)=1.7, p=0.053), but no main effect of Adolescent Access Group (F<1; data not shown). If we included the intermediate rats in the Low Alcohol group, we found a similar pattern, but the Alcohol Consumption Group × Session Block interaction reached significance. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Session Block. We found an effect of Session Block (F(17, 306)=3.6, p<0.01) and an interaction of Alcohol Consumption Group × Session Block (F(17, 306)=2.1, p<0.01), but no main effect of Adolescent Access Group (p>0.05; Figure 3B, right). Post-hoc analyses showed that there was no significant difference in the initial rate of lever-pressing in the first block and that terminal levels of lever-pressing in the final blocks of the lever-reacquisition did not differ. Instead, significant differences emerged between trial blocks 3 and 8, suggesting different responding between 10 and 40 minutes after the start of the session.

We then compared the rats’ lever-press rates in the 1-Month lever-reacquisition session prior to the 1-Month cued fear test based upon whether the rats had access to alcohol or water only in adolescence/early adulthood or during the incubation interval during adulthood. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and Adult Alcohol Group (Water vs. Alcohol) and the within-subjects factor of Session Block (the 18 5-min blocks within the 90-min lever-reacquisition session). We found an effect of Session Block (F(17, 578)=1.8, p<0.05), but no effects or interactions of Adolescent Access Group (all p>0.05; data not shown). We then analyzed the lever-press rates in the 1-Month lever reacquisition session in the rats that received alcohol consumption in adolescence/early adulthood, with a comparison of the high (>2.5 g/kg/24-h) and low (<1.5 or 2 g/kg/24-h) alcohol drinkers. We performed our primary analyses comparing the Low Alcohol Group and the High Alcohol Group, with the intermediate drinkers excluded. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Session Block. We found no significant effects or interactions (all p>0.05; data not shown). If we included the intermediate rats in the Low Alcohol group, we found a similar pattern, but the main effect of Session Block reached significance. We analyzed the lever-press data with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Session Block. We found an effect of Session Block (F(17, 306)=2.2, p<0.01), but no main effect or interaction of Alcohol Consumption Group (all p>0.05).

3.6. Conditioned fear testing

There was no effect of prior alcohol consumption during adolescence/early adulthood or in food-restricted adult rats on conditioned fear in either fear test. However, the sub-groups of the Alcohol group that were divided based on their alcohol consumption behavior differed in their conditioned fear behaviors during testing, suggesting that individual differences in alcohol consumption are associated with individual differences in conditioned fear in the fear incubation procedure.

We first compared the rats based upon whether they had access to alcohol or water only in adolescence/early adulthood or the incubation interval (Figure 3C, left). We analyzed all data with a mixed-factor ANOVA with the between-subjects factors of Adolescent Access Group (Water vs. Alcohol) and Adult Access Group (Water vs. Alcohol) and the within-subjects factor of Test Interval (Day 2 vs. 1 Month). We found an effect of Test Interval (F(1, 34)=24.5, p<0.01), demonstrating our fear incubation effect. There were no significant effects or interactions of Adolescent Access Group or Adult Access group (all p>0.05). An ANOVA analyzing the effect of Adolescent Access Group (Water vs. Alcohol) on day 2 fear (without including Adult Access Group) also found no effect of Adolescent Access Group (F<1). This demonstrates that there was not an effect of adolescent alcohol prior to alcohol/water exposure in the incubation interval that was obscured by the fear incubation effect or by including the future incubation interval alcohol/water access group assignment in the analysis. Analyses of precue lever-pressing also found no effects or interactions of Adolescent Access Group, Adult Access Group, or Test Interval if both tests were included in the ANOVA (all p>0.05) or of Adolescent Access Group on lever-pressing if test day 2 was analyzed alone (F<1).

