Differential involvement of the hippocampus, anterior cingulate cortex, and basolateral amygdala in memory for context and footshock

Abstract

Extensive evidence from contextual fear conditioning experiments suggests that the hippocampus is involved in processing memory for contextual information. Evidence also suggests that the rostral anterior cingulate cortex (rACC) may be selectively involved in memory for nociceptive stimulation. In contrast, many findings indicate that the basolateral amygdala (BLA) is more broadly involved in modulating the consolidation of different kinds of information. To investigate further the differential involvement of these brain regions in memory consolidation, the present experiments used a modified inhibitory avoidance training procedure that took place on 2 sequential days to separate context training from footshock training. Male Sprague–Dawley rats were implanted with unilateral cannulae aimed at the (i) hippocampus, (ii) rACC, or (iii) BLA, and given infusions of the muscarinic cholinergic agonist oxotremorine (OXO) immediately after either context training (day 1) or footshock training in that context (day 2). OXO enhanced retention when infused into the hippocampus after context, but not footshock, training. Conversely, OXO infusions enhanced memory when administered into the rACC immediately after footshock, but not context, training. Lastly, intra-BLA OXO infusions enhanced retention when administered after either context or footshock training. These findings are consistent with evidence that the hippocampus and rACC play selective roles in memory for specific components of training experiences. Additionally, they provide further evidence that the BLA is more liberally involved in modulating memory consolidation for various aspects of emotionally arousing experiences.

Considerable evidence indicates that the hippocampus is involved in memory for tasks, such as inhibitory avoidance (IA) and contextual fear conditioning (CFC), that require learning of contextual cues. Pretraining hippocampus lesions or infusions of cholinergic antagonists or lidocaine administered into the hippocampus in rats impair the acquisition of tasks that involve associating a context with an aversive event (1–8) or a novel object (9). Furthermore, posttraining manipulations of the hippocampus similarly affect the retention of a training context associated with a footshock (10–14). Substantial evidence indicates that retention of CFC training is poor if the animal is given a brief shock immediately after being placed in the training apparatus and not given any subsequent time in the context, a phenomenon known as “immediate shock deficit” (15). However, preexposure to the context 24 h before immediate shock training prevents this deficit and is referred to as the “context preexposure facilitation effect” (16–18). Consistent with evidence that the hippocampus plays a role in contextual memories, pre- or postcontext manipulations of the hippocampus modulate retention of memory of subsequent CFC training in the same context (19–22).

Whereas the hippocampus is implicated in memory for the context of a task, there is evidence suggesting that the rostral anterior cingulate cortex (rACC) may be selectively involved in memory for tasks that involve nociceptive stimulation. We recently found that infusions of the cholinergic agonist oxotremorine (OXO) into the rACC of rats immediately after IA training enhance retention (unpublished findings). OXO infusions administered into the immediately adjacent caudal anterior cingulate cortex (cACC) were ineffective. In addition, previous studies report that pretraining lesions of the rACC impair acquisition of classical conditioning of a noxious CO₂ laser pulse with a tone (23), and that immediate posttraining rACC lesions impair IA retention (24). Importantly, Johansen, Fields, and Manning (25) reported that lesions of the rACC impaired conditioned place avoidance induced by a nociceptive hind paw formalin injection, but not conditioned place avoidance for a non-nociceptive i.p. κ-opioid injection. Other studies have reported that lesions of or infusions into rACC do not affect retention for tasks, such as water maze or appetitively motivated working memory tasks, that do not involve nociceptive stimulation (26–29). Such findings are consistent with the hypothesis that the rACC may be selectively involved in memory for aversive training that uses a nociceptive stimulus.

Whereas the findings discussed above suggest that the hippocampus and rACC may be selectively involved in memory for contextual cues and nociceptive stimulation, respectively, extensive evidence indicates that the basolateral amygdala (BLA) is more liberally involved in modulating the consolidation of memory for a variety of emotionally arousing experiences, including those of IA, CFC, spatial water maze, appetitive T-maze, conditioned taste aversion, and conditioned place preference training (e.g., refs. 11 and 30–35). Previous findings from our laboratory indicate that the BLA modulates memory consolidation via interactions with many efferent brain regions (for review, see refs. (36–38). Lesions or pharmacological inactivations of the BLA disrupt the memory-modulatory effects of posttraining infusions of drugs administered into various brain regions, including the hippocampus (11, 12), insular cortex (34), medial prefrontal cortex (33), entorhinal cortex (39), and nucleus accumbens (40). In experiments using IA training, we recently found that lesions of the BLA also block the memory-enhancing effect of posttraining intra-rACC infusions of the muscarinic cholinergic agonist OXO. We also found, conversely, that rACC lesions blocked the effect of posttraining OXO infusions into the BLA (unpublished findings). Although such findings suggest the possibility that the rACC and BLA may have equivalent functions in memory consolidation, the evidence that BLA effects on memory are not selective for a particular component of aversive training questions that interpretation. This issue is investigated further in the present experiments.

