- Email: [email protected]
Impaired water maze navigation of Wistar rats with retrosplenial cortex lesions: effect of nonspatial pretraining
Behavioural Brain Research 158 (2005) 175–182
Research report
Impaired water maze navigation of Wistar rats with retrosplenial cortex lesions: effect of nonspatial pretraining Nikolai V. Lukoyanov∗ , Elena A. Lukoyanova, Jos´e P. Andrade, Manuel M. Paula-Barbosa Department of Anatomy, Porto Medical School, 4200-319 Porto, Portugal Received 2 July 2004; received in revised form 26 August 2004; accepted 30 August 2004 Available online 6 October 2004
Abstract Damage to the retrosplenial cortex (RC) impairs the performance of rodents on spatial learning and memory tasks, but the extent of these deficits was previously reported to be influenced by the lesion type, rat strain, and behavioral task used. The present study addressed the issue of whether or not cytotoxic damage to RC impairs place navigation of Wistar rats in the Morris water maze and, if so, whether this is merely attributable to spatial learning deficits or to impaired learning of general (nonspatial) behavioral strategies required to correctly perform this task or both. Behaviorally naive rats with bilateral lesions to RC were significantly impaired relative to sham-lesioned rats both during the period of initial learning of the task and during the later phases of training. In addition, these animals showed enhanced thigmotaxis, indicating that the lesion was associated with considerable abnormalities in nonspatial learning. In contrast, RC-lesioned animals that have been previously familiarized with general task rules in a series of shaping trials did not show more thigmotaxis than did their respective controls. Furthermore, although these rats were still impaired in the middle of the training process, their performance during the period of initial learning as well as by the end of training was found to now be normal. Our results confirm those of earlier studies indicating that RC is important for spatial navigation. The findings herein reported are also consistent with the notion that, in addition to spatial information processing, RC is involved in cognitive processes underlying the ability of subjects to properly respond to general task demands. © 2004 Elsevier B.V. All rights reserved. Keywords: Retrosplenial cortex; Lesion; Learning; Spatial navigation; Water maze; Nonspatial pretraining
1. Introduction The retrosplenial cortex (RC) is reciprocally connected with the hippocampal formation and anterior thalamic nuclei [16,17,20–22,26,27], the key elements of the brain circuit that mediates memory formation [1,2,6,10]. That RC constitutes a part of this circuit is supported by evidence from behavioral studies in rodents indicating that damage to this area of cortex impairs spatial learning, which is known to be particularly dependent on the hippocampal formation [7,9,12,18,23,24,30]. The finding that inactivation of RC also impairs navigation in the darkness and alters the firing pattern of hippocampal place cells suggests that RC specifically contribute to spatial
∗
Corresponding author. Tel.: +351 22 509 6808; fax: +351 22 550 5640. E-mail address: [email protected] (N.V. Lukoyanov).
0166-4328/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbr.2004.08.023
behavior by integrating visual cues with information generated by self-motion [4,5]. Although the involvement of RC in spatial navigation is well documented, the results of the previous studies on RClesioned animals are not entirely uniform, particularly with respect to the severity of behavioral changes observed. More specifically, it was found that suction ablations of RC [7,9,30] generally affect spatial abilities of rats to a greater extent than do excitotoxic lesions [9,24,25], suggesting that, in the former case, damage to fibers of passage can partially account for behavioral deficits. It has also been shown that the extent of the spatial impairments is a determinant of whether or not the lesion involves the most caudal portion of RC [23]. Making the picture even more complicated, it has been reported that removal of RC produces robust impairments of spatial navigation in the Long–Evans strain of rats [7,30], while it causes relatively little or no place learning deficits
176
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
in Dark Agouti rats [14,28]. With respect to the latter variable, it has been suggested that the effect of RC lesions on spatial navigation of Dark Agouti rats might be masked by an innate nonspatial impairment, which is characteristic of this strain [7]. Supporting this argument is the observation that intact Dark Agouti rats are impaired, compared to intact Long–Evans rats, both in overall performance on the water maze and in learning of general task requirements [8]. However, the hypothesis that the spatial navigation deficits associated with damage to RC can be masked by some nonspatial impairments remains to be further verified. In the present study, we addressed this issue by assessing the behavioral effects of RC lesions in Wistar rats, which were previously shown to perform spatial tasks at a level comparable with that of Dark Agouti rats [8]. The behavioral assessments were performed using the Morris water maze, a paradigm on which Dark Agouti rats with RC lesions display little or no impairments. Two separate experiments were performed: (1) in behaviorally naive rats and (2) using animals that received nonspatial pretraining before being tested on the place learning task. The latter experiment was intended to unmask the lesion-induced spatial navigation deficits from those related to nonspatial learning.
2. Methods 2.1. Animals and surgery Male Wistar rats, derived from animals obtained from Gulbenkian Institute of Science (Oeiras, Portugal), were used in the present study. Animals were housed two to three per cage, maintained under standard laboratory conditions (20–22 ◦ C and a 12 h:12 h light–dark cycle), and had free access to food and water. At 12 weeks of age, they were subjected either to cytotoxic lesions of RC or to a sham surgical procedure. Rats were anesthetized by subsequent injections of promethazine (10 mg/kg), xylazine (2.6 mg/kg), and ketamine (50 mg/kg), and placed in a Kopf stereotaxic apparatus. Under aseptic conditions, the scalp was incised along the midline and retracted to the side. To expose the brain surface, a piece of parietal bone, 1.8 mm wide and 5 mm long, was removed on either side of the midline using a small drill mounted to the stereotaxic apparatus. However, a narrow bar of bone over the superior sagittal sinus, approximately 1.4 mm wide, was left intact. Bilateral cytotoxic lesions of RC were made by infusing 0.3 l of N-methyl-d-aspartate (NMDA, 0.09 M saline solution) at the following coordinates: AP (from bregma) −2.9, ML (from midline) ±0.8, DV (from dura) −1.8; AP −4.4, ML ±0.8, DV −1.7; AP −5.9, ML ±1.0, DV −2.0. All infusions were delivered over a 3-min period using a 1-l Hamilton syringe equipped with a 33 gauge blunt-tip needle. Five minutes of diffusion time were allowed before the needle was retracted. Sham control rats underwent an identical procedure but received vehicle alone rather than NMDA solution.
