Effects of a synthetic predator odor (TMT) on freezing, analgesia, stereotypy, and spatial memory.

Exposure to laboratory stressors influences a wide range of behavioral and neurophysiological processes in rats. Historically, much of the classic research on stress in rodents has examined how aversive stimulation, such as painful electric shock, affects unlearned defensive responses, stress-induced analgesia, drug reactivity, learned responses, and more complex cognitive behaviors. Because many of the behavioral testing procedures designed to assess the influence of foot shock are similar to those used in the present research on the effects of exposing rats to a predator stimulus, a brief review of some of the specific findings of these earlier studies is important for making possible interstressor comparisons.

Electric shock has been found to be an ideal laboratory stressor because its intensity and duration can be easily and precisely controlled. Many studies have shown that exposure to foot shock results in flight or fleeing reactions (Bolles & Fanselow, 1980), whereas rats typically engage in fear-motivated freezing as a defensive conditioned response when a discrete stimulus or the experimental context has been paired with shock as an unconditioned stimulus (Fanselow & Baackes, 1982; Fanselow, Sigmundi, & Williams, 1987). Exposing rats repeatedly to a series of inescapable tail shocks, as is typically done in studies of learned helplessness (Maier & Seligman, 1976), produced analgesia regulated by the activation of an endogenous opioid system (Maier, 1989). In addition, this type of stress exposure also influenced the reactions of rats to an injection of exogenous opiates in terms of both an increased drug tolerance (Drugan, Grau, Maier, Madden, & Barchas, 1981) and exaggerated withdrawal reactions (Williams, Drugan, & Maier, 1984). Similarly, repeated presentations of inescapable shock increased the frequency of specific reactions, such as response stereotypy, following an injection of a small dosage of d-amphetamine (Anisman, Hahn, Hoffman, & Zacharko, 1985). Prior exposure to shock interfered with the following: one-trial classical conditioning of rats learning to bury a prod that was the source of a single electric shock (Williams, 1987), escape learning by double crossing in a shuttlebox (Maier, 1984), operant learning for a positive reinforcer (Rosellini, 1978), and choice or position learning in a Y maze (Jackson, Alexander, & Maier, 1980). Furthermore, much research has been devoted to the neurophysiological mechanisms that underlie, or are at least correlated with, these shock-induced disruptions in behavior. For example, inescapable shock has been reported to increase serum corticosterone and acetylcholine levels, decrease immune functioning, and alter norepinephrine, dopamine, and serotonin activity (see Maier, 2003; Maier & Watkins, 1998, for review of this literature).

As part of a general shift towards a more ethological or ecologically valid perspective, there has been a renewed interest in how rodents respond to natural stressors. Considerable research has been reported from our laboratory on male rats that have experienced single, versus repeated, sessions of being attacked and defeated by an alpha colony resident. Williams and Scott (1989) found that the odors of alpha residents became powerful conditioned fear stimuli after a rat had been defeated for one session by an alpha resident. The combination of this defeat experience and later testing in an observation chamber, with soiled bedding odors of the alpha rat, produced significant increases in freezing and suppressed the rat's natural tendency to use bedding material to bury a prod that was the source of a single shock. Williams, Worland, and Smith (1990) found that this paradigm of social defeat not only elicited freezing and interfered with prod burying, but the odors of the alpha conspecifics also produced stress-induced analgesia during a post-defeat formalin test for pain sensitivity. In addition, these researchers revealed that this analgesia was mediated by an endogenous opiate system, because it was blocked when the intruder was administered an injection of an opioid antagonist either prior to the defeat or the test session. Further evidence that the alpha odors were conditioned fear stimuli was demonstrated by showing that these behavioral and analgesic effects were extinguished when the rats were given a 12-hr session in the presence of the alpha odors without the dominant resident being present (Williams, Rogers, & Adler, 1990).

Naive rats show unconditioned fear reactions when exposed to natural predators or the odor of a natural predator. For example, rats display pronounced behavioral and physiological changes when exposed to a cat or the odor of a cat (e.g., from its fur, a worn collar, or the soiled bedding of its litter box). This experimental paradigm for studying the effects of natural stressors is receiving attention because it reliably elicits fear and does not involve the use of painful or nociceptive stimulation. The innovative research of the Blanchards distinguished between the behaviors elicited by a live cat and those by cat odor (Blanchard, D. C., & Blanchard, 1988). Exposure to a live cat produces a much stronger defensive response with long durations of retreat inside a burrow system and freezing accompanied by ultrasonic vocalizations. However, the odor of a cat, obtained from a cloth impregnated with cat hair, produced a distinctive risk-assessment pattern of defensive behavior. Rats would leave the burrow system, assume a stretched posture with their backs flat, and make a series of approach-withdrawal responses to the odor source (Blanchard, R. J., & Blanchard, 1989). The persistence of the risk assessment behavior, led these researchers and others (Dielenberg & McGregor, 2001) to conclude that exposure to the full predator stimulus results in rigorous retreat and freezing (fear), whereas the presentation of cat odor represents a "partial predator stimulus," eliciting uncertainty (anxiety) and a motivation to gather more information to assess the degree of predatory imminence.

Some of the findings and conclusions of the Blanchards are somewhat supported by the studies by File and her colleagues, using a different methodology (File, Zangrossi, & Andrews, 1993; Zangrossi & File, 1992). In these studies, a damp cloth was used to procure the cat odor by rubbing it against a cat, and then observing the behavior of rats with the cloth in their home cages. The rats showed an avoidance of the cat-odor cloth and spent considerable time sheltering themselves under the food hopper. These reactions were not found to habituate over repeated exposure sessions (Zangrossi & File, 1994), suggesting that cat odor produced an overexpressed anxiety or phobic-like reaction. In addition, these researchers reported that the odor of cats resulted in an increase in serum corticosterone release (File, Zangrossi, Sanders, & Mabbutt, 1993). However, unlike the research by the Blanchards on anxiolytics and risk assessment, they found that the anxiety responses elicited by cat odor was relatively insensitive to the effects of benzodiazepines (Zangrossi & File, 1994).

Dielenberg and his associates (Dielenberg, Arnold, & McGregor, 1999; Dielenberg & McGregor, 1999) used a novel "hide box" apparatus and a piece of worn cat collar to examine the responses of rats to cat odor. They found that rats spent most of the time hiding when exposed to the cat collar and this response habituated over repeated sessions. Hiding from a worn cat collar was reversed by injecting rats with a low dose of the anxiolytic drug, midazolam, and rats exposed to collars also displayed increased anxiety in an elevated plus maze. These findings also cast doubt on the supposition that predator odors lead to phobic reactions, rather than anxiety. Moreover, their recent research indicates that exposing rats to a cat collar produces risk assessment and increases in blood pressure (Dielenberg, Carrive, & McGregor, 2001) that is accompanied by strong activation of the limbic, hypothalamic, and midbrain circuitry as quantified by Fos immunohistochemistry (Dielenberg, Hunt, & McGregor, 2001).

Research conducted in our laboratory over the past decade has shown that protected exposure to a cat and/or testing with cat odor, obtained from soiled litter-box bedding, profoundly affects a wide range of behavioral and physiological processes in rats. Williams and Scott (1989) examined the effects of noncontact exposure to a cat on subsequent conditioned defensive burying of a wall-mounted prod, which was the source of a single shock, in a box that had bedding mixed or not mixed with soiled cat litter. Cat exposure did not have any effect 1 day later, but the presence of cat odor litter during testing, suppressed the amount of defensive burying of the shock prod and resulted in moderate levels of freezing and risk assessment. Lester and Fanselow (1985) also reported that cat odor, relative to the novel control odor of citronella, elicited freezing and analgesia in naive rats during formalin tests. Furthermore, this analgesia was blocked by administering an opiate antagonist prior to testing, implying that the cat odor resulted in the activation of endogenous opiates. This finding and the results of several foot-shock studies cited earlier, led Williams and Barber (1990) to examine the effects of cat versus no-cat exposure on a single injection of a small dose of d-amphetamine. Observational testing in an activity monitor was done with or without soiled cat litter evenly distributed on the floor of the apparatus. Increases in both the frequency and duration of amphetamine-induced response stereotypy were found for cat-exposed subjects and those only tested with cat litter. Finally, Williams, Baker, Gress, and Givens (1998) examined the effects of cat exposure on the postacquisition performance of rats trained on a 20-s delayed alternation task in an elevated T maze. Stressed, in contrast to nonstressed, rats showed a disruption in their performance, thus implying an impairment in spatial working memory. This conclusion is somewhat supported by another group of investigators (Diamond, Park, Heman, & Rose, 1999), who found that exposure to a cat disrupted subsequent working memory of rats trained and tested in a radial arm water maze. In conclusion, these findings clearly demonstrate that exposure to a cat or cat odor produced fear-motivated alterations in behavior, physiology, and cognition.

Although the use of cats and cat odor as predatory stressors have revealed reliable and interesting effects, some of the findings have been contradictory. For example, several researchers claim that the dominant defensive response to cat odor is risk assessment (e.g., Blanchard, R. J., & Blanchard, 1989), while others believe the freezing occurs along with risk assessment (e.g., Williams, 1999; Williams & Scott, 1989). Probably the major reason for many of these discrepant results is that the stimulus used to induce predator stress has varied considerably. Cat exposure is known to have different effects depending on such factors as the age, diet, and experimental history of the cat. The use of odor from a cat is also problematic because the exact origin of the odor is typically not clearly defined and it is virtually impossible to control or hold experimentally constant. For example, the odor of a cat is typically a mixture of odors produced by the cat's urine, feces, gland secretions, hair, and sometimes the material or litter used as cat bedding. Moreover, the amount, concentration, and age of the substances responsible for these odors are difficult to quantify and calibrate. For these reasons, much of the current research on the effects of predator odors in rats has involved the use of 2,5-Dihydro-2,4,5-trimethylthiazoline (TMT), which is a carbon compound isolated from the feces of red fox (Vulpes, vulpes), a natural predator of rats (Vernet-Maury, 1980). Because this compound is synthesized using a standard biochemical process, its chemical structure and molecular weight are known, and therefore its amount and concentration can be better controlled across subjects, groups, and studies than other sources of natural fear odors.

The seminal research using TMT as a predator stressor (Vernet-Maury, 1980; Vernet-Maury, Polak, & Demael, 1984) demonstrated that this compound produced fear reactions equivalent to responses elicited by whole fecal extracts of fox, in terms of suppressing open-field activity and increasing the corticosterone levels of rats. These investigators also claimed that the reactions were innate because rat pups were found to respond to the odor of TMT from birth to adulthood, without evidence of habituation of fear-elicited behaviors during repeated open-field tests (Vernet-Maury, Constant, & Chanel, 1992).