Despite the lack of a difference between the Alcohol and Water groups, we did find differences between the high (>2.5 g/kg/24-h) and low (<1.5 or 2 g/kg/24-h) alcohol drinkers (Figure 3C, right). We performed our primary analyses comparing the Low Alcohol Group and the High Alcohol Group, with the intermediate drinkers excluded. We analyzed the conditioned suppression data with a mixed-factor ANOVA with the between-subjects factor of Alcohol Consumption Group (High vs. Low) and the within-subjects factor of Test Interval (Day 2 vs. 1 Month). We found significant effects of Alcohol Consumption Group (F( 1, 15)=4.6, p<0.05) and Test Interval (F(1, 15)=5.7, p<0.05), but no interaction (F<1). An ANOVA analyzing the effect of Alcohol Consumption Group (High vs. Low) on day 2 fear also found a significant effect (F(1, 15)=5.7, p<0.05), demonstrating that the individual differences in fear behavior were present before the incubation interval and adult alcohol consumption and were not driven by any psychological or neurobiological changes that occurred during the retention interval. Notably, this pattern was not dependent on our exclusion of the intermediate drinking rats from the analysis. If we included the intermediate rats in the Low Alcohol group, we found the same pattern. An ANOVA examining the day 2 and 1-month tests found significant effects of Alcohol Consumption Group (F(1, 18)=7.6, p<0.05) and Test Interval (F(1, 18)=6.3, p<0.05), but no interaction (F<1). An ANOVA analyzing the effect of Alcohol Consumption Group (High vs. Low) on day 2 fear also found a significant effect (F(1, 18)=10.0, p<0.01). Regardless of whether the intermediate rats were excluded from the analysis or included in the Low Alcohol group, analyses of precue lever-pressing also found no effects or interactions of Alcohol Consumption Group or Test Interval if both tests were included in the ANOVA (all p>0.05) or of Alcohol Consumption Group on lever-pressing if test day 2 was analyzed alone (all F<1).

3.7. Correlations between alcohol consumption and fear assessments

Despite the significant differences in conditioned fear between high and low drinking groups divided based on their consumption in early adulthood (weeks 5 and 6 of the adolescent/early adult assessment), the majority of these relationships were not accompanied by significant correlations if alcohol consumption was directly correlated to the fear measures. We calculated Pearson’s correlation coefficients for eight separate correlations (2 alcohol age periods × 4 fear assessments). The alcohol measures were the average of the last two weeks of alcohol consumption 1) during adolescence/early adulthood and 2) during the incubation period in adulthood. The fear assessment were 1) average conditioned suppression on the first day of conditioned fear training, 2) average conditioned suppression across the 10 days of conditioned fear training, 3) average conditioned suppression in the early cued fear test (day 2), and 4) average conditioned suppression in the late cued fear test (1 month). When we performed these correlations (Figure 3D-3G), the only significant correlation was the relationship of alcohol consumption in weeks 5-6 of adolescent/early adult consumption with conditioned suppression in the first fear conditioning session (r = −0.54, p<0.05). While the other 7 correlations trended in the negative direction, none of them reached significance (correlation coefficients can be found on the figures). The pattern observed for the relationship of alcohol consumption in weeks 5-6 of adolescent/early adult consumption with conditioned suppression in the day 2 and 1-month tests (with significant differences between high and low drinkers when rats are divided into groups, but the correlations missing significance) resembles the pattern observed in our previous paper on the HALF-FIELDER/non-HALF-FIELDER phenotypes [12]. The high alcohol consumption group contains almost exclusively rats with low fear, while the low alcohol consumption group contains some rats with low fear intermixed with some rats with high fear, which resembles the pattern we previously found.

3. Discussion

We found that alcohol consumption had no effect on fear incubation, regardless of whether the alcohol consumption occurred during adolescence/early adulthood or during the post-training incubation interval (even with higher consumption during the incubation interval due to food restriction). Instead, we found that individual differences in alcohol consumption within the alcohol access group were associated with conditioned fear levels. Very low alcohol consumption (whether defined as <1.5 or <2 g/kg/24-h) was associated with high conditioned fear during the first conditioned fear training day and in both the day 2 and 1-month tests. Below, we discuss the individual differences effect and the lack of a group effect of alcohol access.