2.2. Behavioral procedures The behavioral consequences of RC lesions were evaluated using the Morris water maze paradigm. The maze consisted of a black circular tank, 180 cm in diameter and 50 cm deep, and was located in a corner of a room containing extramaze cues, i.e., three posters of different size and shape, and a computer desk. The apparatus was filled with water at room temperature (21 ± 1 ◦ C) to a depth of approximately 35 cm. The water was made opaque by adding a non-toxic paint. The maze was divided, by imaginary lines, into four equal-size quadrants. The swim path was recorded by a computerized video-tracking system (EthoVision V3.0, Noldus, The Netherlands). All rats included in the second experiment were familiarized with the general task requirements in two sessions of shaping trials in which no spatial learning was required [11]. In this procedure, a black acrylic alley, measuring 130 cm × 65 cm × 15 cm, was placed on the bottom of the maze so that its walls extended 30 cm above the surface of the water. A black escape platform, 10 cm in diameter, was placed inside the alley, approximately 25 cm from one of its ends. It was located 2 cm below the surface of the water. Each rat was placed in the water and allowed to swim to the platform and to climb on it. If the rats did not find the escape platform within 20 s, the experimenter guided them to the platform. Each session consisted of three trials, spaced by 15-s inter-trial intervals, in which the rats were released in the water from three different locations, i.e., 30, 60 and 90 cm from the position of the platform. The interval between the two shaping sessions was 24 h. No extramaze cues were present during this phase of training. In the classical place learning task [13], the animals were trained to find the submerged escape platform, which was now located in the center of one of the quadrants. For acquisition, rats were given two trials on each day for 14 consecutive days. Each rat was placed in the water facing the pool wall at one of the four starting points that were used in a pseudo-random order so that each position was used once in each block of four trials. If the rats did not find the escape platform within 60 s, the experimenter guided them to the platform where they were allowed to remain for 15 s. After the first daily trial, the animals were placed in a clean cage, and a 30-s interval was imposed before the beginning of the next trial. The platform location was not changed during the acquisition period. The swim path length, the percentage of swim time spent within 30 cm of the side walls of the pool (thigmotaxis), and the swim speed were calculated. One day after completion of the acquisition, animals were submitted to a single 60-s probe trial in which the platform was removed from the pool. The number of times the rats swam through the zone where the platform had been located (platform crossings) provided a measure of accuracy in recalling the former position of the platform. In order to assess whether the animals use a spatial strategy when searching for the escape platform, the platform crossings measured in the
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
training quadrant were compared with the number of times the animals crossed the corresponding area in the opposite quadrant. The percentage of time spent by rats swimming in the training and opposite quadrants (quadrant preference scores) were also recorded. Performance of animals on the visible platform task was assessed during a 2-day period following the day on which the probe trials were carried out. In this task, the rats were given one block of four trials per day separated by 30-s intertrial intervals. The platform, painted in white, was exposed 3 cm above the water surface. The position of the platform was different in each trial. The distances swum to locate the platform were recorded and averaged across eight trials. All the experiments were performed after at least 30-min habituation of animals to the testing room. Testing was done at the same time of day, beginning at 13:00 h. 2.3. Histology Following the completion of the experiments, animals were deeply anesthetized with sodium pentobarbital and transcardially perfused with buffered saline (pH 7.4) followed by a fixative solution (4% paraformaldehyde in phosphate buffer). The brains were removed from the skulls, codified, and placed in fresh fixative for 3 h. Subsequently, the brains were infiltrated overnight at 4 ◦ C in 10% sucrose solution. After the infiltration, they were cut at a vibratome into 40-m coronal sections. Sections were, mounted on gelatin-coated slides, stained with a Giemsa solution, dehydrated through a graded series of ethanol solutions, and cover-slipped with Histomount. The lesions were identified by missing tissue, gliosis, or obvious abnormalities in the cytoarchitectonic pattern of this region of the cortex. The rat brain atlas by Swanson [19], was used to identify the cytoarchitectonic boundaries of RC throughout its rostrocaudal extent. Behavioral data from animals were only retained for analysis when the RC lesions were bilateral, the underlying hippocampal neurons and the subicular complex were spared, and no substantial and bilateral damage to the adjacent cortical structures was found. 2.4. Statistical analysis Distances swum to find the platform and thigmotaxis scores derived from the daily trials of the place learning task, averaged over four consecutive trials each, were analyzed using repeated measures ANOVA followed by Newman–Keuls post hoc test, where appropriate. The remaining data were analyzed using Student’s t-test. All data are presented as the mean ± S.E.M. Differences were considered as significant at the P < 0.05 level.
177
ther experiment, animals were allowed to recover for at least 20 days after surgery before the beginning of behavioral testing. During the last 5 days of the recovery period, all rats were handled for 5 min per day. The above-described surgical and behavioral procedures were identical in the two experiments with one important exception. The nonspatial pretraining procedure was only employed in the second experiment, while rats from the first experiment were submitted to the place learning task without being first familiarized with the general task requirements i.e., were behaviorally naive at the beginning of testing. Initially, 10 control rats and 15 RClesioned rats were included in each experiment. However, some animals have died shortly after the surgery (one control, two RC-lesioned rats from the first experiment, and one RC-lesioned rat from the second experiment) or few weeks later (one control rat from the second experiment). Further, post-mortem morphological analysis showed that hippocampal CA1 field was lesioned unilaterally in three rats from the first experiment and bilaterally in one rat from the second experiment; given the importance of CA1 pyramidal neurons to spatial navigation, data obtained from these four animals were discarded. Thus, all results described below refer to nine control and 10 RC-lesioned rats in the first experiment, and nine control and 13 RC-lesioned rats in the second experiment. All the rats included in this study were killed at approximately 5 months of age. The handling and care of the animals were conducted according to the European Communities Council guidelines in animal research (86/609/UE).
3. Results 3.1. Histology N-Methyl-d-aspartate (NMDA)-treated rats from both experiments had reliable bilateral lesions in RC, primarily in the anterior and middle portions, with some sparing of the posterior RC (Fig. 1). Like in several previous lesion studies [9,14], the most caudal portion of RC was found to be unaffected. The infusions of excitotoxin did not cause subicular damage, which may be present in animals with the lesions centered in the caudal RC [25]. Likewise, in none of the animals did the lesion extend rostrally into the anterior cingulate cortex. In some cases, however, minor and always unilateral lesions to the neocortical areas adjacent to rostral part of RC were found. The NMDA-induced lesions of RC found in rats included in the first and second experiments were highly similar, both with respect to their general distribution pattern and severity.
2.5. General experimental design
3.2. Effect of RC lesions on the water maze performance in rats unfamiliar with general task requirements
The present study consisted of two independent experiments separated by approximately a 3-month period. In ei-
The mean distances swum by control and RC-lesioned rats to locate the hidden platform in the classical Morris place
178
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
Fig. 1. Site-specific injections of NMDA produced significant bilateral damage to the retrosplenial cortex. (A) Series of Giemsa-stained coronal sections showing representative cytotoxic lesions of the cortex. Scale bar = 600 m. (B) Higher magnification of the inset shown in panel A. Note the lack of neurons at and around the lesion site (the boundaries of the lesions are indicated by arrows). lv: lateral ventricle; cc: corpus callosum. Scale bar = 300 m. (C) Schematic representation of the retrosplenial cortex lesions summarizing the results of the post-mortem histological analysis. This image was generated by superimposing the camera lucida drawings of the lesions in series of level-matched brain sections obtained from all NMDA-infused rats that were included in the first experiment of the present study. The numbers on the right indicate distances (in mm) from bregma. The general distribution pattern and the extent of the lesions observed in rats from the first and second experiments were remarkably similar.
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
179
Table 1 Effect of bilateral cytotoxic lesions to the retrosplenial cortex on the probe trial performance in behaviorally naive and nonspatially pretrained rats Training quadrant
Opposite quadranta
Time spent (%)
Platform crossings
Time spent (%)
Platform crossings
Behaviorally naive rats Control 43.2 ± 2.9 RC-lesioned 43.5 ± 7
2.8 ± 0.4 2.6 ± 0.8
11.8 ± 1.3 9.8 ± 1.4
0.3 ± 0.2 0.4 ± 0.2
Nonspatially pretrained rats Control 51.3 ± 3.7 RC-lesioned 49.8 ± 3.3
3.9 ± 0.6 3.1 ± 0.5
10.4 ± 1.6 11.5 ± 1.4
0.9 ± 0.3 0.3 ± 0.1
a
Fig. 2. Learning the water maze task by behaviorally naive control rats (open circles) and rats with bilateral cytotoxic lesions to the retrosplenial cortex (filled circles). (A) The swim distances in centimeters to locate the submerged platform for each block of four consecutive trials (mean ± S.E.M.). Animals with lesions showed significantly impaired performance on this task relative to control rats (P < 0.05 for the main effect of treatment). (B) The percentage of time spent by rats within 30 cm of the pool walls for each block of four consecutive trials (mean ± S.E.M.). Note that animals with lesions in the cortex were less efficient than control rats in learning to swim away from the side walls of the pool (P < 0.05 for the main effect of treatment).
learning task are shown in Fig. 2A. The overall repeated measures ANOVA showed that the rats in both groups progressively improved their ability to find the platform over the 14 days of acquisition (F6,102 = 54.5, P < 0.00001). However, the animals from RC group performed the task less well than did control animals, as indicated by significant main effect of treatment (F1,17 = 4.46, P < 0.05). On the other hand, analysis of the data showed that there was no significant interaction between the lesion effect and the course of training. This lack of significant interaction effects indicates that RC-lesioned rats were equally impaired during the period of initial learning of the task and on the late phases of training, a finding which is supported by a simple visual inspection of the data. Analysis of the thigmotaxis scores (Fig. 2B) yielded similar results showing that, although the animals in both groups were able to learn to swim away from the side walls of the pool (F6,102 = 80.8, P < 0.00001), those which had received excitotoxin infusions spent, however, significantly more time than controls swimming in the periphery (F1,17 = 5.24, P < 0.05 for main effect of treatment). Again, there was no significant interaction between the lesion effect and the course of training, indicating that RC-lesioned rats were impaired relative to sham operated rats both in the initial acquisition of general behavioral strategies and during the later training stages.