IA training requires animals to associate a specific context with an aversive footshock. In typical IA training, both the context and footshock components of training are experienced during a single training trial, in which rats explore the length of an IA box before they are confined to a dark compartment and given a footshock. However, such IA training does not allow assessment of the relative involvement of a brain region to consolidation of memory for the context to be studied independently from that of the footshock, because both components occur during the single training trial. To address this issue, Liang (41) used a modified IA training procedure, based on the context preexposure findings discussed above (16–18), in which context training alone and brief footshock training alone occur on 2 sequential days. This modified IA training results in subsequent retention latencies comparable to those found with standard one-trial IA training.

The present experiments used the modified IA training to investigate the hypothesis that the hippocampus, rACC, and BLA may be differentially involved in the consolidation of memory for context training or footshock training. We used the muscarinic cholinergic agonist OXO in the present experiments based on previous findings that infusions of muscarinic agonists or antagonists into the hippocampus, rACC, and BLA affect retention of aversive training tasks (e.g., refs. 26, 42, and 43). OXO infusions were administered into the hippocampus, rACC, or BLA immediately after either context or footshock training, and 48-h retention was assessed by latencies to enter the footshock compartment of the training apparatus.

Results

Histology.

(Figs. 1–3) show representative photomicrographs of needle tracks terminating in the hippocampus, rACC, and BLA. Animals with improper injection needle placements were excluded from the analysis.

Fig. 1.

Representative photomicrograph of needle track terminating in thehippocampus with diagram of hippocampus and surrounding structures (58).

Fig. 2.

Representative photomicrograph of needle track terminating in the rACCwith diagram of rACC and surrounding structures (58).

Fig. 3.

Representative photomicrograph of needle track terminating in the BLAwith diagram of BLA and surrounding structures (58).

Experiment 1.

The retention latencies of animals given infusions into the hippocampus immediately after context or footshock training of IA are shown in Fig. 4. A two-way ANOVA revealed a significant main effect of infusion day [F_(1,75) = 12.518, P < 0.001], a significant main effect of the intra-hippocampus infusions [F_(2,75) = 3.352, P < 0.05], and a significant interaction between infusion day and intra-hippocampus infusions [F_(2,75) = 3.560, P < 0.05]. The retention latencies of rats given intra-hippocampus infusions of 10 ng and 100 ng of OXO immediately after context training were significantly longer than those of saline context training controls (P < 0.001 for 10 ng, and P < 0.01 for 100 ng) and their respective footshock training counterparts (P < 0.001 for 10 ng, and P < 0.01 for 100 ng). Furthermore, the retention latencies of rats given intra-hippocampus infusions of saline after context training did not differ from those of post-footshock training saline controls (P > 0.05).

Fig. 4.

Enhanced retention of rats that received unilateral infusions of OXOinto the hippocampus immediately after context training (A) but not footshock training (B). Results represent mean + SEM retention latencies in seconds. ∗, P < 0.01; ∗∗, P < 0.001 (compared with vehicle group). A, n = 16, 14, and 10, respectively; B, n = 13, 14, and 14, respectively.

Experiment 2.

Fig. 5 shows the retention latencies of animals given unilateral infusions into the rACC immediately after either context or footshock IA training. A two-way ANOVA revealed a significant main effect of infusion day [F_(1,55) = 10.636, P < 0.01], a significant main effect of the intra-rACC infusions [F_(2,55) = 3.381, P < 0.05], and a significant interaction between the rACC lesions and the intra-BLA infusions [F_(2,55) = 3.250, P < 0.05]. The retention latencies of rats given intra-rACC infusions of 0.5 ng and 10 ng of OXO immediately after footshock training were significantly longer than those of saline footshock training controls (P < 0.01 for 0.5 ng, and P < 0.05 for 10 ng) and their respective context training counterparts (P < 0.01 for 0.5 ng, and P < 0.05 for 10 ng). The retention latencies of rats given postfootshock intra-rACC saline infusions did not differ from those of postcontext saline controls (P > 0.05).

