Behavioral Neuroscience 2000, Vol. 114, No. 2, 227-240
Copyright 2000 by the American PsychologicalAssociation, Inc. 0735-7044/00/$5.00 DOI: 10.1037//0735-7044.114.2.227
The Effects of Neurotoxic Hippocampal Lesions on Two Effects of Context After Fear Extinction Russell J. Frohardt, Fay A. Guarraci, and Mark E. Bouton University of Vermont Three conditioned suppression experiments with rats examined the role of the hippocampus in 2 effects of context after extinction. Reinstatement is the context-specific recovery of fear to an extinguished conditioned stimulus (CS) that occurs following independent presentations of the unconditioned stimulus (US), after extinction. Renewal is the recovery of fear when the CS is presented in the context in which it was conditioned, after extinction in a different context. Results indicated that neurotoxic lesions of the hippocampus, performed before conditioning, abolished reinstatement, which depends on context-US associations, but not renewal, which does not. This dissociation is not the result of differences in the recentness of context learning that ordinarily governs the 2 effects. The results suggest that the hippocampus is necessary for some, but not all, types of contextual learning.
Extinction is the behavioral phenomenon in which exposure
Recent evidence suggests that the hippocampal formation is important for learning about contextual stimuli (e.g., Holland & Bouton, 1999). For example, in rats, hippocampal lesions can abolish the ability to learn a context-shock association as measured by freezing in the presence of the context after fear conditioning (Kim & Fanselow, 1992; Phillips & LeDoux, 1992). Intrahippocampal infusions of N-methyl-D-aspartate (NMDA) and metabotropic glutamate receptor antagonists, as well as protein kinase C inhibitors, also attenuate contextual fear conditioning (Frohardt, Guarraci, & Young, 1999; Jerusalinsky et al., 1994; Young, Bohenek, & Fanselow, 1994). However, evidence also suggests that not all types of learning about the context are hippocampal dependent. Electrolytic lesions of the dorsal hippocampus do not impair the learning of a simple contextual discrimination in which rats are reinforced in one context and not reinforced in another one (Good & Honey, 1991). In addition, rats with neurotoxic hippocampal lesions or electrolytic lesions of the fimbria-fornix (a major output pathway of the hippocampus) can learn a conditional context discrimination task (McDonald et al., 1997). The overall pattern is thus not entirely consistent with the idea that the hippocampus is necessary to form a coherent representation of the context that might be required for all forms of context learning (cf. Fanselow, DeCola, & Young, 1993). The present experiments further examined the role of the hippocampus in context learning by investigating its role in the contextual control of performance after fear extinction.
to the conditioned stimulus (CS) on its own eliminates performance created after CS-unconditioned stimulus (US) pairings. The present experiments examined two examples of contextual control that suggest that extinction performance depends on the current context (e.g., see Bouton, 1991, 1993, for reviews). Although both are clearly context effects, behavioral research in this laboratory has shown that the two effects are controlled by different kinds of contextual learning. The first effect, reinstatement, occurs when exposure to the US alone after extinction causes a recovery of responding to the CS when it is presented later (e.g., Bouton & Bolles, 1979b; Rescorla & Heth, 1975). A great deal of evidence suggests that reinstatement is mediated by direct conditioning to the context that occurs when the US is presented after extinction. First, if the US is presented in an irrelevant context, reinstatement of responding to the CS is not observed (e.g., Bouton, 1984; Bouton & Bolles, 1979b; Bouton & King, 1983). That is, the US must be presented in the context of testing for reinstatement to occur. Second, extensive exposure to the context after the US presentations, which presumably reduces contextual conditioning through extinction, can abolish reinstatement (Bouton & Bolles, 1979b). Third, the strength of reinstatement is correlated with (and thus predicted by) the amount of contextual conditioning detected by an independent measure (e.g., Bouton, 1984; Bouton & King, 1983). Interestingly, reinstatement is not a simple matter of the CS and context summating. Similar amounts of contextual conditioning have less influence on responding to a CS that has not been extinguished (e.g., Bouton, 1984; Bouton & King, 1986); summation would be expected to occur equally well with these nonextinguished CSs. Contextual conditioning created by US exposure may trigger CS responding by reminding the subject of the original CS-US association after extinction (see Bouton, Rosengard, Achenbach, Peck, & Brooks, 1993). The second effect of context examined in this study is
Russell J. Frohardt, Fay A. Guarraci, and Mark E. Bouton, Department of Psychology, University of Vermont. This research was supported by Grants IBN 9209454 and IBN 9727992 from the National Science Foundation. We thank Mary Cain and William Falls for their comments on the manuscript. Correspondence concerning this article should be addressed to either Russell J. Frohardt or Mark E. Bouton, Department of Psychology, University of Vermont, Burlington, Vermont 05405. Electronic mail may be sent to either
[email protected] or
[email protected]. 227
228
FROHARDT, GUARRACI, AND BOUTON