Some recent studies have failed to observe fear-like behavioral effects of this synthetic predator odor in terms of freezing or risk assessment behavior (Dielenberg & McGregor, 2001; McGregor, Schrama, Ambermoon, & Dielenberg, 2002; Morrow, Redmond, Roth, & Elsworth, 2000). Wallace and Rosen (2000) speculated that the odor of TMT is only a signal for the presence of a predator and, under some conditions, it may not elicit sufficient fear to induce predator defensive responses. When testing was done in a large chamber, Wallace and Rosen (2000) found that rats displayed fear-related behaviors to TMT, including immobility, avoidance, and less exploratory behavior. However, testing with TMT in a smaller chamber induced the species-typical fear response of freezing at a much higher level than for tests using other intense, nonpredator control odors. Furthermore, the use of larger amounts of TMT systematically resulted in more freezing, and there was no habituation of freezing to TMT, within or between sessions, which is consistent with the findings from cat odor studies.

Hotsenpiller and Williams (1997) did a conditioning study using TMT as a potential conditioned stimulus by pairing it with an injection of a anxiogenic drug as an unconditioned stimulus. Although they found TMT to be a very effective conditioned stimulus, they also noted that TMT, by itself, resulted in some levels of freezing and analgesia when rats were tested in a relatively small enclosure with limited ventilation in the ceiling. These findings, which are consistent with Wallace and Rosen (2000) and much of our recent pilot work, suggest that fear-induced responses can be elicited by TMT if rats are tested in an environment in which the odor of TMT is salient, perhaps increasing the perceived imminence of a predator.

Morrow, Redmond, et al. (2000) indicated that TMT, in contrast to a novel control odor, increased dopamine metabolism in the prefrontal cortex, an effect that was also seen with stressors such as foot shock, conditioned fear, and social stress. Using immunocytochemical techniques, Redmond, Morrow, Elsworth, and Roth (2002) reported an activation of the expression of the immediate-early c-fos gene, following exposure to foot shock and TMT, but not a control odor. These researchers postulated that the increased expression of the c-fos gene produced long-term changes in DA neurons that have been implicated in stress-induced disruptions in memory and appears to alter the rat's response to subsequent exposure to stress. Consistent with this conclusion, Morrow, Roth, and Elsworth (2000) have shown that TMT exposure resulted in impaired short-term working memory of rats tested on a delayed, nonmatching-to-sample task. Specifically, nonstressed control rats were exposed to one set of objects and when tested, after a delay interval, they reliably showed good memory of the familiar object by exploring a novel object more. In contrast, rats exposed to TMT, during the delay interval, failed to selectively explore the novel object compared to the familiar one. Definitive conclusions concerning these findings must remain tentative because there was no control odor used in this study and the TMT-stressed rats may have explored the novel object less because of a neophobic reaction, rather than a deficit in working memory.

The primary focus of the present research is to examine a wide range of behavioral and neurophysiological reactions of TMT using paradigms and procedures that have yielded reliable and valid findings with other stressors. This approach is expected to reveal more specific information about the nature of TMT as an unconditioned source of fear and as a predator stressor. Using moderately large samples of TMT and the control odor of citronella, Experiment 1 was done to confirm if exposing rats to TMT resulted in considerable freezing and analgesia. This study also examined if the predicted analgesia to TMT was caused by the activation of endogenous opiates. Experiment 2 examined whether exposing rats to TMT led to augmented stereotypy behavior following a single injection of a small dose of d-amphetamine. The purposes of Experiments 3 and 4 were to see if prior exposure to TMT would influence subsequent cognitive processes involving spatial working memory and spatial reference memory, respectively. Finally, Experiment 4 was also designed to see the anxiolytic effects of a benzodiazepine agonist, during TMT exposure, might mitigate the predicted TMT-induced disruption of spatial reference memory.

Experiment 1

Probably the two most documented effects of exposing rats to stressful situations are defensive freezing behavior and analgesia or pain-reducing reactions. Research has demonstrated that both of these responses are motivated by fear (Fanselow, 1980) and stress-induced analgesia can be regulated by either opioid or nonopioid processes (Maier, 1989). Numerous investigators have assessed the level of stress-induced analgesia reactions to a variety of stressors using the formalin testing procedure (e.g., Fanselow, 1984; Helmstetter & Fanselow, 1987; Williams, 1999). Rats tested with this procedure are administered a subcutaneous injection of a small dose of a diluted solution of formalin in one of their hind paws. Shortly after the injection, the rat engages in considerable licking of the paw as a type of recuperative response. Using a time sampling procedure, the frequency of bouts of paw licking during an observation period is used to infer the level of irritation or nociceptive experience perceived by the rat. In contrast to other techniques (e.g., flicking of the tail or licking a paw exposed to a heat source), the formalin procedure for testing analgesia is particularly effective because it allows subjects to explore their surroundings and engage in various nonrecuperative responses (e.g., freezing). For this reason, the formalin test has been used to examine analgesia and defensive responses elicited by various ethological stressors, such as the odors of dominant or stressed conspecifics (Hotsenpiller & Williams, 1996; Williams, Worland, et al., 1990) or of a cat (Lester & Fanselow, 1985; Williams, 1999). Furthermore, researchers have demonstrated that administering an opiate antagonist (e.g., naloxone, naltrexone), either prior to stress exposure or later during formalin testing, can sometimes block the occurrence of paw licking. Reductions in recuperative licking caused by such antagonists have been assumed to indicate that the analgesic reaction elicited by a particular stressor is regulated by an endogenous opiate system (e.g., Helmstetter & Fanselow, 1987; Williams, 1989).

More relevant to the present research, defensive freezing (Blanchard, Fukunaga, & Blanchard, 1976; Williams & Scott, 1989) and opioid-mediated analgesia (Lester & Fanselow, 1985; Williams, 1999) have been observed when rats were exposed to a cat or soiled cat litter. Hotsenpiller and Williams (1997) reported that the synthetic predator odor of TMT elicited a strong conditioned analgesic reaction after it had been paired with several injections of an anxiogenic drug. These investigators also noted that TMT produced a fairly reliable, but less powerful, unconditioned analgesic reaction when presented alone.

In light of the previously mentioned methodological advantages of using TMT as a predator stressor, as opposed to a cat or the odor of a cat, the present experiment was conducted to confirm if TMT elicits freezing and analgesia during a formalin test session and also to determine if this analgesia might be due to endogenous opiates. Specifically, two groups of rats were first given an injection naltrexone (NALT), a long-lasting opioid antagonist, or a control injection of saline (SAL). After a period elapsed for the naltrexone to become effective, all subjects were injected subcutaneously with formalin in one of their hind paws. Shortly after receiving the formalin injection, subgroups of rats were individually observed in an odor box that contained either TMT or the novel odor of citronella (CIT), which has been used extensively as a control odor because it has been shown not have any emotional or hedonic value for the rat (e.g., Lester & Fanselow, 1985; Williams, 1999). It was predicted that rats injected with saline and exposed to TMT would show less paw licking than all of the other groups, thus indicating that TMT elicited analgesia that was reversed by an opioid antagonist. Furthermore, it was hypothesized that both of the TMT groups would engage in more freezing than the CIT groups, because naltrexone should not affect fear-motivated behavior.

Method

Subjects

The subjects were 40 experimentally naive male rats, approximately 325 g, bred in the Kenyon College Psychology Department laboratory from Sprague-Dawley (Harlan) descent. Two rats were maintained in tub cages with food and water available ad lib until 2 weeks before the experiment, at which time the rats were placed in separate home cages with free access to food and water. All experimental manipulations were conducted during the light phase of a 12:12-hr light:dark schedule.

Apparatus

Habituation and observational testing sessions occurred in 36 X 36 X 41 cm metal boxes. A moderately ventilated clear Plexiglas lid was attached to the top of the boxes, and a 3-cm layer of corn-cob bedding ("Bed-O-Cobs") was evenly spread over the floor. Hanging from the center of the lid was a 2-cm long wire holding a 9-cm Whatman #1 circular filter paper (medium-fast filter speed, medium crystalline retention) for presenting the odors. Formalin injections were given to each rat while it was confined in a rat restrainer that was located in a designated injection area which was separate from the housing and testing rooms. To ensure that the two test odors were isolated, the boxes used to test each odor were located in ventilated rooms at the opposite ends of a long hallway. Each test room had a Panasonic WV-3260 video camera, mounted on a tripod above the boxes for videotaping the behavior of the rats, that was monitored in an adjacent room using a Sony SSM-2010 color monitor. The sensitivity of the camera enabled clear recordings to be obtained by illuminating the boxes with a 5-W red lamp mounted above the camera. A white-noise generator produced a 75-dB SPL background noise that was used to prevent extraneous sounds from distracting the rats during habituation and testing. Most of this apparatus was used previously in our lab for doing formalin tests (Williams, 1999; Williams, Worland, et al., 1990).

Drug

A solution of 7 mg of naltrexone per kg of the subject's weight and the same volume of saline were administered intraperitoneally (IP) to the experimental and control groups, respectively. This dosage of naltrexone, a long-acting opioid antagonist, has been used by many researchers (e.g., Maier, 1986; Helmstetter & Fanselow, 1987) to assess opioid activity resulting from exposure to inescapable shock. More pertinent to our study, this amount of naltrexone has been used to demonstrate opiate involvement in analgesia found during formalin tests following exposure to natural social and predator stressors (Lester & Fanselow, 1985; Williams, Worland, et al., 1990). Furthermore, this dosage of the antagonist was found to block recuperative paw licking, but did not affect defensive responses (e.g., freezing).

Odors

The 2,5-Dihydro-2,4,5-trimethylthiazoline (TMT) was purchased as a neat liquid (molecular weight of 129) from Phero Tech Inc. (British Columbia, Canada). The citronella (CIT), in neat liquid form (citronellal, with a molecular weight of 154.3), was purchased from Sigma-Aldrich Co. (St. Louis, MO). Separate microsyringes were used to apply either 100 [micro]l of TMT or CIT to the filter paper attached to the lids of separate odor boxes. This amount of test odor was used because pilot studies revealed that these odors were absorbed about equally when samples of activated charcoal were placed in the testing box for a 30-min period. Both odors were also judged to be comparable in intensity when they were applied to the filter paper. However, our method of attempting to equate the intensity of the two odors does not mean that they had the same effect on the rat's olfactory system (see Hotsenpiller & Williams, 1997).