P < 0.001 vs. training quadrant for either index.
Behavioral measures derived from the probe trial are summarized in Table 1. The animals from both groups used a spatial strategy when searching for the escape platform during the probe trial, as indicated by the fact that they spent significantly more time in the training quadrant than in the opposite one and crossed the former location of the platform more frequently than the corresponding area of the opposite quadrant (P < 0.001 for both parameters). However, the two probe-trial indexes that are traditionally used to ascertain the possible effects of treatments on retention, i.e., the platform crossings and the percentage of time spent swimming in the training quadrant, did not significantly differ between control and RC-lesioned groups. The swim speeds, averaged over all trials of the place learning task, were 28.7 ± 1.1 cm/s for controls and 28.1 ± 0.6 cm/s for RC-lesioned rats. The distances swum on the visible platform task, averaged across eight testing trials, were 205 ± 59 cm for controls and 242 ± 75 cm for RC-lesioned rats. No significant effects of treatment on these measures were revealed, showing that rats in either group had comparable sensorimotor abilities. 3.3. Performance of the water maze task by RC-lesioned rats following nonspatial pretraining The place learning scores derived from this experiment are shown in Fig. 3A. Similarly to the previous experiment, the animals in both groups progressively improved their ability to find the platform over the 14 days of acquisition (F6,120 = 71.1, P < 0.00001). However, with respect to the treatment effects, the results of repeated-measures ANOVA were different from those reported in the first experiment. More specifically, although the main effect of the lesion was found to be insignificant in this case, there was a significant interaction between the main effect and the course of training (F6,120 = 3.4, P < 0.005), indicating that RC-lesioned and control rats, unlike in the previous experiment, learned the task at different rates. Inspection of the data shows that the animals from RC group performed the task similarly to control animals during the first 6 days of training, but swam longer distances than did controls on the fourth trial block (P < 0.001; post hoc test for Group × Time interaction). Yet,
180
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
The swim speeds, averaged over all trials of the place learning task, were 30.2 ± 1.3 cm/s for controls and 30.1 ± 0.8 cm/s for RC-lesioned rats. The distances swum on the visible platform task, averaged across eight testing trials, were 196 ± 81 cm for controls and 221 ± 90 cm for RC-lesioned rats. No significant effects of treatment on these measures were found.
4. Discussion
Fig. 3. Learning the water maze task by control rats (open circles) and rats with bilateral cytotoxic lesions to the retrosplenial cortex (filled circles) after being first familiarized with general task requirements. (A) The swim distances in centimeters to locate the submerged platform for each block of four consecutive trials (mean ± S.E.M.). Note that the animals with lesions learned the task as well as controls during first 6 days of acquisition and by the end of training, but performed worse on the fourth trial block (P < 0.001, post hoc test for the significant Group × Time interaction). (B) The percentage of time spent by rats within 30 cm of the pool walls for each block of four consecutive trials (mean ± S.E.M). Note that lesioned animals that have been previously familiarized with general task rules did not show more thigmotaxis than did control rats.
by the end of training, their performance was again normal. Analysis of the thigmotaxis scores (Fig. 3B) showed that the animals in both groups were able to learn to swim away from the side walls of the pool (F6,120 = 101.3, P < 0.00001). However, although RC-lesioned rats had a slightly increased level of thigmotaxis on the fourth trial block, ANOVA revealed no statistically significant differences between the two treatment groups. Considered together with the results of the first experiment, these findings suggest that nonspatial pretraining was capable of considerably ameliorating behavioral impairments resulted from damage to RC. Behavioral measures derived from the probe trial are shown in Table 1. Like in the previous experiment, the animals in both groups used a spatial strategy when searching for the escape platform during the probe trial, as indicated by the fact that they spent significantly more time in the training quadrant than in the opposite one and crossed the former location of the platform more frequently than the corresponding area of the opposite quadrant (P < 0.001 for both parameters). Again, the platform crossings and the percentage of time spent swimming in the training quadrant did not significantly differ between control and RC-lesioned groups (Table 1).
The two experiments reported in this communication addressed the issue of whether or not cytotoxic damage to RC impairs water maze performance in Wistar rats and, if so, whether these behavioral alterations reflect spatial navigation deficits or are they merely attributable to impaired learning of nonspatial task requirements or both. The results of the first experiment show that behaviorally naive animals with selective, relatively moderate lesions to RC possess a reduced rate of the water maze acquisition and increased thigmotaxis, thus ruling out the possibility that some innate cognitive deficits of Wistar rats might camouflage the effect of RC lesions upon their performance on the water maze. As shown in the second experiment, nonspatial pretraining was capable of considerably ameliorating the behavioral deficits in RC-lesioned animals, namely on the early trials of the place learning task as well as by the end of the training. The possibility that the behavioral changes observed in RC-lesioned rats from either experiment were due to a non-specific damage to brain structures other than RC can safely be rejected, because post-mortem histological analysis revealed no considerable lesions of the adjacent neocortical areas, as well as of the hippocampus and subiculum. It is worth mentioning that the spatial impairments found in this study in NMDA-lesioned animals appear to be less severe than those described after aspiration RC lesions [7,18,30]. This is consistent with the previously reported observations that the behavioral effects of RC damage can vary as a function of lesion type and size [7,23,25]. The fact that the suction ablations of RC consistently produce more striking impairments of navigation when compared to excitotoxic lesions can be most plausibly explained by concomitant destruction of a diffuse pathway(s) passing through this region and involved in the transmission of information relevant for spatial learning [9]. With respect to the lesion size effect, it has been found that excitotoxic lesions of either rostral or caudal RC disrupt the acquisition of the water maze place task [25,28], whereas more extensive RC lesions, which encompass both its rostral and caudal portions, additionally reduce the retention scores derived from the probe trial, namely the training quadrant preference and the number of target crossings [23]. In the present study, the injections of NMDA produced bilateral destruction of the rostral and middle RC, while sparing its most caudal portion. Consistent with the just reviewed data on the lesion size effect, we detected no differences between the RC-lesioned and control groups in
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
either the time spent in the trained quadrant or the number of platform crossings, despite the fact that RC-lesioned rats showed obvious impairments during the task acquisition. As noted above, there is recent evidence that the extent of behavioral changes associated with RC lesions might also be influenced by the strain of rats used for experimentation [7]. In fact, it has been reported that the damage to RC severely impairs the performance of Long–Evans rats on the standard water maze place task [7,18,30], whereas it fails to do so or causes barely detectable deficits in Dark Agouti rats [14,25,28]. Given that the performance of intact Dark Agouti rats on spatial tasks was shown to be considerably inferior to that of intact Long–Evans rats [8], it has been hypothesized that, in the former strain, the deleterious effects of RC damage on spatial navigation might be masked by an innate cognitive impairment of nonspatial nature [7]. This hypothesis can logically be extended to Wistar rats, whose performance on spatial tasks is known to be quite similar to that of the Dark Agouti strain [8]. However, the results of our study do not support such an extrapolation, because we found that even moderate damage to RC induces well detectable behavioral alterations in Wistar strain. Interestingly, the authors of one of the studies that failed to find changes in the place learning of Dark Agouti rats after cytotoxic damage to RC [28], have recently re-examined their data, by applying a less sophisticated statistical design, and found that the lesion did produce significant spatial deficits [25]. This result has been repeatedly confirmed by the findings of subsequent experiments utilizing the same rat strain and the same lesion type [23,25]. Evidence consistent with this comes from another recent study demonstrating that, although Dark Agouti rats with the suction ablations of RC can indeed perform nearly normally on the place learning task, they are, however, severely impaired on the matching-to-place task [7]. Thus, the results of the studies performed on three different rat strains, i.e., Long–Evans [7,9,18,30], Dark Agouti [23,25,28], and Wistar (this study), agree that RC-lesioned animals possess impaired performance on spatial learning tasks. During the initial phase of the acquisition of the water maze task, animals are required to learn that the pool wall, albeit readily visible, does not afford the opportunity to escape from the cold water and that a small invisible platform, which is typically placed some distance away from the edge of the maze, instead does so. Hence, the amount of time spent by rats swimming along the pool wall (thigmotaxis), which gradually decreases over the course of training, can plausibly be considered as a measure of the animal’s ability to adopt a set of general rules required to normally execute the water maze task [3,15,29]. In the first experiment of the present study, the animals from RC group showed enhanced thigmotaxis compared with sham-operated group, which indicates that the lesion was associated with considerable abnormalities in nonspatial learning. The issue of whether the poor performance of RC-lesioned rats on navigation tasks is attributable to these non-specific cognitive deficits or rather reflects true spatial impairments has been previously addressed
181
by Harker and Whishaw [7], who found that pre-surgical familiarization of animals with general task requirements does not erase the lesion-induced changes in place learning. While this observation removes doubts regarding the spatial nature of behavioral deficits seen in RC-lesioned rats, it does not rule out the possibility that these deficits are of more complex origin, being partially related to non-specific cognitive disabilities. This latter possibility is supported by the results of our second experiment which show that nonspatial pretraining of the animals that have already been subjected to the surgery dramatically reduces the effect of the lesion on both thigmotactic swimming and rate of initial learning. In fact, RC-lesioned rats were still unable to perform the task as well as controls only on the fourth trial block, but showed normal initial learning and normal asymptotic performance. The finding that, despite being familiarized with nonspatial task requirements, RC-lesioned rats swam longer distances than did controls on the fourth trial block may be accounted for by the deleterious effect of the lesion on the processing of spatial information. Alternatively, it is also possible that the animals were still impaired at this point because they were insufficiently familiarized with the general task requirements on the shaping trials (although not at statistically significant level, RC-lesioned rats exhibited more periphery swimming than control rats on the fourth trial block). In addition, the significant Group × Trial block interaction detected in the second experiment can be attributed simply to the unusual behavior of control rats who showed, unlike in the first experiment, unexpectedly great improvement from trial block 3 to trial block 4. Whatever the case, the finding that nonspatial pretraining almost completely normalized the performance of RC-lesioned rats on the place learning task implies that, in addition to spatial navigation, RC may also be involved in cognitive processes underlying the ability of rats to properly respond to general task demands. In summary, the results of the present study confirm and extend the findings of several previous studies by demonstrating that bilateral damage to RC significantly impairs the performance of Wistar rats on the place learning version of the water maze task. The data herein reported additionally show that, at least in this strain of rats, nonspatial pretraining given after RC lesions is capable of considerably mitigating these behavioral deficits.