renewal. In this effect, conditioning that occurs because of CS-US pairings in one context (Context A) is then extinguished in a second context (Context B). When the CS is presented again in the conditioning context (Context A), "renewed" responding is observed (e.g., Bouton & Bolles, 1979a; Bouton & King, 1983). This effect is mediated by a behavioral mechanism that is different from the one mediating reinstatement, and extensive evidence suggests that it does not depend on simple contextual conditioning (e.g., Bouton & King, 1983; Bouton & Swartzentruber, 1986). First, extensive extinction exposure to Context A that occurs between conditioning and testing in Context A does not abolish the renewal effect (Bouton & Swartzentruber, 1989). Second, the strength of the renewal effect is not correlated with direct contextual conditioning as assessed by the methods that correlate with reinstatement (Bouton & King, 1983) or by other methods (Bouton & Swartzentruber, 1986). Third, renewal is known to occur when testing takes place in a context that has never been associated with the US (e.g., Bouton & Bolles, 1979b; Bouton & Brooks, 1993). Fourth, the effect is not due to demonstrable inhibitory conditioning of Context B, the context of extinction (e.g., Bouton & King, 1983; Bouton & Swartzentruber, 1986, 1989). Thus, renewal seems to be independent of direct contextual conditioning. It instead depends on a mechanism similar to occasion setting (e.g., Holland, 1992; Swartzentruber, 1995). The idea is that the subject learns a CS-US association during conditioning and a CS-no-US association during extinction, and that the different contexts determine which of these associations is retrieved (Bouton & Swartzentruber, 1986). This mechanism is different from the simple contextual conditioning that controls the reinstatement effect. Wilson, Brooks, and Bouton (1995) recently extended the behavioral dissociation between reinstatement and renewal with a lesion of the hippocampal system. They examined the influence of radio-frequency lesions of the fornix, produced at the start of the experiment, and used a conditioned suppression (fear conditioning) method in which the US was provided by footshock. Wilson et al. found that fornix lesions abolished reinstatement but left the renewal effect relatively intact. Their results confirm the behavioral dissociation between the two effects and add to other evidence suggesting that at least some forms of context learning do not depend on the hippocampal formation. The fact that the fornix lesion specifically abolished reinstatement suggests that the hippocampus is necessary for the learning or expression of context-footshock associations (e.g., Kim & Fanselow, 1992; Phillips & LeDoux, 1992). In contrast, the lesion spared the occasion setting-like function of context represented by the renewal effect. The purpose of the present experiments was to pursue the dissociation further. They were motivated in part by a recent experiment by Fox and Holland (1998), who examined reinstatement in an appetitive conditioning method that used food pellets as the US (instead of footshock). In contrast to Wilson et al. (1995), Fox and Holland found no effect of hippocampal lesions on the reinstatement effect. However, in addition to studying appetitive conditioning (instead of fear conditioning), Fox and Holland used complete neuro-
toxic lesions of the hippocampus (instead of fornix lesions). Either methodological difference could account for the difference in the results. The present experiments, therefore, tested the effects of complete neurotoxic lesions by using a fear-conditioning procedure similar to the one used by Wilson et al. (1995). They also used neurotoxic lesions of the hippocampus produced by a mixture of ibotenic acid and NMDA (e.g., Han, Gallagher, & Holland, 1995). These lesions destroy the cells of the hippocampus while sparing fibers in its vicinity. In contrast, aspiration, electrolytic, and radio-frequency lesions can cause extensive damage to extrahippocampal neural elements. More specific lesions should help distinguish the role of the hippocampus in mediating the contextual control of extinction effects without implicating other structures of the brain. Experiment 1 The first experiment examined the reinstatement effect. Two groups of rats received neurotoxic lesions of the hippocampus, and two groups received sham lesions. After recovery, the rats were trained to lever-press for a food reward and were then given CS-footshock pairings superimposed on their baseline lever-pressing. After fear of the CS (indexed by the CS's suppression of the baseline) was acquired, the rats underwent extinction, in which the CS was presented repeatedly without the US. This eliminated overt signs of fear. In the crucial reinstatement phase, all rats next received eight noncontingent exposures to the footshock US in the absence of the CS. One group from each lesion condition received this US in the context in which conditioning and extinction had been conducted. The other group, however, received the same US in a different context (counterbalanced). Twenty-four hours later, the CS was then tested again in the context in which conditioning and extinction had occurred. On the basis of previous research, we expected to observe reinstated fear only in the rats that had received the US in the test context (e.g., Bouton, 1984; Bouton & Bolles, 1979b; Bouton & King, 1983; Wilson et al., 1995). However, if the previous fornix lesions had worked by damaging fibers that originated in the hippocampus proper, then no such reinstatement should be observed in the hippocampal-lesioned group (Wilson et al., 1995).