Procedure

All rats were handled for 3 days and habituated for two 30-min daily sessions in the boxes without test odors. Prior to each habituation session and the subsequent test session, a fresh layer of corn-cob bedding was placed on the floor and the white noise and red light turned on. The rats were then randomly assigned to one of four groups of 10 subjects each, involving a 2 (drug) X 2 (odor) factorial design. More specifically, these groups were given the following treatments: injected with saline and tested with the neutral odor of citronella (SAL/CIT), injected with saline and tested with TMT (SAL/TMT), injected with naltrexone and tested with citronella (NALT/CIT), and injected with naltrexone and tested with TMT (NALT/TMT).

On the day of testing, the groups were administered an injection (IP) of either saline or 7 mg/kg of naltrexone (see Maier, 1989, about onset and duration of the effect of naltrexone) and returned to their home cages. Fifteen min later, they were given a 50-[micro]l injection of a diluted solution (15%, in an isotonic saline vehicle) of formalin just under the dorsal surface of the right hind paw (see Fanselow, 1984, for details) and again returned to their home cages. The odor manipulations were then done by dispensing TMT and citronellal in the middle of the Whatman filter paper suspended from the middle of the clear lid of the box. The subjects were then taken from their home cages and individually placed in one of the odor boxes 20 min after the formalin injection. This interval has been shown to be necessary for formalin to become an irritant to the rat (Fanselow, 1984).

Video recordings were made of the subject during the entire 16-min test session, and the experimenter scored responses from a monitor in a separate room. Tabulations were made of the precise response that occurred when an audible signal was heard every 8 sec and recorded on videotape. Throughout the test session, the following responses were recorded: (1) paw licking: holding and licking the formalin-injected paw while sitting or lying, (2) freezing: a crouched position with no movement except for respiration, and (3) general activity: all other behaviors. This time-sampling method has been used extensively by many researchers (e.g., Fanselow, 1984; Fanselow & Helmstetter, 1988; Hotsenpiller & Williams, 1996). High levels of reliability (coefficients above .90) in scoring responses were found during periodic checks done by observers blind to the experimental conditions. After each rat was tested, the filter paper and bedding were removed and the test boxes were cleaned thoroughly with a 5% ammonium hydroxide solution. Further control of the odors was accomplished by conducting the odor tests in separate testing rooms and on different days.

Results and Discussion

For each subject, the number of occurrences of each of the behaviors scored during the 16-min formalin test was summed across the 120 time samples. Preliminary tests indicated that the data met the assumptions of normality and homogeneity of variance, thus a 2 (saline vs. naltrexone) X 2 (citronella vs. TMT) independent-groups analysis of variance (ANOVA) was done for each dependent measure. When the outcomes of the ANOVAs yielded significant main effects or an interaction, Newman-Keuls post hoc tests were done to assess differences between specific groups (Winer, 1971). Statistically significant effects or differences were evaluated, in terms of p values less than .05, .01, or .001, depending on the specific outcome. This general procedure, concerning the analysis and reporting of statistical results, was also followed in the subsequently described experiments.

Figure 1 presents the mean number of occurrences of paw-licking responses that coincided with the time-sampling signal for the four groups during the formalin test. The ANOVA revealed a significant effect for the odor manipulation, F(1, 36) = 4.97, p < .05, and a significant interaction between drugs and odor, F(1, 36) = 4.13, p < .05. Post hoc tests indicated that Group SAL/TMT differed significantly (ps < .01) from the other three groups, which did not differ significantly from one another. The significant reduction in recuperative paw licking in Group SAL/TMT, relative to Group SAL/CIT, indicates that the predator odor of TMT produced a stress-induced analgesia that was not simply the result of the novelty of the odors. Paw-lick suppression during formalin testing has been previously found when rats have been exposed to a cat or cat odor (e.g., Lester & Fanselow, 1985), and Hotsenpiller and Williams (1997) previously reported that TMT produced a moderate, but reliable, analgesic reaction. Also consistent with our hypotheses, the amount of paw licking was found to be similar for Group NALT/CIT and Group NALT/TMT, indicating that naltrexone was capable of completely reversing the analgesia elicited by TMT. This is the first study showing that the predator odor of TMT activates a robust analgesic reaction that appears to be entirely mediated by an endogenous opiate system.

[FIGURE 1 OMITTED]

The mean number of occurrences of freezing for each group during the formalin test are shown in Figure 2. The ANOVA showed that there was a significant effect of the odors, F(1, 36) = 76.44, p < .001, and a significant interaction between the drug and odor manipulations, F(1, 36) = 7.37, p < .05. Post hoc tests revealed that each of the TMT groups differed significantly (ps < .01) from their corresponding CIT groups. In addition, Group SAL/TMT showed significantly more freezing than the two citronella groups (ps < .01) but did not freeze significantly more than Group NALT/TMT. These results demonstrate that freezing occurred to TMT, regardless if rats were previously injected with saline or naltrexone. Furthermore, the combined findings of this study indicate that naltrexone blocked analgesia, as measured by paw licking, but it did not alter the levels of freezing, which is a defensive response that is motivated by fear and not by pain. Some researchers have also reported that TMT elicits freezing (e.g., Wallace & Rosen, 2000), whereas others claim that freezing is not the dominant defensive behavior elicited by TMT or by other predator odors (e.g., Blanchard, Blanchard, & Hori, 1989; and see the General Discussion for more on this issue).

[FIGURE 2 OMITTED]

In terms of other behaviors, rearing and slow head swaying from side to side were also found to occur, particularly in the two TMT groups. However, it was difficult to specify the onset and termination of these responses, thus no statistical analyses were done on these measures. In addition to the ANOVAs, Pearson correlations between paw licking and freezing were conducted for each of the groups. Although all of the correlation coefficients were negative, none were statistically significant, indicating that freezing was not the cause for the suppression of paw licking. Furthermore, the amount of freezing observed for the TMT conditions during testing did not preclude the subjects from engaging in paw licking and other responses.

Experiment 2

Stress exposure has been shown to enhance reactions to d-amphetamine in terms of unilateral circling and stereotypy behavior (e.g., Anisman et al., 1985). It has long been believed that the stress-induced sensitization reactions to amphetamine are the result of increased dopamine activity along with alterations in the level of norepinephrine (Carlson, 1994; Kokkinidis & Anisman, 1981). Of particular relevance to Experiment 2, Williams and Barber (1990) reported a study which examined amphetamine sensitization in rats following exposure to a cat and cat odor. More specifically, these investigators exposed male and female rats to a cat and subsequently observed them in an activity monitor for 1 hour following an injection of a small dose of d-amphetamine. Separate groups were tested with either the clean or soiled bedding from a cat on the floor of the monitor box. The results of this study showed that the rat's reactivity to amphetamine was affected by the sex of the subject, cat exposure, and the presence of cat odor during testing. Overall, female rats engaged in more stereotypy (i.e., repetitive head and forepaw movements) than male rats, but only the males showed significant differences in amphetamine-induced stereotypy as a function of the experimental conditions. Male rats showed significantly more stereotypy, both in terms of the number of occurrences and the durations of stereotypic responses, if they were exposed to a cat and tested with cat odor. Rats that were tested with cat odor, but not exposed to a cat, showed significantly more stereotypy than rats tested with clean bedding.

The purpose of Experiment 2 was to investigate if TMT, like the odor of a cat, would sensitize these same stereotypic reactions after a single amphetamine injection. Following one habituation session in the activity-monitor box, groups of subjects were injected with 1 mg of d-amphetamine (AMP) or saline (SAL). Subgroups of rats were then tested individually with either the neutral odor of citronella (CIT) or the predator odor (TMT) presented in the box. After the time elapsed for amphetamine to be absorbed, recordings of stereotypy frequencies and durations were obtained from an activity monitor during the course of a 1-hr session. It was hypothesized that each of the stereotypy measures would be greater in the AMP than the SAL condition, due to the stimulant effect of the drug. More importantly, these measures of stereotypy were predicted to be the greatest for AMP/TMT condition based on the results of Williams and Barber (1990). Finally, it was speculated that the SAL/TMT condition might show the lowest amount of stereotypy because these rats were not given amphetamine and, based on the results of Experiment 1, they would be expected to engage in freezing behavior.

Method

Subjects

This study used 40 experimentally naive male rats, approximately 350 g, of Sprague-Dawley descent, bred in the Kenyon College Psychology Department laboratory. As was done in Experiment 1, the subjects were housed in pairs and then in separate cages. They were habituated and tested during the light phase of a 12-hr light/dark schedule.

Apparatus

Habituation and behavioral testing was done in two Plexiglas activity boxes (41 cm long X 41 cm wide X 31 cm high) lined with a series of photocells mounted at two levels. After the amphetamine injection, the subject's stereotypic reactions were continuously monitored throughout the test session using an Omnitech Activity Monitor interfaced with a PC. Taped to the ceiling lid in the center of the boxes was a half piece of a 9.0 cm in diameter Whatman #1 filter paper which was used to present the test odors. A sound generator produced a 75-dB SPL white noise that was used to prevent extraneous sounds from disturbing the subjects during habituation and testing, and a 5-W red lamp remained on throughout these sessions. The activity boxes used for testing the rats with the two test odors were located in separate rooms adjacent to a monitoring room that contained the Omnitech Digiscan Analyzer and the PC.

Standardization studies have been conducted using the measures recorded by Omnitech's Digiscan Analyzer (Creese & Iverson, 1973; Sanberg, Moran, Kubos, & Coyle, 1983). The photocell-activated analyzer defines two stereotypic variables: (a) the number of stereotypic responses, which corresponds to the number of times the monitor observes the rat breaking the same beam repeatedly, and (b) stereotypy time, which consists of the total number of seconds that such behavior is detected by the monitor. The stereotypy-inducing effects of amphetamine, as measured by the analyzer, correlate significantly with visual stereotypic ratings developed by Creese and Iverson (1973). The analyzer was also programmed to record the overall number of general movements by the rats to further examine the behavioral effects of amphetamine and TMT.

Drug

A solution of 1 mg of d-Amphetamine in a volume of 1 ml/kg of the subject's weight was administered IP. An equal volume of saline was injected IP as the control condition.

Odors

Prior to observational testing, 100 [micro]l of CIT or TMT was applied to a piece of Whatman filter paper that was taped to the lid of the activity boxes.

Procedure

All subjects were handled for 3 days and on the next day they were habituated for one session in the activity boxes for a 60-min period. During habituation, the red light was on, the white background noise was presented, but the filter paper that was used to apply the test odors was not attached to the lids of the boxes.