Fig. 5.

Enhanced retention of rats that received unilateral infusions of OXOinto the rACC immediately after footshock training (B) but not context training (A). Results represent mean + SEM retention latencies in seconds. ∗, P < 0.05; ∗∗, P < 0.01 (compared with vehicle group). A, n = 12, 11, and 10, respectively; B, n = 12, 9, and 7, respectively.

Experiment 3.

The retention latencies of animals given infusions into the BLA immediately after context or footshock training of IA are shown in Fig. 6. A two-way ANOVA revealed only a significant main effect of intra-BLA infusions [F_(2,68) = 8.630, P < 0.001]. The retention latencies of rats given intra-BLA infusions of 10 ng and 100 ng of OXO immediately after both context and footshock training were significantly longer than those of their respective saline controls (P < 0.05 for postcontext 10 ng and postcontext 100 ng, P < 0.05 for post-footshock 10 ng, and P < 0.01 for post-footshock 100 ng). Lastly, the retention latencies of rats given intra-BLA infusions of saline after context training did not differ from those of post-footshock training saline controls (P > 0.05).

Fig. 6.

Enhanced retention of rats that received unilateral infusions of OXOinto the BLA immediately after context training (A) and footshock training (B). Results represent mean + SEM retention latencies in seconds. ∗, P < 0.05; ∗∗, P < 0.01 (compared with vehicle group). A, n = 11, 10, and 9, respectively; B, n = 13, 16, and 15, respectively.

Discussion

These findings provide evidence that, (i) the hippocampus is selectively involved in memory consolidation for training of the context component of IA, (ii) the rACC is selectively involved in memory consolidation for the footshock component of IA training, and (iii) the BLA is involved in memory consolidation for both context and footshock IA training. Furthermore, these findings provide additional evidence that memory consolidation for the context and footshock components of IA can be independently studied by using Liang’s (41) modified IA training procedures. Unilateral OXO infusions into the hippocampus enhanced retention for an IA task when given immediately after context training but not when given after footshock training. In contrast, infusions of OXO into the rACC enhanced retention when given after the footshock component of IA training but not when administered after the context training. Finally, intra-BLA OXO infusions enhanced retention when given after either context or footshock training. These results are consistent with previous evidence suggesting that the hippocampus and rACC are selectively involved in contextual and nociceptive memory consolidation, respectively, whereas the BLA is more generally involved in modulating the consolidation of memory for emotionally arousing experiences.

As discussed above, there is considerable evidence that pretraining treatments affecting hippocampal functioning influence acquisition of a variety of tasks, including CFC (2–7), context preexposure CFC (20–22), and the context component of a novelty-preference paradigm (9). When these tasks used other cues during training, such as a tone, manipulations of the hippocampus selectively impaired retention for the context and did not disturb memory of the other cues. Moreover, posttraining drug infusions or stimulation of the hippocampus affect retention of IA training (10–12, 14), CFC (7, 13), and postcontext preexposure CFC (19, 22). However, inactivation of the hippocampus after standard IA or immediate footshock CFC training does not impair retention if the rats are preexposed to the context (19, 44). These findings suggest that the hippocampus consolidates contextual memories after the initial exposure to the context and is not necessary for the subsequent association of the context with a footshock (45, 46). The present finding that OXO infusions into the hippocampus enhanced IA retention only when administered after context training provides further evidence that the hippocampus is selectively involved in memory consolidation for contextual components of tasks after the initial exposure to the context.