Method Subjects Subjects were 38 female Wistar rats obtained from Charles River Canada (Saint Constant, Quebec). They were between 75 and 80 days old at the start of the experiment and were individually housed in standard stainless steel cages in a room maintained on a 12-hr light-dark cycle. One week after surgery, the rats were food deprived and kept at 80% of their free-feeding weights throughout the remainder of the experiment.
Surgery All surgeries were performed under sufficient sodium pentobarbital (65 mg/kg) anesthesia, in aseptic conditions. Before any
HIPPOCAMPUS AND CONTEXT EFFECTS injection, the rats were transported to the surgical area in a covered cage, where they remained undisturbed for 30 min. Before anesthetic injections, rats were pretreated with atropine sulfate (0.05 mg/kg) to prevent respiratory distress. Neurotoxic lesions of the hippocampus were produced by injections of a mixture of 0.5 lag ibotenic acid and 0.5 lag N M D A per 0.5 ~ of normal saline. The solution was injected into the hippocampus at three injection sites per hemisphere. The coordinates (from dura) were: P, 3.3, L, 1.5, V, 2.5; P, 4.8, L, 3.5, V, 3.0; and P, 5.8, L, 5.0, V, 6.3. Each microinjection (0.5 lal) was infused through a 28-gauge cannula over 5 min via polyethylene tubing (PE 50) attached to a 10-lal gas-tight Hamilton syringe mounted on a syringe pump (Model 200, Stoelting, Wood Dale, IL). After each injection, the cannula was left in place for another 5 min. Sham rats received identical operative treatment except no skull holes were drilled and no cannulas were lowered. Immediately after surgery, the rats were placed under a heat lamp to prevent hypothermia. They were observed continuously until locomotion returned, at which time they were returned to their home cage. The postoperative recovery period was 7 days. Rats received analgesics (Buprenorphine, 0.05 mg/kg). The first postoperative analgesic was given when the rats had begun to stand, before any signs of pain or discomfort, and was 0.005 mg/kg. The remaining three doses were given at 8:00 a.m. and 8:00 p.m. for 36 hr. The rats were examined daily for any signs of illness.
Histology After behavioral testing, the rats were deeply anesthetized with sodium pentobarbital (120 mg/kg, ip) and perfused with 0.9% saline, followed by a 10% formol-saline solution. Their brains were frozen, sectioned at 70 larn, and stained with thionin for microscopic examination and verification of lesion sites.
Apparatus The apparatus consisted of two sets of four Skinner boxes that were located in different rooms of the laboratory. The boxes in one set measured 26 cm wide x 25 cm long x 19 cm high. The front, back, and one side wall were made of aluminum, and the remaining side wall and ceiling consisted of clear acrylic. The floor was made of tubular steel bars 16 mm in diameter, spaced 3.2 cm from center to center. The bars were perpendicular to the front wall where the operant manipulanda were located. A 4 × 4-cm food cup was located in the front wall, 1 cm above the floor and I cm to the left of a 2.5 x 2.5-cm lever. The lever was centered 5 cm from the right side of the wall and 5 cm above the floor. Rats were placed in these boxes through a door in the right side wall. A dish containing 10 ml of Heinz white vinegar, which was placed outside the box and below the food cup, provided a distinctive olfactory cue. The boxes in the second set measured 24 cm wide × 22 cm long × 18 cm high. The front and back walls were aluminum, and the side wall and ceiling consisted of clear acrylic, with 2-cm-wide vertical black stripes spaced 2.5 cm apart. The floor consisted of stainless steel bars 3 mm in diameter, spaced 1.5 cm from center to center. The bars were parallel to the front wall where the operant manipulanda were located. The 4 X 4-cm food cup was recessed in the front wall and was centered 3.5 cm from the right side wall, 1 cm above the floor. A 1-cm-wide lever protruded 4 cm from the front wall into the box and was located 5 cm above the floor and 6 cm to the left of the food cup. Rats were placed in these boxes through a door in the ceiling. A dish containing approximately 1 g of Vicks Vaporub, which was placed outside the box and below the food cup, provided a distinctive olfactory cue.