On the following test day, 10 rats were randomly assigned to four conditions and tested for four successive 15-min intervals in the activity boxes. Thus, this experiment involved a 2 (odor) X 2 (drug) X 4 (interval) factorial design, with two independent-groups variables and one repeated-measures variable, respectively. Two groups of rats were administered saline and tested in the activity boxes with citronella (SAL/CIT) or TMT (SAL/TMT), and the two remaining groups were administered amphetamine before testing with citronella (AMP/CIT) or TMT (AMP/TMT). A 20-min interval, between the IP injection of the drug or saline and the start of testing in the boxes, was given in order for the amphetamine to be absorbed and the subjects to recover from the injections (Williams & Barber, 1990). Immediately before placing the rats into the boxes, the previously described amount of CIT or TMT was applied to the filter paper and taped to the center of the lid of each box. The Digiscan activity monitors recorded the rat's behavior and printed the response output, every 15 min, throughout the 60-min test session. After each subject was tested, the scented filter paper was removed and the boxes were thoroughly cleaned with a 5% ammonium hydroxide solution, as described in Experiment 1.

Results and Discussion

Figure 3 presents the mean number of stereotypic responses for each group over the four successive test intervals. The number of stereotypy scores was subjected to a 2 X 2 X 4 mixed-design analysis of variance (ANOVA) which revealed statistically significant between-groups effects of drug, F(1, 36) = 288.85, p < .001; odor, F(1, 36) = 55.31, p < .001; and a two-way interaction of drug by odor, F(1, 36) = 144.95, p < .001. Significant within-subjects effects were found for test interval, F(3, 108) = 52.62, p < .001; and the two-way interactions of drug by interval, F(3, 108) = 10.59, p < .001; and odor by interval, F(3, 108) = 6.04, p < .01, but the three-way interaction was not significant. Newman-Keuls post hoc tests indicated that Group AMP/TMT showed significantly (ps < .01) more stereotypic responses, over the four test intervals, than the three other groups. This finding, in conjunction with the drug by odor interaction, indicates that exposure to TMT potentiated the rats' stereotypy to a small injection of amphetamine. Post hoc tests, examining group differences at the various intervals, showed that Group SAL/TMT made significantly (ps < .05) fewer stereotypic responses during Interval 1 and Interval 2 than the other three groups. Group AMP/CIT showed significantly (ps < .05) more stereotypy during Interval 3 and Interval 4 than either of the two saline groups (i.e., Groups SAL/CIT and SAL/TMT), which reflected the sustained effect of the stimulant. Finally, significant (ps < .05) decreases in stereotypy were found between Interval 1, Interval 2, and Interval 4 for Group SAL/CIT, and between Interval 1 and the remaining intervals for Group SAL/TMT.

[FIGURE 3 OMITTED]

The amount of time that the rats showed stereotypy was analyzed by means of the same type of mixed-design analysis of variance as was used for the number of stereotypic responses. The significant outcomes paralleled, and were just as robust as, those found for the number of responses. Within-groups Pearson correlations between the two measures were found to be highly significant for all groups (rs (8) > .90; ps < .01). Because of the redundancy between the major findings for the number and time measures of stereotypy, it was thought that the reporting of the stereotypy time results would not be very informative. In addition, the results of an ANOVA and a series of post hoc tests on the movement data revealed that both the amphetamine groups (i.e., Groups AMP/CIT and AMP/TMT) showed significantly (ps < .01) more movement than the other groups during the final three test intervals, thus confirming the effect of amphetamine as a stimulant. In contrast, the effects of saline and TMT (i.e., Group SAL/TMT) resulted in significantly (p < .01) less movement than the other groups during the first two intervals, which is in agreement with previous reports of TMT decreasing locomotion and exploration (e.g., Vernet-Maury, 1980).

The results of the stereotypy data of this experiment are unique in demonstrating that the predator odor of TMT augmented both the number and duration of amphetamine-induced stereotypy responses. These findings are consistent with previous research showing that exposure to various stressors, such as electric shock, exacerbates amphetamine reactions (e.g., Anisman et al., 1985). They also represent an important addition to the research by Williams and Barber (1990), which indicated that prior exposure to a cat and later tests with cat odor sensitized both the frequency and duration of stereotypy, using the same dose (1 mg) of amphetamine and similar monitoring procedures. Finally, the significant within-group correlations between the number and time measures of stereotypy, reported earlier, were similar to those found by Williams and Barber (1990) using cat and cat odor as predator stressors.

Experiment 3

The findings of Experiments 1 and 2 indicate that exposure to the odor of the synthetic compound TMT produces a number of significant reactions in rats: opioid-mediated analgesia, defensive freezing, and exaggerated stereotypic responses to amphetamine. As discussed, these types of stress-induced reactions have been reported in rats using a variety of stressors, ranging from nociceptive stimuli to harmless exposure to a natural predator. Furthermore, a wide variety of chronic and acute stressors have also been found to influence the performance of rats on more complex tasks involving learning and memory. For example, long-term physical restraint has been shown to cause spatial memory impairments in rats using maze-learning tasks (Luine, Villegas, Martinez, & McEwen, 1994; Meaney, Aitken, Bhatnager, van Berkel, & Sapolsky, 1988). Moreover, a single session of exposure to inescapable shock was found to interfere with many types of associative learning tasks, such as shuttlebox escape learning (e.g., Maier, 1984), appetitive instrumental learning (e.g., Rosellini, 1978), and position escape learning in a Y maze (Jackson et al., 1980). More recently, Williams et al. (1998) reported that an acute ethological stressor of exposing rats, in a protective cage, to a cat significantly disrupted their asymptotic-level performance on a spatial delayed-alternation task used to assess working memory.

Experiment 3 incorporated some of the methodology used by Williams et al. (1998) to determine if exposure to TMT might effectively interfere with spatial working memory (i.e., SWM) on a delayed-alternation task using water as a reinforcer. Another objective of this research was to examine if low versus high amounts of TMT, relative to equivalent amounts of the novel odor of citronella (CIT), would have differential effects on the subsequent delayed-alteration performance. During training and a post-odor-exposure test session, subjects were reinforced for selecting the arm of an elevated T maze that they did not enter on the previous trial. Each training and testing session consisted of 30 trials: 15 involving virtually no delay (0-s delay) and 15 involving a 30-s delay between successive trials. It was hypothesized that exposure to a low amount of TMT would disrupt delayed-alternation performance on trials preceded by a 30-s delay, as opposed to a 0-s delay. Finally, exposure to a high amount of TMT was predicted to cause greater interference with test performance for both intervals, indicating that these decrements were not specifically related to a deficit in memory.

Method

Subjects

Thirty-six experimentally naive Sprague-Dawley male rats, approximately 360 g in weight, bred in the Kenyon College laboratory facilities, were used as subjects. The rats were maintained in individual cages on a regular light-dark cycle (12:12 hr) and were run during the light phase. They had free access to food, but were water restricted so that water could be used as a reinforcer. In order to prepare the rats for their subsequent restriction regime, they were given 2 hr of water per day for 7 days before shaping sessions. Throughout the remainder of the experiment, the rats were given 5-10 min of water after each daily session, which resulted in each rat maintaining a weight during training that was approximately 85% of its pretraining weight (Williams et al., 1998). All rats were housed in the same room during the training phase of the experiment. During the subsequent odor treatment phase, the rats in the two experimental TMT groups and the CIT control group were housed in individual cages located in separate rooms.

Apparatus

Two identical elevated wooden T mazes with two arms (60 cm X 10 cm) at the end of a runway (60 cm X 10 cm) were used to assess working memory. The maze walls were low, 3.5 cm high, to facilitate spatial learning of the room cues. Each elevated maze was supported by four 75-cm long table legs, and all the surfaces of the mazes were painted with a neutral gray enamel. On the top of the walls of the runway, the start position was indicated by a thin piece of black tape located 15 cm from the start of the runway. A gray wooden platform 10 cm X 10 cm X 1 cm with a recessed plastic cup was placed at the end of each arm so the water was not visible by the rat from the choice point. The mazes were located in separate small rooms, which contained an identical set of extra-maze cues. The rooms were illuminated by a 60-W white lamp placed to the right of the maze on top of a speaker used to present white noise of 75 decibels SPL to mask extraneous noise.

At the start of every training session the rats were individually placed in large boxes (41 cm long X 41 cm wide X 31 cm high) constructed of Plexiglas with removable lids. These boxes, which were used during the odor-exposure sessions, were located in isolated rooms illuminated with a 5-W red lamp. A white-noise generator was again used to mask extraneous sounds.

Procedure

After several days of handling, the subjects were randomly assigned to three independent groups. All the rats underwent equivalent handling, shaping, and testing procedures using one of the elevated T mazes.

Shaping. During the shaping phase, two drops of water were placed in the cup at the end of each arm of the maze. Each rat was placed in front of one of the cups, randomly determined, and allowed to drink the water. Over successive runs, the rat was started farther away from the cup after drinking the water and gradually the amount of reinforcement was reduced to one drop. Each rat eventually ran consistently from the start of the runway to the cup and drank the water. By means of a movable gray-wall barrier, the rat was forced randomly to go to either the right or left arm to receive water on consecutive shaping runs. Shaping sessions were given until each rat completed 30 trials of running from the start position of the maze to the water cup within a 25-min period, with an intertrial interval of 10-20 s.

Training. Prior to every delayed-alternation training session, the rat was individually placed in one of the Plexiglas odor boxes for a 30-min period, without attaching the filter paper that was later to be used to provide odors during testing. Each training session was initiated by giving the rat one forced trial, followed by 30 choice trials. On a forced trial, one drop of water was randomly placed in either the right or left cup, and the movable barrier was inserted at the choice point so that it forced the rat to go down the arm with the water. On subsequent choice trials, a drop of water was placed in the cup of the arm opposite to the one that the rat entered on the previous trial. After each trial, the rat was either returned to the vicinity of its home cage, but not placed inside it (i.e, virtually a 0-s delay), or it was placed in its home cage for a 30-s interval (30-s delay) before the start of the next trial. During training sessions, the home cage was placed inside a white plastic tub cage so that the rat could not see the maze or many of the room cues during the delay interval. Each training session consisted of a total of 30 choice trials: 15 trials with a 0-delay interval and 15 trials with a 30-s delay interval prior to the next trial. The 0-s and 30-s delay trials were randomized throughout the training session. All animals underwent the same sequence of 0-s and 30-s delay trials on a particular training session, but a new sequence was used for each day of training. As mentioned, a correct response was recorded when the rat went to the arm opposite from the one chosen on the previous trial, and it did not gain access to the water after an error was made. The mazes were rapidly cleaned with a diluted solution of alcohol and water (1:3) after each trial to control for the rat's odor, and they were thoroughly cleaned with a 15% solution of ammonium hydroxide at the end of each session (see Experiments 1 and 2 for more details).