Acknowledgements This work was supported by Fundac¸a˜ o para a Ciˆencia e a Tecnologia, Grant POCTI/NSE/42794/2001, and Centro de Morfologia Experimental (POCTI/SFA/121).
References [1] Aggleton JP, Brown MW. Episodic memory, amnesia, and the hippocampal–anterior thalamic axis. Behav Brain Sci 1999;22:425–89.
182
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
[2] Aggleton JP, Pearce JM. Neural systems underlying episodic memory: insights from animal research. Philos Trans R Soc Lond B Biol Sci 2001;356:1467–82. [3] Cain DP, Saucier D. The neuroscience of spatial navigation: focus on behavior yields advances. Rev Neurosci 1996;7:215–31. [4] Cooper BG, Mizumori SJ. Retrosplenial cortex inactivation selectively impairs navigation in darkness. Neuroreport 1999;10:625–30. [5] Cooper BG, Mizumori SJ. Temporary inactivation of the retrosplenial cortex causes a transient reorganization of spatial coding in the hippocampus. J Neurosci 2001;21:3986–4001. [6] Gabriel M, Smith DM. What does the limbic memory circuit actually do? Behav Brain Sci 1999;22:451. [7] Harker KT, Whishaw IQ. Impaired spatial performance in rats with retrosplenial lesions: importance of the spatial problem and the rat strain in identifying lesion effects in a swimming pool. J Neurosci 2002;22:1155–64. [8] Harker KT, Whishaw IQ. Place and matching-to-place spatial learning affected by rat inbreeding (Dark Agouti, Fischer 344) and albinism (Wistar, Sprague–Dawley) but not domestication (wild rat vs. Long–Evans, Fischer–Norway). Behav Brain Res 2002;134:467–77. [9] Harker KT, Whishaw IQ. Impaired place navigation in place and matching-to-place swimming pool tasks follows both retrosplenial cortex lesions and cingulum bundle lesions in rats. Hippocampus 2004;14:224–31. [10] Kang E, Gabriel M. Hippocampal modulation of cingulo-thalamic neuronal activity and discriminative avoidance learning in rabbits. Hippocampus 1998;8:491–510. [11] Lukoyanov NV, Sa MJ, Madeira MD, Paula-Barbosa MM. Selective loss of hilar neurons and impairment of initial learning in rats after repeated administration of electroconvulsive shock seizures. Exp Brain Res 2004;154:192–200. [12] Meunier M, Jaffard R, Destrade C. Differential involvement of anterior and posterior cingulate cortices in spatial discriminative learning in a T-maze in mice. Behav Brain Res 1991;44:133–43. [13] Morris R. Developments of a water–maze procedure for studying spatial learning in the rat. J Neurosci Methods 1984;11:47–60. [14] Neave N, Lloyd S, Sahgal A, Aggleton JP. Lack of effect of lesions in the anterior cingulate cortex and retrosplenial cortex on certain tests of spatial memory in the rat. Behav Brain Res 1994;65:89–101. [15] Saucier D, Hargreaves EL, Boon F, Vanderwolf CH, Cain DP. Detailed behavioral analysis of water maze acquisition under systemic NMDA or muscarinic antagonism: nonspatial pretraining eliminates spatial learning deficits. Behav Neurosci 1996;110:103–16.
[16] Shibata H. Efferent projections from the anterior thalamic nuclei to the cingulate cortex in the rat. J Comp Neurol 1993;330:533– 42. [17] Shibata H. Organization of projections of rat retrosplenial cortex to the anterior thalamic nuclei. Eur J Neurosci 1998;10:3210–9. [18] Sutherland RJ, Whishaw IQ, Kolb B. Contributions of cingulate cortex to two forms of spatial learning and memory. J Neurosci 1988;8:1863–72. [19] Swanson LW. Brain maps: structure of the rat brain. Amsterdam: Elsevier; 1992. [20] Van Groen T, Wyss JM. Connections of the retrosplenial granular a cortex in the rat. J Comp Neurol 1990;300:593–606. [21] Van Groen T, Wyss JM. Connections of the retrosplenial dysgranular cortex in the rat. J Comp Neurol 1992;315:200–16. [22] Van Groen T, Wyss JM. Connections of the retrosplenial granular b cortex in the rat. J Comp Neurol 2003;463:249–63. [23] Vann SD, Aggleton JP. Extensive cytotoxic lesions of the rat retrosplenial cortex reveal consistent deficits on tasks that tax allocentric spatial memory. Behav Neurosci 2002;116:85–94. [24] Vann SD, Brown MW, Aggleton JP. Fos expression in the rostral thalamic nuclei and associated cortical regions in response to different spatial memory tests. Neuroscience 2000;101:983–91. [25] Vann SD, Kristina Wilton LA, Muir JL, Aggleton JP. Testing the importance of the caudal retrosplenial cortex for spatial memory in rats. Behav Brain Res 2003;140:107–18. [26] Vogt BA, Miller MW. Cortical connections between rat cingulate cortex and visual, motor, and postsubicular cortices. J Comp Neurol 1983;216:192–210. [27] Wang B, Gonzalo-Ruiz A, Morte L, Campbell G, Lieberman AR. Immunoelectron microscopic study of glutamate inputs from the retrosplenial granular cortex to identified thalamocortical projection neurons in the anterior thalamus of the rat. Brain Res Bull 1999;50:63–76. [28] Warburton EC, Aggleton JP, Muir JL. Comparing the effects of selective cingulate cortex lesions and cingulum bundle lesions on water maze performance by rats. Eur J Neurosci 1998;10:622– 34. [29] Whishaw IQ. Cholinergic receptor blockade in the rat impairs locale but not taxon strategies for place navigation in a swimming pool. Behav Neurosci 1985;99:979–1005. [30] Whishaw IQ, Maaswinkel H, Gonzalez CL, Kolb B. Deficits in allothetic and idiothetic spatial behavior in rats with posterior cingulate cortex lesions. Behav Brain Res 2001;118:67–76.