4.1. Alcohol access had no effect on conditioned fear

Alcohol access during adolescence/early adulthood (PND26-66) had no effect on fear conditioning during the ten fear conditioning sessions or in the day 2 or 1-month tests. Likewise, alcohol consumption during the 1-month incubation interval had no effect on conditioned fear in the 1-month test, even though this consumption was increased compared to that in adolescence/early adulthood (likely due to food restriction). We have already reviewed reasons for a lack of an effect of adolescence/early adult alcohol consumption on fear conditioning in a previous paper [12]. To summarize, it is likely that the dose/blood ethanol concentrations (BECs) reached were insufficient to alter conditioned fear and/or the interval between the end of alcohol access and the start of fear conditioning was too long (see [12] for more details).

The lack of an effect of post-training alcohol access on the day 30 test requires more explanation. The level of alcohol consumption during the incubation interval was higher and the interval between the end of alcohol access and the fear test was shorter than the comparable values for adolescence/early adulthood (day 2 test occurring 28-29 days after the end of adolescent/early adult alcohol vs. 1-month test occurring 6 days post-alcohol for post-training). For this reason, alcohol exposure during the incubation interval might be expected to have a stronger effect. An examination of the trial-by-trial data in our 1-month fear test, however, found no effect of alcohol exposure on the trial-by-trial extinction (no main effects or interactions of Adolescent Alcohol Group or Adult Alcohol Group, all F<1), so post-training alcohol had no effect on either overall fear or the rate of fear extinction in our test.

Although it is possible that the delay between post-training alcohol and the 1-month test is responsible for our lack of an effect, we believe that a more likely explanation is that the level of alcohol exposure was insufficient to lead to a long-term change in conditioned fear. Five once-daily injections of 1.5 g/kg ethanol (i.p.) given 1-5 days after auditory fear conditioning in rats led to increased fear expression in a test 3 days after the last injection, although this effect was not seen in a test 11 days after the last injection [8]. This suggests that post-fear conditioning alcohol exposure can increase expression of previously learned fear, but it can be a relatively short-lived effect. However, another experiment found effects at a longer post-alcohol interval. After fear conditioning, Ripley and colleagues gave rats an alcoholic diet as the exclusive food source for 28–30 days, leading to average consumption of ~12.5-13 g/kg/24-h [9]. They found slower extinction, but no effect on fear expression on the first trial, in a test 12 days after the end of fear expression. As the alcohol diet slowed fear extinction at a longer alcohol-test interval (12 days) than the one we used (6 days) [9], it is unlikely the alcohol-test interval alone can account for our lack of effect.

An alternative possibility for the lack of an effect of alcohol in our fear test is that the level of alcohol exposure/BECs in our study were not high enough to lead to alterations in fear if given post-training. The alcohol injections that led to increased fear expression 3 days later were intended to cause peak BECs of ~200 mg/dl [8, 40] and the liquid diet that led to slower fear extinction 12 days later caused average consumption of ~12.5-13 g/kg/24-h [9]. These represent consumption levels and likely peak BECs higher than those in our current experiment. In previous work from our laboratory, food-restricted adult rats given CIA consumed an average of 10.5-11 g/kg/24-h and reached blood alcohol levels of 85-90 mg/dl [41]. During the incubation interval in the current experiment, the rats consumed an average of 6.5-8.5 g/kg/24-h in weeks 3 and 4 (depending on group), suggesting the BECs in the current study were lower than 85-90 mg/dl. Since the rats in the current study likely had much lower BECs or consumption levels than studies that found increased fear expression (~200 mg/dl BEC [8]) or slower extinction (~12.5-13 g/kg/24-h consumption [9]), it is possible that our level of alcohol exposure was not high enough to cause alterations in fear if given after training. Additional research will need to be conducted to determine the parameters under which post-training alcohol exposure affects fear conditioning.