At the conclusion of each daily training session, the percentage of choice accuracy for each subject was calculated separately for the two delay conditions. Specifically, this measure was obtained by dividing the number of correct responses for a given delay interval by 15 and multiplying this value by 100. A training criterion of 80% correct choice accuracy on the 30-s delay trials was used to ensure that subjects had learned the delayed-alternation task to an equivalent level before testing. On Day 13 of training, 10 subjects in each group achieved this criterion and only these rats were tested during the remaining phases of the experiment. During all sessions, records were also made of the total amount of time that each rat spent in the maze, from the start of the first trial to completion of the last trial.

Odor exposure and testing. On Day 14, all rats received the same procedure that was used during training except that, depending on the group, a specific amount of citronella (CIT) or TMT was applied to a piece of Whatman filter paper that was taped to the lid of the odor-exposure box. All subjects were individually exposed to the odors in the box for a period of 30 min. One group of rats (Group CIT-CONT) was exposed to 132 [micro]l citronellal that was reapplied after every 4 subjects. For the second group (Group TMT-LOW), the same quantity (132 [micro]l) of TMT was applied to filter paper and it was also replenished after every 4 subjects. The third group (Group TMT-HIGH) was exposed to 264 [micro]l of TMT which was reapplied after every 2 subjects.

Following the 30-min period of odor exposure, subjects were tested in the T maze using the identical delayed-alternation procedure as the one used during training. Each of the three groups was exposed to the test odor and tested in the maze on separate days in the following order: Group CIT-CONT, Group TMT-LOW, and Group TMT-HIGH. This order was used to prevent the possibility that the odor of TMT might be present during subsequent tests with the control odor of CIT. In addition, the odor boxes and T mazes were cleaned with a 15% solution of ammonium hydroxide at the end of each session to eliminate the presence of residual test odors and conspecific odors from previously tested rats.

Results and Discussion

The mean pretreatment acquisition scores of the three groups revealed that their percentage of choice accuracy was near chance on Day 1, ranging from 48% to 59% for both delay intervals. The results of a repeated-measures ANOVA and a series of post hoc tests, examining the effects of groups, delay intervals, and sessions, indicated that all three groups showed comparable and significant (ps < .001) increases in accuracy, for each of the delay intervals, over the successive training sessions. The results of an ANOVA and post hoc tests confirmed that the mean performance levels of the rats which met the training criterion for three groups were virtually the same (ps > .10).

More importantly, a comparison was made of the percentage of choice accuracy of each subject on the last session of training with that subject's choice accuracy on the test session. Specifically, for each of the two delay conditions, a given rat's mean percentage of correct accuracy score during testing (Day 14) was subtracted from that rat's mean percentage of accuracy score on the last day of training (Day 13). These difference scores were used as a dependent measure to assess the size of the odor disruption effect for the 0-s and 30-s delay trials during testing.

Figure 4 presents the mean percentage values of the odor disruption effect for the three odor groups during the two delay conditions. These scores were examined by a 3 X 2 mixed-design analysis of variance (ANOVA) which showed a statistically significant between-groups effect of odor, F(2, 27) = 10.46, p < .01, a significant within-groups effect of the duration of delay, F(1, 27) = 5.98, p < .05, and a significant interaction between odor and delay, F(2, 27) = 8.11, p < .01. The results of Newman-Keuls post hoc tests indicated that significantly (p < .01) more odor disruption was found for the 30-s delay trials, as compared to the 0-s delay trials, for Group TMT-LOW. However, no significant differences were found between the delay conditions for the other two odor groups. The disruption on each of the delay conditions for Group TMT-HIGH was significantly (p < .01) greater than the corresponding delay trials for Group CIT-CONT and during the 0-s delay trials for Group TMT-LOW. In addition, a series of directional single-sample t tests showed significant levels of disruption above 0% (i.e., preexposure performance) for only three conditions: Group TMT-LOW during the 30-s trials, t(9) = 2.24, p < .05; and Group TMT-HIGH during the 0-s trials, t(9) = 4.60, p < .01; and during the 30-s trials, t(9) = 6.18, p < .01.

[FIGURE 4 OMITTED]

Although stress has been previously shown by many investigators to result in impairments in memory, this is the first study to demonstrate that the odor of TMT produces a deficit in spatial working memory. Recently, Morrow, Roth, et al. (2000) claimed that TMT interfered with working memory because their TMT-exposed rats equally explored a familiar versus a novel visual object, whereas their nonstressed rats explored the novel object more. This reduction in exploration of a novel stimulus could be because of a neophobic reaction following exposure to TMT and not because of a disruption in memory. In addition, the findings from the present experiment represent an important extension of our previous research showing that cat odor also impaired delayed-alternation performance using the same apparatus and general design as the current study (Williams et al., 1998). A second new finding of this experiment was that a larger amount of TMT was found to impair alternation performance during testing on both the 0-s and 30-s delay trials. This result suggests that a larger amount of TMT leads to a greater fear reaction and an interference with general performance.

Although freezing was frequently seen during the periods of exposure to TMT, as opposed to citronella, virtually no instances of freezing were observed during the test session in the T maze. The results of an additional analyses revealed that the TMT and CIT groups did not differ significantly in the total time required for subjects to finish their 30 test trials. In addition, within-groups correlational analyses failed to find a significant relationship between each subject's total testing time and choice performance during testing, which was consistent with the results from our cat-odor study of memory (Williams et al., 1998). The above findings suggest that the working memory disruptions observed for the TMT groups in this study can not be attributed to slower running times and therefore longer intertrial intervals.

Experiment 4

A number of researchers have reported that stress levels, often along with changes in corticosterone, impaired the ability of rats to process complex spatial information as measured in the Morris water maze (Conrad, Galea, Kuroda, & McEwen, 1996; Luine et al., 1994; Lupien & McEwen, 1997; Sandi, Loscertales, & Guaza, 1997). This maze is particularly useful to examine spatial learning because it requires that the rat learn to swim to a specific location using visual cues present in the room (Morris, 1984). The Morris maze is effective not only for testing the acquisition of spatial information, but also for determining the ability of the subject to recall previously learned spatial information. Studies that have focused on spatial working memory (SWM) have typically varied the location of the platform throughout the trials. In addition, investigators (e.g., de Quervain, Roozendaal, & McGaugh, 1998) have examined spatial reference memory (SRM) in rats using a procedure in which the escape platform was in a fixed location during both training and subsequent test sessions.

Benzodiazepines have long been documented to reduce fear and anxiety reactions in response to stress (Westerbrook, Greeley, Nabke, & Swinbourne, 1991). Fanselow and Helmstetter (1988) found that midazolam and diazepam in foot-shock-formalin tests with rats were capable of reducing conditioned analgesia and the level of freezing behavior. In addition, Hotsenpiller and Williams (1996) showed that diazepam successfully blocked the conditioned fear in rats caused by sessions of being defeated by a dominant male rat. More relevant to the present experiment, Dielenberg et al. (1999) found that the avoidance response of hiding in a box by rats, exposed to cat odors, was reversed when these subjects were a given low-dose (0.375 mg/kg) of the benzodiazepine agonist, midazolam (MDZ). MDZ was also revealed to have a rapid onset time, a brief duration of action, and did not produce motoric side effects (Dielenberg et al., 1999). Although some benzodiazepines have been shown to have anterograde effects on memory (e.g., Cahill, Brioni, & Izquierdo, 1986; McNaughton & Morris, 1987), retrograde amnesia was not found using a low dose of MDZ which effectively reduced anxiety to cat odors (Dielenberg et al., 1999).

Experiment 4 was designed to address two questions: (a) if prior exposure to TMT would interfere with the recall of SRM during continued testing in a Morris water maze, and (b) if the administration of a low dose of MDZ prior to TMT-exposure affected later spatial recall in the maze. Separate groups of rats were trained to find the submerged platform in the water maze and were injected with MDZ or saline prior to exposure to TMT or the neutral odor citronella (CIT). Following odor exposure, the subjects were tested in the maze and, as during training, their latency to reach the hidden platform was measured. The SAL/CIT control group was used to assess possible stress effects of the injection and of exposure to the novel control odor of citronella. The MDZ/CIT group was run to check that midazolam, per se, did not lead to an increase in latencies due to possible amnesic or motoric effects. Finally, the SAL/TMT group was designed to show the hypothesized stress effects of TMT on spatial recall, and the MDZ/TMT group was tested to determine if midazolam was able to reduce stress during TMT exposure and prevent the predicted impairment in recall of the location of the platform.

Method

Subjects

Thirty-two, experimentally naive Sprague-Dawley male rats, approximately 370 g in weight, bred in Kenyon College laboratory facilities were used as subjects. The rats were housed individually in tub cages in a large room isolated from a smaller room containing the water maze and from two separate cubicles in which Plexiglas odor boxes were located. The subjects were kept on a 12:12-hr light/dark cycle and training and testing sessions were run during the light phase. The rats were provided with free access to food and water throughout the experiment.

Apparatus

The Morris water maze was a black metal circular tank, with a diameter of 1.7 m, filled to a depth of 18 cm with water at room temperature (62-65 [degrees]F). The circular tank was divided into four equal quadrants. In the middle of one of the quadrants, a black escape platform 11 cm in diameter was located 2 cm below the surface of the water. For a given subject, this hidden platform was located in the same quadrant of the maze throughout the experiment, and the rat was released into the maze from a different quadrant on each trial. To facilitate learning and memory the various extra-maze cues and their position within the room, as well as the experimenter, remained constant during the training and test sessions. The maze room was illuminated by a 60-W white lamp placed to the right of the maze on top of a speaker from which white noise of 75 decibels SPL was emitted to mask extraneous noise.

Before each training and the testing session, the rats were individually placed in large boxes (41 cm long X 41 cm wide X 31 cm high) made of Plexiglas with removable lids. As was the case in Experiment 3, the boxes were used to provide odors during later testing. They were located in isolated rooms illuminated with a 5-W red lamp and sounds were masked by the previously described white noise.