Research report
Impaired water maze navigation of Wistar rats with retrosplenial cortex lesions: effect of nonspatial pretraining Nikolai V. Lukoyanov∗ , Elena A. Lukoyanova, Jos´e P. Andrade, Manuel M. Paula-Barbosa Department of Anatomy, Porto Medical School, 4200-319 Porto, Portugal Received 2 July 2004; received in revised form 26 August 2004; accepted 30 August 2004 Available online 6 October 2004
Abstract Damage to the retrosplenial cortex (RC) impairs the performance of rodents on spatial learning and memory tasks, but the extent of these deficits was previously reported to be influenced by the lesion type, rat strain, and behavioral task used. The present study addressed the issue of whether or not cytotoxic damage to RC impairs place navigation of Wistar rats in the Morris water maze and, if so, whether this is merely attributable to spatial learning deficits or to impaired learning of general (nonspatial) behavioral strategies required to correctly perform this task or both. Behaviorally naive rats with bilateral lesions to RC were significantly impaired relative to sham-lesioned rats both during the period of initial learning of the task and during the later phases of training. In addition, these animals showed enhanced thigmotaxis, indicating that the lesion was associated with considerable abnormalities in nonspatial learning. In contrast, RC-lesioned animals that have been previously familiarized with general task rules in a series of shaping trials did not show more thigmotaxis than did their respective controls. Furthermore, although these rats were still impaired in the middle of the training process, their performance during the period of initial learning as well as by the end of training was found to now be normal. Our results confirm those of earlier studies indicating that RC is important for spatial navigation. The findings herein reported are also consistent with the notion that, in addition to spatial information processing, RC is involved in cognitive processes underlying the ability of subjects to properly respond to general task demands. © 2004 Elsevier B.V. All rights reserved. Keywords: Retrosplenial cortex; Lesion; Learning; Spatial navigation; Water maze; Nonspatial pretraining
1. Introduction The retrosplenial cortex (RC) is reciprocally connected with the hippocampal formation and anterior thalamic nuclei [16,17,20–22,26,27], the key elements of the brain circuit that mediates memory formation [1,2,6,10]. That RC constitutes a part of this circuit is supported by evidence from behavioral studies in rodents indicating that damage to this area of cortex impairs spatial learning, which is known to be particularly dependent on the hippocampal formation [7,9,12,18,23,24,30]. The finding that inactivation of RC also impairs navigation in the darkness and alters the firing pattern of hippocampal place cells suggests that RC specifically contribute to spatial
∗
Corresponding author. Tel.: +351 22 509 6808; fax: +351 22 550 5640. E-mail address: [email protected] (N.V. Lukoyanov).
0166-4328/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbr.2004.08.023
behavior by integrating visual cues with information generated by self-motion [4,5]. Although the involvement of RC in spatial navigation is well documented, the results of the previous studies on RClesioned animals are not entirely uniform, particularly with respect to the severity of behavioral changes observed. More specifically, it was found that suction ablations of RC [7,9,30] generally affect spatial abilities of rats to a greater extent than do excitotoxic lesions [9,24,25], suggesting that, in the former case, damage to fibers of passage can partially account for behavioral deficits. It has also been shown that the extent of the spatial impairments is a determinant of whether or not the lesion involves the most caudal portion of RC [23]. Making the picture even more complicated, it has been reported that removal of RC produces robust impairments of spatial navigation in the Long–Evans strain of rats [7,30], while it causes relatively little or no place learning deficits
176
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
in Dark Agouti rats [14,28]. With respect to the latter variable, it has been suggested that the effect of RC lesions on spatial navigation of Dark Agouti rats might be masked by an innate nonspatial impairment, which is characteristic of this strain [7]. Supporting this argument is the observation that intact Dark Agouti rats are impaired, compared to intact Long–Evans rats, both in overall performance on the water maze and in learning of general task requirements [8]. However, the hypothesis that the spatial navigation deficits associated with damage to RC can be masked by some nonspatial impairments remains to be further verified. In the present study, we addressed this issue by assessing the behavioral effects of RC lesions in Wistar rats, which were previously shown to perform spatial tasks at a level comparable with that of Dark Agouti rats [8]. The behavioral assessments were performed using the Morris water maze, a paradigm on which Dark Agouti rats with RC lesions display little or no impairments. Two separate experiments were performed: (1) in behaviorally naive rats and (2) using animals that received nonspatial pretraining before being tested on the place learning task. The latter experiment was intended to unmask the lesion-induced spatial navigation deficits from those related to nonspatial learning.
2. Methods 2.1. Animals and surgery Male Wistar rats, derived from animals obtained from Gulbenkian Institute of Science (Oeiras, Portugal), were used in the present study. Animals were housed two to three per cage, maintained under standard laboratory conditions (20–22 ◦ C and a 12 h:12 h light–dark cycle), and had free access to food and water. At 12 weeks of age, they were subjected either to cytotoxic lesions of RC or to a sham surgical procedure. Rats were anesthetized by subsequent injections of promethazine (10 mg/kg), xylazine (2.6 mg/kg), and ketamine (50 mg/kg), and placed in a Kopf stereotaxic apparatus. Under aseptic conditions, the scalp was incised along the midline and retracted to the side. To expose the brain surface, a piece of parietal bone, 1.8 mm wide and 5 mm long, was removed on either side of the midline using a small drill mounted to the stereotaxic apparatus. However, a narrow bar of bone over the superior sagittal sinus, approximately 1.4 mm wide, was left intact. Bilateral cytotoxic lesions of RC were made by infusing 0.3 l of N-methyl-d-aspartate (NMDA, 0.09 M saline solution) at the following coordinates: AP (from bregma) −2.9, ML (from midline) ±0.8, DV (from dura) −1.8; AP −4.4, ML ±0.8, DV −1.7; AP −5.9, ML ±1.0, DV −2.0. All infusions were delivered over a 3-min period using a 1-l Hamilton syringe equipped with a 33 gauge blunt-tip needle. Five minutes of diffusion time were allowed before the needle was retracted. Sham control rats underwent an identical procedure but received vehicle alone rather than NMDA solution.