A major aim of the current experiments was to investigate the hypothesis that the rACC plays a role in memory for nociceptive stimuli. The ACC is involved in the emotional components of pain (for review, see ref. 47), and administration of noxious stimuli alone, such as hind paw formalin injections or footshock, as well as during training, induces c-Fos expression and increased ATP hydrolysis in the ACC (48, 49). Studies of both monkeys and humans have shown that the ACC is activated by nociceptive stimuli as well as by the anticipation of pain (50–52). Recent findings indicate that the ACC is composed of at least two functionally distinct regions. fMRI studies of human subjects find that the rostral and caudal portions of ACC respond differentially to tasks that vary in emotional content. In these experiments, the cACC is activated by cognitive tasks and inhibited by emotional tasks, whereas the rACC is activated by emotional tasks and inhibited by cognitive tasks (53–56). Furthermore, lesions of the rACC, but not cACC, disrupt conditioned place avoidance induced by nociceptive hind paw formalin injections (25). We chose to study the rostral portion of ACC in the current experiments based on these studies, as well as our previous finding that posttraining OXO infusions enhance IA retention when administered into the rACC but not when infused into the adjacent cACC (unpublished findings). Recently, Johansen, Fields, and Manning (25) found that lesions of the rACC impair conditioned place avoidance when a context is paired with a nociceptive hind paw formalin injection but not when the aversive stimulus was a non-nociceptive κ-opioid agonist, suggesting that the rACC may play a selective role in memory for nociceptive events. As discussed above, we previously found that posttraining OXO infusions into the rACC enhanced retention for standard IA training. Our present finding that OXO infusions into the rACC enhanced memory when given after footshock training, but not after context training of IA, provides evidence that the rACC is specifically involved in learning about experiences that involve nociceptive stimuli. Because infusions of OXO into the hippocampus after context training presumably enhanced the memory representation of the context, it is likely that OXO infusions administered into the rACC after footshock training enhanced the memory representation of the footshock. However, our findings do not exclude the alternative interpretation that infusions of OXO into the rACC may have strengthened the association between the context representation and the footshock representation. This is an important distinction that cannot be resolved by the findings of the present experiments alone.

The findings of experiments 1 and 2 indicate that the hippocampus and rACC can be dissociated based on their involvement in memory consolidation for the context and footshock components of IA training. Substantial evidence indicates that the BLA is involved in the consolidation of memory of training tasks that are emotionally arousing, regardless of task modality. As discussed above, lesions of or drug infusions administered into the BLA affect retention of a wide variety of tasks, including spatial water maze, CFC, appetitive T-maze tasks, conditioned taste aversion, and conditioned place preference (e.g., refs. 11, 13, and 30–35). The present finding that OXO infusions administered into the BLA enhanced IA retention regardless of whether the infusions were given after context training or footshock training is consistent with previous evidence that the BLA is liberally involved in memory consolidation when the task is emotionally arousing.

Our previous finding that lesions of the BLA or rACC block the memory-modulatory effects of infusions into the other region (unpublished findings) might be interpreted as evidence that the BLA and rACC can play comparable roles during memory consolidation. According to this interpretation, infusions of OXO into the rACC in the current experiments should have had effects comparable to those produced by OXO infusions administered into the BLA. Our findings very clearly fail to support this view, because OXO infusions into the rACC enhanced retention only for footshock training, whereas intra-BLA OXO infusions enhanced retention of both footshock training and context training. In our previous study, rACC lesions may have blocked the effect of OXO infusions into the BLA because IA involves learning about a nociceptive footshock. Our current findings suggest that lesions of the rACC may not block the effect of intra-BLA OXO infusions on memory of training that does not involve nociceptive stimulation, such as object recognition or spatial water maze training. This implication has not, as yet, been investigated.

In summary, the present findings provide evidence indicating that the hippocampus and rACC are differentially involved in modulating the consolidation of memory for context and footshock stimulation used in IA training and provide additional evidence that the BLA is liberally involved in modulating the consolidation of different components of emotionally arousing experiences.

Materials and Methods

Subjects.

Male Sprague–Dawley rats (≈300 g at the time of surgery) from Charles River Laboratories were housed individually in a temperature-controlled (22°C) vivarium and maintained on a 12-h light/dark cycle (7 a.m. to 7 p.m.) with food and water available ad libitum. Rats were given 1 week to acclimate to the vivarium before surgery. Behavioral procedures began 6–10 days after surgery. All training and testing occurred between 10:00 a.m. and 5:00 p.m. All procedures were in accordance with National Institutes of Health guidelines and were approved by the University of California at Irvine Institutional Animal Care and Use Committee.

Surgery.