Drug

Midazolam HCI (from Sigma-Aldrich Co.) was injected SC behind the neck at a dose of 0.375 mg/kg for the experimental groups, and an equivalent volume of saline was used for the control subjects. Dielenberg, Arnold, et al., (1999) reported 0.375 mg/kg as being an optimal dose of MDZ, with lower doses (0.2 mg/kg) having little effect on cat odor avoidance and higher doses (0.5 mg/kg) resulting in clear signs of sedation.

Odors

During the odor-exposure phase of this experiment, subjects were exposed to the test odor by placing 120 [micro]l of CIT or TMT, with a microsyringe, on a piece of Whatman filter paper suspended from the lid of the odor box.

Procedure

All of the rats were handled for a short time each day for a week prior to training. Training sessions occurred on consecutive days until the subjects were proficient at the water maze task. At the start of each session, all rats received a SC injection of saline to control for the stress of an injection. Fifteen min later, the rats were transported to an isolated room and placed in the Plexiglas odor box, without the addition of any test odor, for a 30-min habituation session.

Each training session in the maze consisted of eight trials, two trials beginning from each of the four starting points, at the marked quadrants, along the wall of the circular maze. Subjects were given a total of 120 sec to locate the platform on each trial. If they did not do so in this time period, they were guided to the platform by the experimenter. All rats were allowed an intertrial interval of 30 sec after they were on the platform. The time, in sec, for the rat to find the platform on each trial was obtained using an electronic stopwatch and recorded by the experimenter. After the last trial of a session, the rats were towel-dried and returned to their home cages.

Consistent with the findings of Vunderwolf and Cain (1994), all of the rats showed rapid learning of the location of the platform and their performance was stable by Day 3 of training. The subjects were then randomly assigned to four groups of 8 rats each and given one of the following procedures: a saline injection followed by an exposure session using either citronella (SAL/CIT) or TMT (SAL/TMT) or a midazolam injection followed by an odor session of citronella (MDZ/CIT) or TMT (MDZ/TMT). As was done during training, the rats were returned to their home cages for a 15-min period after the injection. This interval was given to allow time for the MDZ to have an anxiolytic effect (Dielenberg et al., 1999). The rats were then placed inside one of the Plexiglass odor boxes where they were exposed for 30 min to either the samples of CIT or TMT applied to Whatman filter paper that was suspended from the lid of each box. This duration of odor exposure was the same as that used in the Experiment 3 and in a previous memory study involving cat odor (Williams et al., 1998).

Following odor exposure, the rats were returned to their home cages for an additional 30-min period in order for the low dose of MDZ to no longer have an anxiolytic effect (Dielenberg et al., 1999). The rats were then tested in the water maze with the same procedure as one used during training, and the amount of time each rat required to locate the platform was recorded after each trial. The odor boxes were throughly cleaned after the exposure session (see Experiments 1-3 for details), and both the CIT groups were trained and tested prior to the TMT groups to ensure the isolation of the odor conditions.

Results and Discussion

Figure 5 presents the mean latency scores to reach the hidden platform in the water maze for the four groups on the final training session (Day 3) and during the post-odor test session (Day 4). These data were analyzed by a 2 (drug) X 2 (odor) X 2 (session) mixed-design ANOVA, with the within-subjects variable being the final training versus the test session. The results of this analysis indicated that the main effects both of the between-groups variables were statistically significant: drug, F(1, 28) = 4.29, p < .05; and odor, F(1, 28) = 6.29, p < .05. In addition, the two-way interaction of drug by odor, F(1, 28) = 4.38, p < .05, and three-way interaction of drug by odor by session, F(1, 28) = 6.75, p < .05, were found to be significant. Newman-Keuls post hoc tests indicated that the mean latency for Group SAL/TMT was significantly (ps < .01) longer during testing. Additional post hoc tests revealed that none of the other groups showed statistically significant differences in their latencies between training and testing, nor were there any significant differences between these groups at the .05 level. In fact, with the exception of Group SAL/TMT, all of the other groups showed nonsignificant decreases in their latencies from training to the test session. This suggests that, even though all of the groups had similar latencies on the final session of training, the rats may not have quite reached their asymptotic level of acquisition on Day 3.

[FIGURE 5 OMITTED]

These data demonstrate that prior exposure to TMT produced a deficit in spatial reference memory of the location of the maze platform. This represents an important extension of the results of Experiment 3, which showed that TMT disrupted working memory in rats trained and tested on a delayed-alternation task. This finding is also somewhat consistent with a recent study by Diamond et al. (1999) in which exposure to a cat interfered with the spatial memory of rats tested in a radial arm water maze. Another unique finding of this experiment was that injecting rats with midazolam, before exposing them to TMT (i.e., Group MDZ/TMT) was found to counteract the deficit in recall observed for Group SAL/TMT. This result illustrates that this benzodiazepine agonist was able to protect rats from showing an interference in reference memory caused by the stress of TMT exposure. Finally, the similar levels of test performance found for the two control groups (i.e., Groups SAL/CIT and MDZ/CIT) confirm that the small dose of MDZ used in this study failed to have amnesic or motoric effects (Dielenberg et al., 1999).

General Discussion

The results of Experiment 1 revealed that rats exposed to TMT, after receiving a small injection of painful formalin in their rear paw, engaged in less recuperative paw licking than subjects tested with the control odor of citronella, which is known to be an hedonically neutral odor for rats (e.g., Lester & Fanselow, 1985; Williams, 1999; Williams, Worland, et al., 1990). Suppression in recuperative paw licking has been interpreted by numerous researchers to be indicative of a stress-induced analgesic reaction. Furthermore, it was found that this analgesia was controlled by an endogenous opioid system because it was blocked when rats were given an injection of an opiate antagonist prior to the formalin test (Fanselow & Baackes, 1982). The stress-induced, opioid analgesia that we observed with TMT has also been reported in studies using the following stressors: conditioned stimuli paired with foot shock (Fanselow, 1984; Fanselow & Baackes, 1982; Helmstetter & Fanselow, 1987), unconditioned odors of previously shocked conspecifics (Williams, 1987), conditioned odors of intruder rats that were attacked and defeated during a single resident-intruder session (Williams & Scott, 1989; Williams, Worland, et al., 1990), and exposure to the unconditioned fear of cat odor (Lester & Fanselow, 1985). Finally, the present results of TMT-elicited analgesia support our previous study which revealed that TMT was not only an effective conditioned stimulus, when paired with an injection of an anxiogenic, but it also elicited a moderate level of freezing and analgesia in the control condition in which the anxiogenic was not given as an unconditioned stimulus (Hotsenpiller & Williams, 1997).

During the time that the rats were exposed to TMT, in Experiment 1, they were observed to display significant durations of freezing. This is consistent with the findings of Wallace and Rosen (2000), who reported that rats predominately engaged in freezing when tested with TMT in a relatively small chamber, but not in an open field or a long alley. Furthermore, these investigators noted that the use of larger amounts of TMT resulted in more freezing, which also may be why more freezing was found in the present research than in some of the previously reported TMT studies. However, the levels of freezing were not high enough (i.e., about 42% of the total time of testing) to have prevented the rats from engaging in recuperative licking or other responses. In fact, the TMT-exposed subjects were observed to spend considerable time displaying a form of risk assessment behavior that consisted of the rats staring at the suspended TMT filter, swaying their heads from side to side, while assuming a semi-rearing posture. This behavior differed from the approach-avoidance, flat-back type of risk assessment pattern, reported by the Blanchards (e.g., Blanchard, R. J., & Blanchard, 1989), for rats tested with cat hair on a cloth located at the end of a straight-alley maze with a relatively low ceiling. The odor boxes, used in Experiment 1 and most of the remaining experiments, were much higher and rearing may have enabled the rat to more effectively explore the filter paper scented with TMT. This represents a good example of how the size and configuration of the exposure environment are critical in determining the precise nature, or typography, of the observed pattern of defensive behavior.

A number of researchers have claimed that TMT reliably produces response immobility and less exploration in open-field and hide-box testing situations, as opposed to species-typical freezing and risk assessment (Dielenberg & McGregor, 2001; McGregor et al., 2002; Morrow, Redmond, et al., 2000). Even though some of these researchers favor the use of the cat-odor paradigm, rather than TMT to study predator fear, many of these investigators reported that TMT elicits a host of neurophysiological alterations that have also been observed with other stressors, such as foot shock, conditioned fear stimuli, and exposure to dominant conspecifics or predators. Some of the physiological reactions elicited by TMT include: increments in serum levels of corticosterone (Dielenberg & McGregor, 2001; Vernet-Maury, 1980; Vernet-Maury et al., 1984), more rapid dopamine metabolism in the prefrontal cortex (Morrow, Redmond, et al., 2000), and the activation of the expression of an immediate-early c-fos gene that is involved in memory changes (Redmond et al., 2002).

One of the main reasons for the variability in the findings of studies on predator odors is because researchers have used different odor presentation techniques and amounts of odors. A second critical factor, which was just mentioned in terms of risk assessment behavior, is that the size of the odor-exposure areas has varied widely, thus affecting ventilation and odor loss. A third factor is that the observation of the rats' behavior is sometimes done when rats are actually experiencing the odor of TMT, whereas other studies involve tests occurring after TMT exposure. Because all of these variables affect the salience of the odor and the animal's perception of the imminence of a possible predator, it is not surprising that they would influence whether or not certain responses occur. In terms of Experiment 1, and most of the other experiments, testing involved a fairly large amount of TMT that was in close proximity to the rat. Under these conditions, it is speculated that TMT would not simply elicit anxiety-motivated behaviors, but result in more intense fear-motivated reactions, such as freezing and opioid analgesia. Finally, differences in the defensive responses of rats to TMT, as opposed to cat odor, may be because the red fox is larger and perhaps a more threatening predator than a cat and it engages in a different pattern of species-typical predatory behavior.

The data from Experiment 2 clearly revealed that rats, injected with a small dose of d-amphetamine, showed augmentations in stereotypic movements when exposed to TMT, but not the neutral control odor. Stress-sensitization effects to a small dosage of amphetamine were noted previously using repeated presentations of inescapable foot shock (Anisman et al., 1985). Our prior research also found an exaggeration in amphetamine-induced stereotypy in rats exposed to a cat and tested in the presence of soiled cat litter (Williams & Barber, 1990). It is interesting that the data from the activity monitors, in Experiment 2, did not indicate that TMT produced long periods of immobility. In fact, if immobility increased with TMT exposure, it might have precluded the augmentations in the frequency and durations of stereotypy. Unfortunately, our automated monitors were not capable of reliably recording freezing behavior. However, many researchers claim that freezing behavior is very different from response immobilization (Fanselow, 1984; Fanselow et al., 1987). Finally, the TMT-induced sensitization of amphetamine reactions is not surprising in light of researchers who have reported that TMT elevates levels of corticosterone (Dielenberg & McGregor, 2001; Vernet-Maury, 1980; Vernet-Maury et al., 1984) and increases dopamine activity (Morrow, Redmond, et al., 2000). More psychopharmacological studies are needed to examine if the TMT-produced alterations in reaction to amphetamine, and possibly other stimulants, are due to dopaminergic changes and possible interactions with other neurotransmitter mechanisms, such as those affecting norepinephrine activity.