4.2. Individual differences in alcohol access are related to conditioned fear behavior

We replicated our previous finding that conditioned fear on the first day of fear conditioning was associated with the level of drinking during early adulthood (PND 54-66), although this relationship was not seen across the entire 10-day conditioned fear training phase. We also found that the level of consumption during early adulthood was associated with the level of fear in both the day 2 and 1 month tests, with parallel increases in the high and low drinkers so that the groups never converged. While these findings could be explained by high alcohol consumption having effects that are not seen with lower alcohol consumption levels, there were no significant differences between the water-only group and the alcohol access group as a whole. In addition, the average suppression ratio for the water group was always in-between the average value for the high and low alcohol consumption sub-groups during the first conditioning session, the average of the 10 fear conditioning sessions, the day 2 test and the 1-month test (Table 1). This pattern mirrors our previous finding that, in a procedure with a single fear conditioning session followed by testing, the average values for water-only access group were between those of the high and low alcohol consumption sub-group for all fear conditioning measures [12]. We also replicated the general finding from our previous report [12] that, while the high alcohol consumption rats had predominately low fear expression during test, the low consumption rats had a mix of rats with high fear expression with some rats with low fear expression. This led to a pattern, just as in our previous report [12], where there was a significant difference between the high drinking vs. low drinking rats when they were divided into groups, but there was no significant correlation between early adult alcohol consumption and conditioned fear in the two cued fear extinction tests. However, to our surprise, we did find a significant correlation between early adult alcohol consumption and conditioned fear during the first day of fear training. Although later training and testing was different between this experiment and our previous investigation of relationships between alcohol consumption and conditioned fear [12], the experiments were largely identical up to first day of conditioning. Additional experiments will be needed to determine whether the significant correlation (found only in the current experiment) is reliable. Regardless, we have found a difference between high drinking and low drinking rats in fear expression during the first day of conditioning in both the current experiment and this previous experiment, so this pattern appears reliable. These findings suggest that the association between high alcohol consumption and low conditioned fear in the putative HALF-FIELDER rat is reliable and generalizes to other conditioned fear procedures.

Table 1:

Suppression ratios for adolescent/early adult alcohol consumption sub-groups and adolescent/early adult water-only group

	Low Alcohol Consumption Group*	Water-Only Group	High Alcohol Consumption Group
Day 1 Fear Conditioning	0.32±0.07	0.14±0.05	0.08±0.07
10 Day Fear Conditioning (average)	0.42±0.10	0.30±0.06	0.18±0.08
Day 2 Test	0.28±0.07	0.21±0.03	0.02±0.07
1-Month Test	0.45±0.16	0.44±0.09	0.22±0.08

^*Intermediate rats were excluded

In addition, we found some evidence for differences in contextual fear in the high vs. low drinkers, with the high drinkers having higher lever-press rates midway through the Day 1 lever-reacquisition session. This could suggest that the high drinkers exhibited lower contextual fear and/or faster extinction to the context. However, these findings must be interpreted with caution for two reasons. First, baseline lever-pressing is a crude measure of contextual fear, as baseline lever-press rates exhibit individual differences in the absence of fear conditioning. Second, this pattern only reached significance if the intermediate drinking rats were included in the low drinking group (as in our secondary analysis) and not if the intermediate rats were excluded from the analysis. Additional studies will be required to determine if the HALF-FIELDER/non-HALF-FIELDER phenotypes have reliable effects on contextual fear.