2.2. Behavioral procedures The behavioral consequences of RC lesions were evaluated using the Morris water maze paradigm. The maze consisted of a black circular tank, 180 cm in diameter and 50 cm deep, and was located in a corner of a room containing extramaze cues, i.e., three posters of different size and shape, and a computer desk. The apparatus was filled with water at room temperature (21 ± 1 ◦ C) to a depth of approximately 35 cm. The water was made opaque by adding a non-toxic paint. The maze was divided, by imaginary lines, into four equal-size quadrants. The swim path was recorded by a computerized video-tracking system (EthoVision V3.0, Noldus, The Netherlands). All rats included in the second experiment were familiarized with the general task requirements in two sessions of shaping trials in which no spatial learning was required [11]. In this procedure, a black acrylic alley, measuring 130 cm × 65 cm × 15 cm, was placed on the bottom of the maze so that its walls extended 30 cm above the surface of the water. A black escape platform, 10 cm in diameter, was placed inside the alley, approximately 25 cm from one of its ends. It was located 2 cm below the surface of the water. Each rat was placed in the water and allowed to swim to the platform and to climb on it. If the rats did not find the escape platform within 20 s, the experimenter guided them to the platform. Each session consisted of three trials, spaced by 15-s inter-trial intervals, in which the rats were released in the water from three different locations, i.e., 30, 60 and 90 cm from the position of the platform. The interval between the two shaping sessions was 24 h. No extramaze cues were present during this phase of training. In the classical place learning task [13], the animals were trained to find the submerged escape platform, which was now located in the center of one of the quadrants. For acquisition, rats were given two trials on each day for 14 consecutive days. Each rat was placed in the water facing the pool wall at one of the four starting points that were used in a pseudo-random order so that each position was used once in each block of four trials. If the rats did not find the escape platform within 60 s, the experimenter guided them to the platform where they were allowed to remain for 15 s. After the first daily trial, the animals were placed in a clean cage, and a 30-s interval was imposed before the beginning of the next trial. The platform location was not changed during the acquisition period. The swim path length, the percentage of swim time spent within 30 cm of the side walls of the pool (thigmotaxis), and the swim speed were calculated. One day after completion of the acquisition, animals were submitted to a single 60-s probe trial in which the platform was removed from the pool. The number of times the rats swam through the zone where the platform had been located (platform crossings) provided a measure of accuracy in recalling the former position of the platform. In order to assess whether the animals use a spatial strategy when searching for the escape platform, the platform crossings measured in the
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
training quadrant were compared with the number of times the animals crossed the corresponding area in the opposite quadrant. The percentage of time spent by rats swimming in the training and opposite quadrants (quadrant preference scores) were also recorded. Performance of animals on the visible platform task was assessed during a 2-day period following the day on which the probe trials were carried out. In this task, the rats were given one block of four trials per day separated by 30-s intertrial intervals. The platform, painted in white, was exposed 3 cm above the water surface. The position of the platform was different in each trial. The distances swum to locate the platform were recorded and averaged across eight trials. All the experiments were performed after at least 30-min habituation of animals to the testing room. Testing was done at the same time of day, beginning at 13:00 h. 2.3. Histology Following the completion of the experiments, animals were deeply anesthetized with sodium pentobarbital and transcardially perfused with buffered saline (pH 7.4) followed by a fixative solution (4% paraformaldehyde in phosphate buffer). The brains were removed from the skulls, codified, and placed in fresh fixative for 3 h. Subsequently, the brains were infiltrated overnight at 4 ◦ C in 10% sucrose solution. After the infiltration, they were cut at a vibratome into 40-m coronal sections. Sections were, mounted on gelatin-coated slides, stained with a Giemsa solution, dehydrated through a graded series of ethanol solutions, and cover-slipped with Histomount. The lesions were identified by missing tissue, gliosis, or obvious abnormalities in the cytoarchitectonic pattern of this region of the cortex. The rat brain atlas by Swanson [19], was used to identify the cytoarchitectonic boundaries of RC throughout its rostrocaudal extent. Behavioral data from animals were only retained for analysis when the RC lesions were bilateral, the underlying hippocampal neurons and the subicular complex were spared, and no substantial and bilateral damage to the adjacent cortical structures was found. 2.4. Statistical analysis Distances swum to find the platform and thigmotaxis scores derived from the daily trials of the place learning task, averaged over four consecutive trials each, were analyzed using repeated measures ANOVA followed by Newman–Keuls post hoc test, where appropriate. The remaining data were analyzed using Student’s t-test. All data are presented as the mean ± S.E.M. Differences were considered as significant at the P < 0.05 level.
177
ther experiment, animals were allowed to recover for at least 20 days after surgery before the beginning of behavioral testing. During the last 5 days of the recovery period, all rats were handled for 5 min per day. The above-described surgical and behavioral procedures were identical in the two experiments with one important exception. The nonspatial pretraining procedure was only employed in the second experiment, while rats from the first experiment were submitted to the place learning task without being first familiarized with the general task requirements i.e., were behaviorally naive at the beginning of testing. Initially, 10 control rats and 15 RClesioned rats were included in each experiment. However, some animals have died shortly after the surgery (one control, two RC-lesioned rats from the first experiment, and one RC-lesioned rat from the second experiment) or few weeks later (one control rat from the second experiment). Further, post-mortem morphological analysis showed that hippocampal CA1 field was lesioned unilaterally in three rats from the first experiment and bilaterally in one rat from the second experiment; given the importance of CA1 pyramidal neurons to spatial navigation, data obtained from these four animals were discarded. Thus, all results described below refer to nine control and 10 RC-lesioned rats in the first experiment, and nine control and 13 RC-lesioned rats in the second experiment. All the rats included in this study were killed at approximately 5 months of age. The handling and care of the animals were conducted according to the European Communities Council guidelines in animal research (86/609/UE).
3. Results 3.1. Histology N-Methyl-d-aspartate (NMDA)-treated rats from both experiments had reliable bilateral lesions in RC, primarily in the anterior and middle portions, with some sparing of the posterior RC (Fig. 1). Like in several previous lesion studies [9,14], the most caudal portion of RC was found to be unaffected. The infusions of excitotoxin did not cause subicular damage, which may be present in animals with the lesions centered in the caudal RC [25]. Likewise, in none of the animals did the lesion extend rostrally into the anterior cingulate cortex. In some cases, however, minor and always unilateral lesions to the neocortical areas adjacent to rostral part of RC were found. The NMDA-induced lesions of RC found in rats included in the first and second experiments were highly similar, both with respect to their general distribution pattern and severity.
2.5. General experimental design
3.2. Effect of RC lesions on the water maze performance in rats unfamiliar with general task requirements
The present study consisted of two independent experiments separated by approximately a 3-month period. In ei-
The mean distances swum by control and RC-lesioned rats to locate the hidden platform in the classical Morris place
178
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
Fig. 1. Site-specific injections of NMDA produced significant bilateral damage to the retrosplenial cortex. (A) Series of Giemsa-stained coronal sections showing representative cytotoxic lesions of the cortex. Scale bar = 600 m. (B) Higher magnification of the inset shown in panel A. Note the lack of neurons at and around the lesion site (the boundaries of the lesions are indicated by arrows). lv: lateral ventricle; cc: corpus callosum. Scale bar = 300 m. (C) Schematic representation of the retrosplenial cortex lesions summarizing the results of the post-mortem histological analysis. This image was generated by superimposing the camera lucida drawings of the lesions in series of level-matched brain sections obtained from all NMDA-infused rats that were included in the first experiment of the present study. The numbers on the right indicate distances (in mm) from bregma. The general distribution pattern and the extent of the lesions observed in rats from the first and second experiments were remarkably similar.
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
179
Table 1 Effect of bilateral cytotoxic lesions to the retrosplenial cortex on the probe trial performance in behaviorally naive and nonspatially pretrained rats Training quadrant
Opposite quadranta
Time spent (%)
Platform crossings
Time spent (%)
Platform crossings
Behaviorally naive rats Control 43.2 ± 2.9 RC-lesioned 43.5 ± 7
2.8 ± 0.4 2.6 ± 0.8
11.8 ± 1.3 9.8 ± 1.4
0.3 ± 0.2 0.4 ± 0.2
Nonspatially pretrained rats Control 51.3 ± 3.7 RC-lesioned 49.8 ± 3.3
3.9 ± 0.6 3.1 ± 0.5
10.4 ± 1.6 11.5 ± 1.4
0.9 ± 0.3 0.3 ± 0.1
a
Fig. 2. Learning the water maze task by behaviorally naive control rats (open circles) and rats with bilateral cytotoxic lesions to the retrosplenial cortex (filled circles). (A) The swim distances in centimeters to locate the submerged platform for each block of four consecutive trials (mean ± S.E.M.). Animals with lesions showed significantly impaired performance on this task relative to control rats (P < 0.05 for the main effect of treatment). (B) The percentage of time spent by rats within 30 cm of the pool walls for each block of four consecutive trials (mean ± S.E.M.). Note that animals with lesions in the cortex were less efficient than control rats in learning to swim away from the side walls of the pool (P < 0.05 for the main effect of treatment).
learning task are shown in Fig. 2A. The overall repeated measures ANOVA showed that the rats in both groups progressively improved their ability to find the platform over the 14 days of acquisition (F6,102 = 54.5, P < 0.00001). However, the animals from RC group performed the task less well than did control animals, as indicated by significant main effect of treatment (F1,17 = 4.46, P < 0.05). On the other hand, analysis of the data showed that there was no significant interaction between the lesion effect and the course of training. This lack of significant interaction effects indicates that RC-lesioned rats were equally impaired during the period of initial learning of the task and on the late phases of training, a finding which is supported by a simple visual inspection of the data. Analysis of the thigmotaxis scores (Fig. 2B) yielded similar results showing that, although the animals in both groups were able to learn to swim away from the side walls of the pool (F6,102 = 80.8, P < 0.00001), those which had received excitotoxin infusions spent, however, significantly more time than controls swimming in the periphery (F1,17 = 5.24, P < 0.05 for main effect of treatment). Again, there was no significant interaction between the lesion effect and the course of training, indicating that RC-lesioned rats were impaired relative to sham operated rats both in the initial acquisition of general behavioral strategies and during the later training stages.
P < 0.001 vs. training quadrant for either index.