229
Both sets of boxes were placed in sound attenuation chambers, and illumination was provided by two 7.5-W white incandescent bulbs mounted on the ceilings of the sound attenuation chambers, 25 cm above the floor of the boxes. The CS was the 60-s offset of these lights, which produced complete darkness. Darkness has been used routinely as a CS in this laboratory to ensure that the rats perceived the CS as the same stimulus in the two contexts. Food pellets (45 rag, Formula A improved; P. J. Noyes, Lancaster, NH) were used to reinforce the operant baseline in each Skinner box. The footshock US (0.6 mA, 0.5 s) was provided by Grason-Stadler (West Concord, MA) shock sources and scramblers. The apparatus was controlled by computer equipment located in an adjacent room.
Procedure The contexts are referred to as Context A and Context B. Context A refers to the context in which Pavlovian fear conditioning was conducted, and Context B refers to the alternative context. The actual set of Skinner boxes that provided Contexts A and B were counterbalanced within each group. The experiment was conducted over consecutive days (one session per day) unless otherwise noted. All sessions were 90 min in duration unless otherwise specified. Baseline training. In the first baseline training session, conducted in Context A, each rat was put in a box, and a food pellet was delivered for each of the first 15 to 30 responses. The rat then progressed through leaner schedules of reinforcement until a 90-s variable-interval schedule (VI-90) was reached. The next session was conducted in Context B, where the VI-90 schedule was in effect. Two additional sessions of VI-90 training then took place in each context on alternate days. Training was suspended for one day between Sessions 4 and 5 of baseline training. Conditioning. After baseline training, all rats received two 2-day cycles of Pavlovian conditioning in Context A and exposure to Context B. The procedure was designed to provide all rats with equal exposure to the two contexts. On the 1st day of each cycle, half of the rats from each surgical condition received four presentations of the light-off CS, which terminated with the onset of the footshock US in Context A; the other half of the rats were permitted to lever-press in Context B on the VI-90 reinforcement schedule. On the 2nd day of each cycle, the treatments were reversed. Each group received two presentations of this A-to-B (or B-to-A) cycle, for a total of eight conditioning trials. The first trial always occurred approximately 16 min after the start of the session. The intertrial interval (ITI) was variable, with a mean of 20 min. Extinction. All of the groups then underwent extinction in the original conditioning context (Context A). Extinction was carried out in four 2-day cycles. On the 1st day of each cycle, half of the rats from both surgical conditions were presented with eight presentations of the light-off CS in Context A, and the remaining rats were permitted to lever-press on the VI-90 reinforcement schedule in Context B. On the 2nd day of each cycle, the rats that received conditioning trials on the 1st day were permitted to lever-press in Context B, and the remaining rats received extinction in Context A. This procedure continued to provide all rats with equal exposure to the two contexts. The first extinction trial always occurred approximately 9 min after the start of the session. The ITI was variable, with a mean of 9 min; there were 8 trials per extinction session, for a total of 32 extinction trials. Reinstatement. New groups were created after extinction. One of the new groups from each surgical condition (Sham/Same and Hippocampal/Same) received unsignaled shocks in the same context in which extinction had occurred (Context A); the other groups (Sham/Different and Hippocampal/Different) receiyed the
230
FROHARDT, GUARRACI, AND BOUTON
same shocks in the alternate, nonextinction context (Context B). During the single, 90-min reinstatement shocking session, every rat received eight unsignaled presentations of the 0.5-s, 0.6-mA shock while lever-pressing. The intershock interval was variable, with a mean of 9 min. Reinstatement test. Reinstatement testing was conducted the next day in Context A for all groups. All rats received four presentations of the light-off CS alone during the 60-min test session. The ITI was variable, with a mean of 12 min. Data analysis. Suppression to the CS was indexed in terms of standard suppression ratios of the form xl(x + y), in which x represents the number of lever presses made during the CS and y represents the number of lever presses during the 60-s period immediately before the CS (the pre-CS period). A suppression ratio of 0 indicates complete suppression of responding during the CS, whereas a suppression ratio of 0.5 indicates no response suppression during the CS. Suppression ratios were analyzed with analyses of variance (ANOVAs). Corresponding analyses were also conducted on the pre-CS scores. Throughout, statistical tests were considered significant ifp was less than .05. To further assess possible contextual conditioning during crucial parts of the experiment, pre-CS scores were converted into context suppression ratios in the following way. In the equation x/(x + y), x represented the number of lever presses made during the focal pre-CS periods and y represented the average number of lever presses made during the pre-CS periods of the previous session in the same context. The suppression ratio has been shown to be especially sensitive to baseline response suppression (Church, 1969). It provides more power than raw pre-CS scores because it normalizes variability in baseline response rates.