Our data complement previous data showing that alcohol preferring P rats exhibit greater contextual fear-induced increases in fear-potentiated startle and require less training to exhibit light cue-induced fear-potentiated startle than alcohol non-preferring NP rats [42], although the relationship between alcohol consumption and fear in that study was opposite to our finding. It is unclear what different mechanisms would be responsible for the different fear conditioning pattern in HALF-FIELDER/non-HALF-FIELDER vs. P/NP rats, but this could reflect different neural mechanisms responsible for the different patterns of consumption in these 2 phenotypic dichotomies. Some mechanisms associated with higher drinking could lead to higher conditioned fear and other mechanisms associated with higher drinking could lead to lower conditioned fear. Many neurotransmitter systems have been shown to differ in P vs. NP rats [36, 43], so any one of these other systems (or a combination of them) could be responsible for the different pattern of fear conditioning in the P vs. NP rats with a different mechanism responsible for the conditioned fear pattern in HALF-FIELDER vs. non-HALF-FIELDER rats.

One caveat, however, is that the rats in current study were shipped for arrival on PND21 and individually housed from their arrival until the end of experimentation. Stress can be caused both by shipping (as reviewed in [44]) and by individual housing (as reviewed in [45]). This means our rats could be considered to have experienced pre-adolescent stress due to shipping and individual housing immediately upon arrival in our facility and have experienced adolescent stress due to individual housing throughout the alcohol access period. Stress during pre-adolescence/adolescence (including stress due to social isolation) can affect levels of alcohol consumption [46, 47] and can have long-term effects during adulthood on fear conditioning (with probabilistic cue-shock pairings) and fear extinction [48–50], It is possible, therefore, that our behavioral assessments were affected by this pre-adolescent/adolescent stress and the patterns we observed would not be replicated in animals that were bred in a laboratory (avoiding transportation stress) and pair- or group-housed (avoiding stress from individual housing).

It is unclear what neural mechanisms would lead to lower fear conditioning in rats with higher alcohol consumption, and all proposed mechanisms are speculative at the current time. However, one possibility is that greater endogenous opioid activity (at the ligand- or receptor-level) could lead to increased alcohol consumption and lower fear conditioning. Opioid receptors are implicated in alcohol’s reinforcing properties [51], Likewise, activity of endogenous opioids (acting at mu-opioid receptors) has been shown to limit fear learning in training procedures with or without cue-competition [52–56], Thus, greater endogenous opioid activity/sensitivity in some individuals could lead to greater alcohol consumption and lower conditioned fear. Notably, humans with a family history of alcoholism, but who are not alcoholics themselves, exhibit greater sensitivity to naloxone than people without a family history of alcoholism [57], There might also be related structural/functional differences in anterior cingulate cortex (ACC) associated with the propensity to consume alcohol and sensitivity to pain. Notably, ACC is involved in pain processing [58], cue-induced analgesia (with alterations in ACC opioid signaling) [59], and avoidance of neuropathic pain (with a role of ACC opioids in this avoidance) [60, 61], In humans, people with a family history of alcoholism (but who are not alcoholics themselves) exhibit neurochemical differences in ACC compared to people without this history [62] and these 2 populations exhibit differences in ACC activation (measured with fMRI) during a go/no-go task [63, 64], In our laboratory, errors in the reversal learning phase of a go/no-go task correlated with high alcohol consumption as the “LDER” (“low discontinuation errors in reversal”) part of the HALF-FIELDER phenotype [15], However, more research will need to be conducted to assess whether differences in activity of the opioid system or ACC structure/function are associated with the HALF-FIELDER/non-HALF-FIELDER phenotypes.