Behavioral measures derived from the probe trial are summarized in Table 1. The animals from both groups used a spatial strategy when searching for the escape platform during the probe trial, as indicated by the fact that they spent significantly more time in the training quadrant than in the opposite one and crossed the former location of the platform more frequently than the corresponding area of the opposite quadrant (P < 0.001 for both parameters). However, the two probe-trial indexes that are traditionally used to ascertain the possible effects of treatments on retention, i.e., the platform crossings and the percentage of time spent swimming in the training quadrant, did not significantly differ between control and RC-lesioned groups. The swim speeds, averaged over all trials of the place learning task, were 28.7 ± 1.1 cm/s for controls and 28.1 ± 0.6 cm/s for RC-lesioned rats. The distances swum on the visible platform task, averaged across eight testing trials, were 205 ± 59 cm for controls and 242 ± 75 cm for RC-lesioned rats. No significant effects of treatment on these measures were revealed, showing that rats in either group had comparable sensorimotor abilities. 3.3. Performance of the water maze task by RC-lesioned rats following nonspatial pretraining The place learning scores derived from this experiment are shown in Fig. 3A. Similarly to the previous experiment, the animals in both groups progressively improved their ability to find the platform over the 14 days of acquisition (F6,120 = 71.1, P < 0.00001). However, with respect to the treatment effects, the results of repeated-measures ANOVA were different from those reported in the first experiment. More specifically, although the main effect of the lesion was found to be insignificant in this case, there was a significant interaction between the main effect and the course of training (F6,120 = 3.4, P < 0.005), indicating that RC-lesioned and control rats, unlike in the previous experiment, learned the task at different rates. Inspection of the data shows that the animals from RC group performed the task similarly to control animals during the first 6 days of training, but swam longer distances than did controls on the fourth trial block (P < 0.001; post hoc test for Group × Time interaction). Yet,
180
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
The swim speeds, averaged over all trials of the place learning task, were 30.2 ± 1.3 cm/s for controls and 30.1 ± 0.8 cm/s for RC-lesioned rats. The distances swum on the visible platform task, averaged across eight testing trials, were 196 ± 81 cm for controls and 221 ± 90 cm for RC-lesioned rats. No significant effects of treatment on these measures were found.
4. Discussion
Fig. 3. Learning the water maze task by control rats (open circles) and rats with bilateral cytotoxic lesions to the retrosplenial cortex (filled circles) after being first familiarized with general task requirements. (A) The swim distances in centimeters to locate the submerged platform for each block of four consecutive trials (mean ± S.E.M.). Note that the animals with lesions learned the task as well as controls during first 6 days of acquisition and by the end of training, but performed worse on the fourth trial block (P < 0.001, post hoc test for the significant Group × Time interaction). (B) The percentage of time spent by rats within 30 cm of the pool walls for each block of four consecutive trials (mean ± S.E.M). Note that lesioned animals that have been previously familiarized with general task rules did not show more thigmotaxis than did control rats.
by the end of training, their performance was again normal. Analysis of the thigmotaxis scores (Fig. 3B) showed that the animals in both groups were able to learn to swim away from the side walls of the pool (F6,120 = 101.3, P < 0.00001). However, although RC-lesioned rats had a slightly increased level of thigmotaxis on the fourth trial block, ANOVA revealed no statistically significant differences between the two treatment groups. Considered together with the results of the first experiment, these findings suggest that nonspatial pretraining was capable of considerably ameliorating behavioral impairments resulted from damage to RC. Behavioral measures derived from the probe trial are shown in Table 1. Like in the previous experiment, the animals in both groups used a spatial strategy when searching for the escape platform during the probe trial, as indicated by the fact that they spent significantly more time in the training quadrant than in the opposite one and crossed the former location of the platform more frequently than the corresponding area of the opposite quadrant (P < 0.001 for both parameters). Again, the platform crossings and the percentage of time spent swimming in the training quadrant did not significantly differ between control and RC-lesioned groups (Table 1).
The two experiments reported in this communication addressed the issue of whether or not cytotoxic damage to RC impairs water maze performance in Wistar rats and, if so, whether these behavioral alterations reflect spatial navigation deficits or are they merely attributable to impaired learning of nonspatial task requirements or both. The results of the first experiment show that behaviorally naive animals with selective, relatively moderate lesions to RC possess a reduced rate of the water maze acquisition and increased thigmotaxis, thus ruling out the possibility that some innate cognitive deficits of Wistar rats might camouflage the effect of RC lesions upon their performance on the water maze. As shown in the second experiment, nonspatial pretraining was capable of considerably ameliorating the behavioral deficits in RC-lesioned animals, namely on the early trials of the place learning task as well as by the end of the training. The possibility that the behavioral changes observed in RC-lesioned rats from either experiment were due to a non-specific damage to brain structures other than RC can safely be rejected, because post-mortem histological analysis revealed no considerable lesions of the adjacent neocortical areas, as well as of the hippocampus and subiculum. It is worth mentioning that the spatial impairments found in this study in NMDA-lesioned animals appear to be less severe than those described after aspiration RC lesions [7,18,30]. This is consistent with the previously reported observations that the behavioral effects of RC damage can vary as a function of lesion type and size [7,23,25]. The fact that the suction ablations of RC consistently produce more striking impairments of navigation when compared to excitotoxic lesions can be most plausibly explained by concomitant destruction of a diffuse pathway(s) passing through this region and involved in the transmission of information relevant for spatial learning [9]. With respect to the lesion size effect, it has been found that excitotoxic lesions of either rostral or caudal RC disrupt the acquisition of the water maze place task [25,28], whereas more extensive RC lesions, which encompass both its rostral and caudal portions, additionally reduce the retention scores derived from the probe trial, namely the training quadrant preference and the number of target crossings [23]. In the present study, the injections of NMDA produced bilateral destruction of the rostral and middle RC, while sparing its most caudal portion. Consistent with the just reviewed data on the lesion size effect, we detected no differences between the RC-lesioned and control groups in
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
either the time spent in the trained quadrant or the number of platform crossings, despite the fact that RC-lesioned rats showed obvious impairments during the task acquisition. As noted above, there is recent evidence that the extent of behavioral changes associated with RC lesions might also be influenced by the strain of rats used for experimentation [7]. In fact, it has been reported that the damage to RC severely impairs the performance of Long–Evans rats on the standard water maze place task [7,18,30], whereas it fails to do so or causes barely detectable deficits in Dark Agouti rats [14,25,28]. Given that the performance of intact Dark Agouti rats on spatial tasks was shown to be considerably inferior to that of intact Long–Evans rats [8], it has been hypothesized that, in the former strain, the deleterious effects of RC damage on spatial navigation might be masked by an innate cognitive impairment of nonspatial nature [7]. This hypothesis can logically be extended to Wistar rats, whose performance on spatial tasks is known to be quite similar to that of the Dark Agouti strain [8]. However, the results of our study do not support such an extrapolation, because we found that even moderate damage to RC induces well detectable behavioral alterations in Wistar strain. Interestingly, the authors of one of the studies that failed to find changes in the place learning of Dark Agouti rats after cytotoxic damage to RC [28], have recently re-examined their data, by applying a less sophisticated statistical design, and found that the lesion did produce significant spatial deficits [25]. This result has been repeatedly confirmed by the findings of subsequent experiments utilizing the same rat strain and the same lesion type [23,25]. Evidence consistent with this comes from another recent study demonstrating that, although Dark Agouti rats with the suction ablations of RC can indeed perform nearly normally on the place learning task, they are, however, severely impaired on the matching-to-place task [7]. Thus, the results of the studies performed on three different rat strains, i.e., Long–Evans [7,9,18,30], Dark Agouti [23,25,28], and Wistar (this study), agree that RC-lesioned animals possess impaired performance on spatial learning tasks. During the initial phase of the acquisition of the water maze task, animals are required to learn that the pool wall, albeit readily visible, does not afford the opportunity to escape from the cold water and that a small invisible platform, which is typically placed some distance away from the edge of the maze, instead does so. Hence, the amount of time spent by rats swimming along the pool wall (thigmotaxis), which gradually decreases over the course of training, can plausibly be considered as a measure of the animal’s ability to adopt a set of general rules required to normally execute the water maze task [3,15,29]. In the first experiment of the present study, the animals from RC group showed enhanced thigmotaxis compared with sham-operated group, which indicates that the lesion was associated with considerable abnormalities in nonspatial learning. The issue of whether the poor performance of RC-lesioned rats on navigation tasks is attributable to these non-specific cognitive deficits or rather reflects true spatial impairments has been previously addressed
181
by Harker and Whishaw [7], who found that pre-surgical familiarization of animals with general task requirements does not erase the lesion-induced changes in place learning. While this observation removes doubts regarding the spatial nature of behavioral deficits seen in RC-lesioned rats, it does not rule out the possibility that these deficits are of more complex origin, being partially related to non-specific cognitive disabilities. This latter possibility is supported by the results of our second experiment which show that nonspatial pretraining of the animals that have already been subjected to the surgery dramatically reduces the effect of the lesion on both thigmotactic swimming and rate of initial learning. In fact, RC-lesioned rats were still unable to perform the task as well as controls only on the fourth trial block, but showed normal initial learning and normal asymptotic performance. The finding that, despite being familiarized with nonspatial task requirements, RC-lesioned rats swam longer distances than did controls on the fourth trial block may be accounted for by the deleterious effect of the lesion on the processing of spatial information. Alternatively, it is also possible that the animals were still impaired at this point because they were insufficiently familiarized with the general task requirements on the shaping trials (although not at statistically significant level, RC-lesioned rats exhibited more periphery swimming than control rats on the fourth trial block). In addition, the significant Group × Trial block interaction detected in the second experiment can be attributed simply to the unusual behavior of control rats who showed, unlike in the first experiment, unexpectedly great improvement from trial block 3 to trial block 4. Whatever the case, the finding that nonspatial pretraining almost completely normalized the performance of RC-lesioned rats on the place learning task implies that, in addition to spatial navigation, RC may also be involved in cognitive processes underlying the ability of rats to properly respond to general task demands. In summary, the results of the present study confirm and extend the findings of several previous studies by demonstrating that bilateral damage to RC significantly impairs the performance of Wistar rats on the place learning version of the water maze task. The data herein reported additionally show that, at least in this strain of rats, nonspatial pretraining given after RC lesions is capable of considerably mitigating these behavioral deficits.