The results of Experiment 3 revealed that TMT exposure has a subsequent disruptive effect on rats performing a previously learned delayed-alternation task in an elevated T maze, as a test of working memory. TMT-induced difficulty in selecting the arm (i.e., spatial location), not chosen on the previous trial, was not found for control rats exposed to the novel odor of citronella. The subjects in Experiment 3 were also found not to be impaired after TMT exposure when no delay interval was imposed between trials during maze testing. This finding is consistent with our previous research demonstrating cat exposure altered performance of a 20-s, delayed-alternation task (Williams et al., 1998). In addition, exposing rats to a larger amount of TMT was found to result in greater deficits in response alternation following both the 30-s delay trials and the no-delay trials, suggesting that a high level of TMT influences performance factors that do not involve memory. Recently, Morrow, Roth, et al. (2000) have also claimed that TMT affects working memory, but this study did not include a control odor and rats that were exposed to TMT may have failed to explore a novel versus a familiar visual object because of fear following exposure to TMT (i.e., neophobia), rather than a memory impairment.

The results of Experiment 4 support and expand on those of Experiment 3 by demonstrating that TMT interfered with recall of the fixed location of an escape platform in a Morris water maze. This impairment presumably reflected a deficit in spatial reference memory because the subjects showed good performance on this task prior to stress exposure. Thus, our findings indicate that TMT, as a predator stressor, is capable of altering both spatial working memory as well as spatial reference memory. Furthermore, the data from Experiment 4 suggest that a moderate dose of the benzodiazepine agonist, midazolam, is effective in eliminating TMT-induced memory deficits without any evidence of retrograde amnesic or motoric effects. Future studies are needed to establish the duration and the possible causes of the reported TMT-induced interferences in memory. On the basis of our past research, we suspect that these deficits might be for a relatively short duration because disruptions in memory were found to recover, within 24 hr after cat exposure, when rats were subsequently given reinforced trials on a spatial delayed-alternation task (Williams et al., 1998). The short duration of these predator-elicited lapses in memory are probably not caused by glucocorticoid changes in the brain that have been noted with repeated or more chronic exposure to stress (Luine et al., 1994; Luine, Spencer, & McEwen, 1993). However, some researchers have observed a more transient form of memory loss, which involves precursors to corticosterone and glucocorticoids changes, that occurs following an acute exposure to stress (e.g., McEwen & Sapolsky, 1995; Sapolsky, 1992) or opioid modulation (Wan, Givens, & Olton, 1995).

Another possible mechanism responsible for the memory disruptions caused by TMT has been presented by Morrow and his colleagues (Morrow, Elsworth, & Roth, 2002; Morrow, Roth, et al., 2000). According to these investigators, TMT activation of DA metabolism in mPFC results in an enhanced reaction and duration of fear, accompanied by reduced activity in an open field. The enhancement of this fear response is considered to cause the rat to focus its attention on the possibility of aversive stimuli in its environment and to reduce the processing of information not related to a potential threat. Consistent with this interpretation, it is speculated that prior exposure to TMT resulted in our rats not attending to environmental cues critical for making the correct instrumental response for appetitive reinforcement (i.e., water reward in Experiment 3) or escape reinforcement (i.e., the hidden platform in Experiment 4).

The previously described attentional-disruption explanation is particularly appealing in view of the data from a recent pilot study that was conducted in our lab (Richardson & Williams, 2003). In this study, rats were given extensive training in an eight-arm radial maze to go to four specific arms for a piece of chocolate-coated cereal as a reinforcer. The arms of the maze were rotated after each training session so that the rats would have to rely on remembering the room cues. After the rats reached an asymptotic level of learning, they were exposed to either TMT or citronella. As expected, during the session of TMT exposure all of the rats showed considerable freezing, whereas they did not engage in freezing during subsequent recall testing in the maze. However, the rats exposed to TMT had no problem in selecting the correct arms during testing and performed as well as the control group. Our current explanation of these negative findings is consistent with the theorizing of Morrow et al. (2002). We believe that the stressed rats failed to show disruptions in memory because the test environment had become very familiar to them after their extensive training and they experienced only mild food restriction. Thus, these subjects experienced virtually no stress or fear during the test session. Currently, we are examining if a TMT-induced interference in memory might occur if the test session is conducted in the same maze as the one used during training, but novel or mildly threatening stimuli are also present in the test environment. These findings would support the interpretation that TMT disrupts memory, and perhaps other cognitive processes, because subjects are attending to threatening stimuli as opposed to the task-relevant spatial cues.

This research was partially supported by endowment funds from the Samuel B. Cummings, Jr. Professorship in Psychology awarded to the senior author and also by stipends from the Kenyon College Summer Science Program. Correspondence concerning this article should be addressed to Jon L. Williams, Department of Psychology, 115 Samuel Mather Building, Kenyon College, Gambier, OH 43022. (E-mail: [e-mail]).

References

ANISMAN, H., HAHN, B., HOFFMAN, D., & ZACHARKO, R. M. (1985). Stressor invoked exacerbation of amphetamine-elicited perseveration. Pharmacology, Biochemistry, & Behavior, 23, 173-185.

BLANCHARD, D. C., & BLANCHARD, R. J. (1988). Ethobiological approaches to the biology of emotion. Annual Reviews of Psychology, 39, 43-68.

BLANCHARD, R. J., & BLANCHARD, D. C. (1989). Attack and defensive behaviors in rodents as ethoexperimental models for the study of emotion. Progress in Neuropsycho-pharmacological & Biological Psychiatry, 13, 3-14.

BLANCHARD, R. J., BLANCHARD, D. C., & HORI, K. (1989). An ethoexperimental approach to the study of defense. In R. J. Blanchard, P. F. Brain, D. C. Blanchard, & S. Parmigiani (Eds.), Ethoexperimental approaches to the study of behavior (pp. 114-136). Holland: North Holland Publishing Company.

BLANCHARD, R. J., FUKUNAGA, K. K., & BLANCHARD, D. C. (1976). Environmental control of defensive reactions to a cat. Bulletin of the Psychonomic Society, 8, 129-130.

BOLLES, R. C., & FANSELOW, M. S. (1980). A perceptual-defensive-recuperative model of fear and pain. Behavioral and Brain Sciences, 3, 291-323.

CAHILL, L., BRIONI, J., & IZQUIERDO, I. (1986). Retrograde memory enhancement by diazepam: Its relation to anterograde amnesia, and some clinical implications. Psychopharmacology, 90, 554-556.

CARLSON, N. R. (1994). Physiology of behavior (5th ed.). Boston: Allyn and Bacon.

CONRAD, C. D., GALEA, L. A. M., KURODA, Y., & MCEWEN, B. S. (1996). Chronic stress impairs spatial memory on the Y maze, and this effect is blocked by tianeptine pretreatment. Behavioral Neuroscience, 100, 1321-1334.

CREESE, I., & IVERSEN, S. D. (1973). Blockage of amphetamine-induced motor stimulation and stereotypy in the adult rat following neonatal treatment with 6-hydroxydopamine. Brain Research, 55, 368-382.

DE QUERVAIN, D. J., ROOZENDAAL, B., & MCGAUGH, J. C. (1998). Stress and glucocorticoids impair retrieval of long-term spatial memory. Nature, 394, 787-790.

DIAMOND, D. M., PARK, C. R., HEMAN, K. L., & ROSE, G. M. (1999). Exposing rats to a predator impairs spatial working memory in the radial arm water maze. Hippocampus, 9, 542-552.

DIELENBERG, R. A., ARNOLD, J. C., & MCGREGOR, I. S. (1999). Low dose midazolam attenuates predator odor avoidance in rats. Pharmacology, Biochemistry, & Behavior, 67, 197-201.

DIELENBERG, R. A., CARRIVE, P., & MCGREGOR, I. S. (2001). The cardiovascular and behavioral response to cat odor in rats: Unconditioned and conditioned effects. Brain Research, 897, 228-237.

DIELENBERG, R. A., HUNT, G. E., & MCGREGOR, I. S. (2001). 'When a rat smells a cat': The distribution of Fos immunoreactivity in rat brain following exposure to a predator odor. Neuroscience, 104, 1085-1097.

DIELENBERG, R. A., & MCGREGOR, I. S. (1999). Habituation of the hiding response to cat odor in rats (Rattus norvegicus). Journal of Comparative Psychology, 113, 376-387.

DIELENBERG, R. A., & MCGREGOR, I. S. (2001). Defensive behavior in rats towards predatory odors: A review. Neuroscience & Biobehavioral Reviews, 25, 597-609.

DRUGAN, R. C., GRAU, J. W., MAIER, S. F., MADDEN, J., & BARCHAS, J. D. (1981). Cross tolerance between morphine and longer-term analgesia reaction to inescapable shock. Pharmacology, Biochemistry, & Behavior, 14, 677-682.

FANSELOW, M. S. (1980). Conditioned and unconditioned components of post-shock freezing. Pavlovian Journal of Biological Sciences, 15, 177-182.

FANSELOW, M. S. (1984). Shock-induced analgesia on the formalin test: Effects of shock severity, naloxone, hypophysectomy, and associative variables. Behavioral Neuroscience, 98, 79-95.

FANSELOW, M. S., & BAACKES, M. P. (1982). Conditioned fear-induced opiate analgesia on the formalin test: Evidence for two aversive motivational systems. Learning & Motivation, 13, 200-221.

FANSELOW, M. S., & HELMSTETTER, F. J. (1988). Conditioned analgesia, defensive freezing, and benzodiazepines. Behavioral Neuroscience, 102, 233-243.

FANSELOW, M. S., SIGMUNDI, R. A., & WILLIAMS, J. L. (1987). Response selection and the hierarchical organization of species-specific defense reactions: The relationship between freezing, flight, and defensive burying. The Psychological Record, 37, 381-386.

FILE, S. E., ZANGROSSI, H., Jr., & ANDREWS, N. (1993). Novel environment and cat odor change GABA and 5-HT release and uptake in the rat. Pharmacology, Biochemistry, & Behavior, 45, 931-934.