Elsewhere we have discussed strain differences in conditioned fear in inbred and outbred rats and individual differences in conditioned fear within outbred strains (see [12] for more details). However, the relationship of one pair of phenotypes requires further discussion. Our findings that the high-drinking HALF-FIELDER rat shows lower conditioned fear than the lower drinking non-HALF-FIELDER rat (here and in [12]) resemble a pattern observed in another pair of phenotypes, the sign-tracking and goal-tracking rats [39, 65, 66], The sign-tracking/goal-tracking phenotypes, in which sign-tracking rats predominately approach a cue that predicts food reward and the goal-tracking rats predominately approach the site of food delivery before the food is present, exhibit differences in a number of motivated behaviors, including many cue-related behaviors for a cocaine reinforcer or reward (as reviewed in [66]). Notably, there is evidence for a relationship between the propensity to consume alcohol and sign-tracking/goal-tracking, although the direction of the relationship is mixed in different assessments. When Lister rats were given CIA during adulthood and divided into the top 12.5% (high drinkers) and bottom 12.5% (low drinkers) in terms of alcohol consumption, the high drinkers exhibited more sign-tracking behaviors (lever-presses) and fewer goal-tracking responses (head-entries) than the low drinking rats [27]. Conversely, previous research has found that alcohol-preferring inbred P rats exhibit fewer sign-tracking behaviors (lever-presses) and more goal-tracking responses (head-entries) than non-alcohol preferring NP rats [67]. One additional behavioral trait that co-varies with sign-tracking/goal-tracking is conditioned fear learning. In two experiments, Morrow and colleagues found higher conditioned fear to a tone cue (in a test session after a single fear conditioning session) in sign-tracker than in goal-tracker rats [26]. In another experiment, Morrow and colleagues also examined whether sign-tracker and goal-tracker rats would differ in conditioned fear in an extended training fear incubation procedure. Morrow and colleagues found that sign-tracking rats or intermediate rats (which exhibit a mix of behaviors towards to the food-predictive cue and the site of food delivery) exhibited the typical fear incubation pattern, with low fear 3 days after end of fear conditioning that was increased in a test 33 days after the end of fear conditioning [22]. Conversely, their goal-tracking rats exhibited high fear 3 days after the end of fear conditioning that remained at the same level at the 33-day test, such that all three groups exhibited similar fear at the 33-day interval.

Our current experiment represented a test of whether the goal-tracking and HALF-FIELDER phenotypes (which both exhibit low conditioned fear to an auditory CS after a single-session fear conditioning) represent the same phenotype that has simply been identified using non-overlapping behavioral assessments, with the sign-tracking and non-HALF-FIELDER phenotypes representing another phenotype. In this case, we would expect that HALF-FIELDER rats would exhibit high fear soon after extended fear conditioning, with no increase in fear across a one-month interval, while the non-HALF-FIELDER rats would exhibit low fear soon after fear conditioning that grows over the one-month retention interval. However, this is not the pattern that was found. Both the high drinking and low drinking rats exhibited parallel increases in conditioned fear, with no convergence in fear levels at the 1-month test. This suggests that the HALF-FIELDER phenotype is not the same as the goal-tracking phenotype, but these represent 2 independent phenotypes that each have effects on conditioned fear, with non-HALF-FIELDER and sign-tracking phenotypes similarly independent. Future research will be needed to clarify similarities and differences between the two phenotypic dichotomies.

4.3. Conclusions

In conclusion, we found that low levels of voluntary alcohol consumption did not alter conditioned fear during fear conditioning or testing in the extended training fear incubation procedure. However, the level of alcohol consumption was associated with the level of conditioned fear in the first day of conditioning and during the early day 2 test (when fear is low) and the 1-month test (when fear increases). Our results suggest that some of the relationship between alcohol consumption and conditioned fear may be caused by individual differences, rather than a neurotoxic effect of alcohol on the neural substrates of fear. Notably, the pattern seen in the high vs. low drinking rats was different from the pattern seen in sign-tracking vs. goal-tracking rats. This suggests that our HALF-FIELDER and non-HALF-FIELDER phenotypes may represent a different dimension of individual differences than that observed in sign-tracking vs. goal-tracking rats. Future research will examine other behavioral traits and neurobiological substrates associated with these putative phenotypes.

Highlights.

• Alcohol consumption in adolescence and adulthood did not affect conditioned fear

• In the adolescent alcohol group, high consumption was associated with lower fear

• High drinkers exhibited lower conditioned fear in the first conditioning session

• High drinking, low drinking and water-only groups all exhibited fear incubation

• High drinking rats exhibited lower fear in the day 2 and 1-month tests