Acknowledgements This work was supported by Fundac¸a˜ o para a Ciˆencia e a Tecnologia, Grant POCTI/NSE/42794/2001, and Centro de Morfologia Experimental (POCTI/SFA/121).
References [1] Aggleton JP, Brown MW. Episodic memory, amnesia, and the hippocampal–anterior thalamic axis. Behav Brain Sci 1999;22:425–89.
182
N.V. Lukoyanov et al. / Behavioural Brain Research 158 (2005) 175–182
[2] Aggleton JP, Pearce JM. Neural systems underlying episodic memory: insights from animal research. Philos Trans R Soc Lond B Biol Sci 2001;356:1467–82. [3] Cain DP, Saucier D. The neuroscience of spatial navigation: focus on behavior yields advances. Rev Neurosci 1996;7:215–31. [4] Cooper BG, Mizumori SJ. Retrosplenial cortex inactivation selectively impairs navigation in darkness. Neuroreport 1999;10:625–30. [5] Cooper BG, Mizumori SJ. Temporary inactivation of the retrosplenial cortex causes a transient reorganization of spatial coding in the hippocampus. J Neurosci 2001;21:3986–4001. [6] Gabriel M, Smith DM. What does the limbic memory circuit actually do? Behav Brain Sci 1999;22:451. [7] Harker KT, Whishaw IQ. Impaired spatial performance in rats with retrosplenial lesions: importance of the spatial problem and the rat strain in identifying lesion effects in a swimming pool. J Neurosci 2002;22:1155–64. [8] Harker KT, Whishaw IQ. Place and matching-to-place spatial learning affected by rat inbreeding (Dark Agouti, Fischer 344) and albinism (Wistar, Sprague–Dawley) but not domestication (wild rat vs. Long–Evans, Fischer–Norway). Behav Brain Res 2002;134:467–77. [9] Harker KT, Whishaw IQ. Impaired place navigation in place and matching-to-place swimming pool tasks follows both retrosplenial cortex lesions and cingulum bundle lesions in rats. Hippocampus 2004;14:224–31. [10] Kang E, Gabriel M. Hippocampal modulation of cingulo-thalamic neuronal activity and discriminative avoidance learning in rabbits. Hippocampus 1998;8:491–510. [11] Lukoyanov NV, Sa MJ, Madeira MD, Paula-Barbosa MM. Selective loss of hilar neurons and impairment of initial learning in rats after repeated administration of electroconvulsive shock seizures. Exp Brain Res 2004;154:192–200. [12] Meunier M, Jaffard R, Destrade C. Differential involvement of anterior and posterior cingulate cortices in spatial discriminative learning in a T-maze in mice. Behav Brain Res 1991;44:133–43. [13] Morris R. Developments of a water–maze procedure for studying spatial learning in the rat. J Neurosci Methods 1984;11:47–60. [14] Neave N, Lloyd S, Sahgal A, Aggleton JP. Lack of effect of lesions in the anterior cingulate cortex and retrosplenial cortex on certain tests of spatial memory in the rat. Behav Brain Res 1994;65:89–101. [15] Saucier D, Hargreaves EL, Boon F, Vanderwolf CH, Cain DP. Detailed behavioral analysis of water maze acquisition under systemic NMDA or muscarinic antagonism: nonspatial pretraining eliminates spatial learning deficits. Behav Neurosci 1996;110:103–16.
[16] Shibata H. Efferent projections from the anterior thalamic nuclei to the cingulate cortex in the rat. J Comp Neurol 1993;330:533– 42. [17] Shibata H. Organization of projections of rat retrosplenial cortex to the anterior thalamic nuclei. Eur J Neurosci 1998;10:3210–9. [18] Sutherland RJ, Whishaw IQ, Kolb B. Contributions of cingulate cortex to two forms of spatial learning and memory. J Neurosci 1988;8:1863–72. [19] Swanson LW. Brain maps: structure of the rat brain. Amsterdam: Elsevier; 1992. [20] Van Groen T, Wyss JM. Connections of the retrosplenial granular a cortex in the rat. J Comp Neurol 1990;300:593–606. [21] Van Groen T, Wyss JM. Connections of the retrosplenial dysgranular cortex in the rat. J Comp Neurol 1992;315:200–16. [22] Van Groen T, Wyss JM. Connections of the retrosplenial granular b cortex in the rat. J Comp Neurol 2003;463:249–63. [23] Vann SD, Aggleton JP. Extensive cytotoxic lesions of the rat retrosplenial cortex reveal consistent deficits on tasks that tax allocentric spatial memory. Behav Neurosci 2002;116:85–94. [24] Vann SD, Brown MW, Aggleton JP. Fos expression in the rostral thalamic nuclei and associated cortical regions in response to different spatial memory tests. Neuroscience 2000;101:983–91. [25] Vann SD, Kristina Wilton LA, Muir JL, Aggleton JP. Testing the importance of the caudal retrosplenial cortex for spatial memory in rats. Behav Brain Res 2003;140:107–18. [26] Vogt BA, Miller MW. Cortical connections between rat cingulate cortex and visual, motor, and postsubicular cortices. J Comp Neurol 1983;216:192–210. [27] Wang B, Gonzalo-Ruiz A, Morte L, Campbell G, Lieberman AR. Immunoelectron microscopic study of glutamate inputs from the retrosplenial granular cortex to identified thalamocortical projection neurons in the anterior thalamus of the rat. Brain Res Bull 1999;50:63–76. [28] Warburton EC, Aggleton JP, Muir JL. Comparing the effects of selective cingulate cortex lesions and cingulum bundle lesions on water maze performance by rats. Eur J Neurosci 1998;10:622– 34. [29] Whishaw IQ. Cholinergic receptor blockade in the rat impairs locale but not taxon strategies for place navigation in a swimming pool. Behav Neurosci 1985;99:979–1005. [30] Whishaw IQ, Maaswinkel H, Gonzalez CL, Kolb B. Deficits in allothetic and idiothetic spatial behavior in rats with posterior cingulate cortex lesions. Behav Brain Res 2001;118:67–76.