FILE, S. E., ZANGROSSI, H., Jr., SANDERS, F. L., & MABBUTT, P. S. (1993). Dissociation between behavior and corticosterone responses on repeated exposures to cat odor. Physiology & Behavior, 54, 1109-1111.

HELMSTETTER, F. J., & FANSELOW, M. S. (1987). Effects of naltrexone on learning and performance of conditional fear-induced freezing and opioid analgesia. Physiology & Behavior, 39, 501-505.

HOTSENPILLER, G., & WILLIAMS, J. L. (1996). Conditioned anxiety and analgesia: Benzodiazepine and 5-H[T.sub.1A] agonists. Psychobiology, 24, 118-126.

HOTSENPILLER, G., & WILLIAMS, J. L. (1997). A synthetic predator odor (TMT) enhances conditioned analgesia and fear when paired with a benzodiazepine receptor inverse agonist (FG-7142). Psychobiology, 25, 83-88.

JACKSON, R. L., ALEXANDER, J. H., & MAIER, S. F. (1980). Learned helplessness, inactivity, and associative deficits: Effects of inescapable shock of response choice escape learning. Journal of Experimental Psychology: Animal Behavior Processes, 6, 1-20.

KOKKINIDIS, L., & ANISMAN, H. (1981). Amphetamine psychosis and schizophrenia: A dual model. Neuroscience & Biobehavior Review, 5, 449-461.

LESTER, L. S., & FANSELOW, M. S. (1985). Exposure to a cat produces opioid analgesia in rats. Behavioral Neuroscience, 99, 756-759.

LUINE, V., SPENCER, R., & MCEWEN, B. S. (1993). Effects of chronic corticosterone ingestion on spatial memory performance and hippocampal serotonergic function. Brain Research, 616, 65-70.

LUINE, V., VILLEGAS, M., MARTINEZ, C., & MCEWEN, B. S. (1994). Repeated stress causes reversible impairments of spatial memory performance. Brain Research, 639, 167-170.

LUPIEN, S. J., & MCEWEN, B. S. (1997). The acute effects of corticosteroids on cognition: Integration of animal and human model studies. Brain Research & Review, 24, 1-27.

MAIER, S. F. (1984). Learned helplessness and animal models of depression. Progress in Neuropsychopharmacological and Biological Psychiatry, 83, 435-446.

MAIER, S. F. (1986). Stressor controllability and stress-induced analgesia. In D. D. Kelley (Ed.), Annals of the New York Academy of Sciences, 467, 55-71.

MAIER, S. F. (1989). Determinants of the nature of environmentally induced hypoalgesia. Behavioral Neuroscience, 103, 131-143.

MAIER, S. F. (2003). Bidirectional immune-brain communication: Implications for understanding stress, pain, and cognition. Brain, Behavior, & Immunity, 17, 69-85.

MAIER, S. F., & SELIGMAN, M. E. P. (1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology: General, 105, 3-46.

MAIER, S. F., & WATKINS, L. R. (1998). Stressor controllability, anxiety, and serotonin. Cognitive Therapy Research, 6, 595-613.

MCEWEN, B. S., & SAPOLSKY, R. (1995). Stress and cognitive function. Recent Opinion in Neurobiology, 5, 205-216.

MCGREGOR, I. S., SCHRAMA, L., AMBERMOON, P., & DIELENBERG, R. A. (2002). Not all epredator odours' are equal: Cat odour but not 2, 4, 5 trimethylthiazoline (TMT: fox odour) elicits specific defensive behaviours in rats. Behavior Brain Research, 129, 1-16.

MCNAUGHTON, N., & MORRIS, R. G. M. (1987). Chlordiazepoxide, and anxiolytic benzodiazepine, impairs place navigation in rats. Behavior Brain Research, 24, 39-46.

MEANEY, M., AITKEN, D., BHATNAGER, S., VAN BERKEL, C., & SAPOLSKY, R. (1988). Effects of neonatal handling on age-related impairments associated with the hippocampus. Science, 239, 766-769.

MORRIS, R. G. M. (1984). Developments of a water-maze procedure for studying spatial learning in the rat. Journal of Neuroscience Methods, 11, 47-60.

MORROW, B. A., ELSWORTH, J. D., & ROTH, R. H. (2002). Fear-like biochemical and behavioral responses in rats to the predator odor, TMT, are dependent on the exposure environment. Synapse, 46, 11-18.

MORROW, B. A., REDMOND, A. J., ROTH, R. H., & ELSWORTH, J. D. (2000). The predator odor, TMT, displays a unique stress-like pattern of dopaminergic and endrocrinological activation in the rat. Brain Research, 864, 146-151.

MORROW, B. A., ROTH, R. H., & ELSWORTH, J. D. (2000). TMT, a predator odor, elevates mesoprefrontal dopamine metabolic activity and disrupts short-term working memory in the rat. Brain Research Bulletin, 52, 519-523.

REDMOND, A. J., MORROW, B. A., ELSWORTH, J. D., & ROTH, R. H. (2002). Selective activation of the A10, but not the A9, dopamine neurons in the rat by predator odor, 2,5, dihydro-2, 4, 5-trimethylthiazoline. Neuroscience Letters, 328, 209-212.

RICHARDSON, E. M., & WILLIAMS, J. L. (2003, October). Effects of predator odor on spatial reference and working memory. Presented at the Kenyon Summer Science Scholars' Poster Session. Kenyon College, Gambier, OH.

ROSELLINI, R. A. (1978). Inescapable shock interferes with the acquisition of an appetitive operant. Animal Learning & Behavior, 6, 155-159.

SANBERG, P. R., MORAN, T. H., KUBOS, K. L., & COYLE, J. T. (1983). Automated measures of stereotypic behavior in rats. Behavioral Neuroscience, 97, 830-832.

SANDI, C., LOSCERTALES, M., & GUAZA, C. (1997). Experience-dependent facilitating effect of corticosterone on spatial memory formation in the water maze. European Journal of Neuroscience, 9, 637-642.

SAPOLSKY, R. (1992). Stress, the aging brain, and the mechanisms of neuron death. Cambridge, MA: The MIT Press.

VERNET-MAURY, E. (1980). Trimethylthiazoline in fox feces: A natural alarming substance for the rat. In H. van der Starre (Ed.), Olfaction and Taste, 7, (p. 407). Washington, DC: IRL Press.

VERNET-MAURY, E., CONSTANT, B., & CHANEL, J. (1992). Repellent effect of trimethylthiazoline in the wild rat Rattus norvegicus Berkenhout. In R. L. Doty & D. Muller-Schwarze (Eds.), Chemical signals in vertebrates, 6, (pp. 305-310). New York: Plenum Press.

VERNET-MAURY, E., POLAK, E. H., & DEMAEL, A. (1984). Structure-activity relationship of stress-inducing odorants in the rat. Journal of Chemical Ecology, 10, 1007-1018.

VUNDERWOLF, C. H., & CAIN, D. P. (1994). The behavioral neurobiology of learning and memory: A conceptual reorientation. Brain Research & Review, 19:264-297.

WALLACE, K. J., & ROSEN, J. B. (2000). Predator odor as an unconditioned fear stimulus in rats: Elicitation of freezing by trimethylthiazoline, a component of fox feces. Behavioral Neuroscience, 114, 912-922.

WAN, R. Q., GIVENS, B., & OLTON, D. S. (1995). Opioid modulation of working memory: Intraseptal, but not intraamygdaloid, infusions of B-endorphin impair performance in spatial alternation. Neurobiology of Learning and Memory, 63, 74-86.

WESTERBROOK, R. F., GREELEY, J. D., NABKE, C. P., & SWINBOURNE, A. L. (1991). Aversive conditioning in the rat: Effects of a benzodiazepine and of an opioid agonist and antagonist on conditioned hypoalgesia and fear. Journal of Experimental Psychology: Animal Behavior Processes, 17, 219-230.

WILLIAMS, J. L. (1987). Influence of conspecific stress odors and shock controllability on conditioned defensive burying behavior. Animal Learning & Behavior, 15, 333-341.

WILLIAMS, J. L. (1989). Ethoexperimental analysis of stress, contextual odors, and defensive behaviors. In R. J. Blanchard, P. F. Brain, D. C. Blanchard, & S. Parmigiani, Ethoexperimental approaches to the study of behavior (pp. 214-228). Dordrecht: Martinus Nijhoff.

WILLIAMS, J. L. (1999). Effects of conspecific and predator odors on defensive behavior, analgesia, and spatial working memory. The Psychological Record, 49, 493-536.

WILLIAMS, J. L., BAKER, S. L., GRESS, J. E., & GIVENS, B. (1998). Effects of predator-induced stress and age on memory in rats. The Psychological Record, 48, 355-372.

WILLIAMS, J. L., & BARBER, R. G. (1990). Effects of cat exposure and cat odors on subsequent amphetamine-induced stereotypy. Pharmacology, Biochemistry, & Behavior, 36, 375-380.

WILLIAMS, J. L., DRUGAN, R. R., & MAIER, S. F. (1984). Exposure to uncontrollable stress alters withdrawal from morphine. Behavioral Neuroscience, 98, 836-846.

WILLIAMS, J. L., ROGERS, A. G., & ADLER, A. P. (1990). Prolonged exposure to conspecific and predator odors reduce fear reactions to these odors during subsequent prod-shock tests. Animal Learning & Behavior, 18, 453-461.

WILLIAMS, J. L., & SCOTT, D. K. (1989). Influence of conspecific and predatory stressors and their associated odors on defensive burying and freezing. Animal Learning & Behavior, 17, 383-393.

WILLIAMS, J. L., WORLAND, P. D., & SMITH, M. G. (1990). Defeat-induced hypoalgesia in the rat: Effects conditioned odors, naltrexone, and extinction. Journal of Experimental Psychology: Animal Behavior Processes, 16, 345-357.

WINER, B. J. (1971). Statistical principles in experimental design. New York: McGraw-Hill.

ZANGROSSI, H., Jr., & FILE, S. E. (1992). Behavioral consequences in animal tests of anxiety and exploration of exposure to cat odor. Brain Research Bulletin, 29, 381-388

ZANGROSSI, H., Jr., & FILE, S. E. (1994). Habituation and generalization of phobic responses to cat odors. Brain, Research, & Behavior, 33, 189-194.

JON L. WILLIAMS, CATHERINE BAEZ, KATHERINE J. HLADKY, and CHERI A. CAMACHO

Kenyon College