Results Histology Rats were included in data analyses if there was extensive cell loss in all subfields of the hippocampal formation (CA1-CA3, dentate gyrus). Some rats sustained minimal damage to either the subiculum, entorhinal cortex, or cortex directly dorsal to the infusion sites. Rats were excluded if substantial portions of the hippocampal neuronal population were left intact or if the cortical damage was extensive. Figure 1 shows the representative largest and smallest lesion of the hippocampal formation for all lesioned rats retained in the three experiments. All 19 rats lesioned in Experiment 1 showed sufficient damage of the hippocampus (with minimal extra-hippocampal damage) to be included in the analyses.
Conditioning The left panel of Figure 2 presents suppression to the CS during each two-trial block of the conditioning phase. All groups acquired conditioning, and there were no reliable differences among the groups. A Surgical Group X Shock Context X Trial Block ANOVA revealed a nonsignificant main effect of surgical group, F(1, 34) < 1. There was a significant Trial Block X Surgical Group interaction, F(3, 102) = 5.28. Further comparisons of each trial block during conditioning showed that the sham rats exhibited significantly more suppression to the CS than did lesioned rats during the first trial block, t(36) = 3.98. However, there
Figure 1. Coronal sections through the hippocampus (2.3045.30 mm posterior to bregma). Gray areas illustrate the representative largest lesion, and black areas illustrate the representative smallest lesion of the hippocampus from all three experiments. From The Rat Brain in Stereotaxic Coordinates, by G. Paxinos and C. Watson, 1998, San Diego, CA: Academic Press. Copyright 1998 by Academic Press. Adapted with permission.
were no differences between the surgical groups on the final trial block, t(36) = 0.23. All other main effects and interactions failed to reach statistical significance (Fs < 1), except for the main effect of trial block, which was highly significant, F(3, 102) = 48.16. These analyses confirm that rats with hippocampal lesions learned to suppress ongoing operant responding in the presence of the light-off CS when it was paired with shock. The mean pre-CS baseline response rates for each group during the conditioning sessions in Context A are shown in Table 1. A Surgical Group X Shock Context x Trial Block ANOVA on the number of lever-press responses observed during the pre-CS period revealed nonsignificant main effects of surgical group and extinction context, Fs(1, 34) < 3.12, as well as trial block, F(3, 102) < 1. Similarly, all interactions failed to reach statistical significance.
Extinction The right panel of Figure 2 presents the suppression ratios to the CS for each four-trial block during the extinction phase. The data suggest that extinction of suppression
HIPPOCAMPUS AND CONTEXT EFFECTS 0.6
Acquisition
231
The mean pre-CS baseline response rates for each group during the extinction sessions are listed in Table 1. A Surgical Group X Shock Context X Trial Block A N O V A on the number of lever-press responses observed during the pre-CS period revealed a nonsignificant main effect of extinction context, F(1, 34) < 2.22. The main effects of trial block, F(7, 238) = 3.80, and surgical group, F(1, 34) = 5.69, were statistically significant. The baseline rate of lever-pressing increased in all groups over the course o f extinction, but the hippocampal-lesioned rats responded more during baseline throughout extinction.
Extinction
0.5 •~ 0.4 •~ 0.3 ~0.2 e
I
O.1
2
4
2-Trial Blocks
"-