The rats were anesthetized with sodium pentobarbital (50 mg/kg, i.p.) and given atropine sulfate (0.1 mg/2 ml, i.p.) to maintain respiration and 0.9% sterile saline to prevent dehydration (3.0 ml). The animals were then placed in a small animal stereotaxic frame (Kopf Instruments, Tujunga, CA). In experiment 1, animals were implanted unilaterally with stainless-steel guide cannulae (11 mm) aimed above the hippocampus. The following stereotaxic coordinates were used: anteroposterior (AP), −3.4 mm from bregma; mediolateral (ML), ±1.7 mm from midline; dorsoventral (DV), −2.7 mm from skull surface. In experiment 2, animals were implanted with unilateral cannulae (10 mm; 23 gauge) aimed above the rACC (AP, +2.7; ML, ±0.5; DV, −1.5). For experiments 1 and 2, the cannulae were counterbalanced for side. In experiment 3, unilateral cannulae (15 mm) were implanted above the BLA (AP, −2.8; ML, −5.0; DV, −6.5). The cannulae were limited to the right side based on recent evidence from our laboratory that infusions into the right BLA are more effective than those into the left side (57). The cannulae were affixed to the skull with dental cement and two anchoring surgical screws (Small Parts, Miami Lakes, FL), and insect pins (10-, 11-, or 15-mm-long 00 insect dissection pins) were inserted into the cannulae to maintain patency. After the surgery, the rats were kept in an incubator until they awoke from the anesthesia.

Behavioral Procedures.

Each rat was handled for at least 1 min per day for 3 days before the start of training. The rats were trained and tested in a modified IA task developed by Liang (41). The IA apparatus consisted of a trough-shaped alley (91 cm long, 15 cm deep) divided into two compartments by a retractable door: an illuminated safe compartment (31 cm long) and a dark shock compartment (60 cm long). On day 1 (context training), each rat was placed in the lit start compartment facing away from the shock compartment and was allowed to freely explore the IA box for 3 min. On day 2 (footshock training), each rat was placed into the dark compartment, facing away from the lit compartment, with the retractable door closed. The rat then received an immediate inescapable mild footshock (0.8 mA for 1.0 s for experiments 1 and 2; 1.0 mA for 2.0 s for experiment 3); OXO was administered into the hippocampus, rACC, or BLA immediately after either the context or footshock training. On the retention test 48 h later, each rat was placed into the light compartment with the retractable door open and was permitted to explore the box freely. The latency to enter the dark (shock) compartment with all four paws was recorded as a measure of retention. A maximum of 600 s was recorded on the retention test.

Drugs and Infusion Procedures.

For each experiment, the drug solutions were infused directly into the hippocampus, rACC, or BLA immediately after context or footshock training to selectively affect the consolidation phase of memory for either component of training. Control animals received infusions of the vehicle solution (0.9% sterile saline). All other animals received infusions of the nonselective muscarinic cholinergic agonist OXO (10 or 100 ng in 0.5 μl for hippocampus; 0.5 or 10 ng in 0.5 μl for rACC; 10 or 100 ng in 0.2 μl for BLA). OXO was obtained from Sigma.

The drug solutions were infused at a constant rate (over 57.5 s for hippocampal and rACC infusions, and 32 s for BLA infusions) by an automated syringe pump (Sage Instruments, Boston) through a 30-gauge needle (extending 1 mm past the cannulae for rACC and hippocampus, and 2 mm past the cannulae for BLA) that was attached by polyethylene tubing to a 10-μl Hamilton syringe.

Histology.

After the behavioral tests were completed, the rats were anesthetized with an overdose of sodium pentobarbital (100 mg/kg i.p.) and perfused intracardially with 0.9% saline and then 10% formaldehyde. After decapitation, the brains were removed and placed in 10% formaldehyde for a minimum of 24 h and were then cryoprotected in a 30% sucrose solution. Coronal slices of 50 μm were taken with a freezing microtome, mounted on gelatin-coated slides, and stained with thionin. The sections were examined under a light microscope to determine injection needle placement according to the standardized atlas plates of Paxinos and Watson (58).

Statistics.

The behavioral data from each experiment were analyzed with a two-way ANOVA, with intra-hippocampus, rACC, or BLA drug treatment as the repeated measure and infusion day (context or footshock) as the between-subjects variable. For each experiment, Fisher’s post hoc tests were performed to determine the source of detected significance. p values of <0.05 were considered significant. All measures are expressed as mean ± SEM.

Abbreviations

BLA

basolateral amygdala

cACC

caudal anterior cingulate cortex

CFC

contextual fear conditioning

IA

inhibitory avoidance

OXO

oxotremorine

rACC

rostral anterior cingulate cortex.