- Email: [email protected]
Task solving by procedural strategies in the Morris water maze
Physiology & Behavior 78 (2003) 785 – 793
Task solving by procedural strategies in the Morris water maze Elisabetta Baldi, Carlo Ambrogi Lorenzini, Corrado Bucherelli* Department of Physiological Sciences, University of Florence, Viale G.B. Morgani 63, I-50134, Florence, Italy Received 7 May 2002; received in revised form 17 February 2003; accepted 3 March 2003
Abstract The aim of the present work was to assess the importance of the ‘‘general procedural’’ components, when for rats it was impossible to employ extramaze allothetic information to reach the goal in the Morris water maze (MWM). Groups of Long – Evans rats (males, 70 days old) were trained (10 trials per day, over five consecutive days) following seven paradigms. Four paradigms differed in context (extramaze cues available; extramaze cues not available) and in platform location (constantly at the center of one quadrant of the water maze; at random at the center of any one of the quadrants). In the fifth paradigm, there were no extramaze cues available, and the platform was located at random distances from the maze wall. In the sixth paradigm, rats underwent the standard MWM training (extramaze cues available, invisible platform constantly placed in the center of one quadrant) but they were administered with scopolamine before the daily trials. In a seventh paradigm, the platform was visible. In all paradigms, the starting point was randomized with respect to the goal. When platform distance from the wall was random, there was no significant better performance after the trials. In all the six paradigms in which platform location was at a constant distance from the wall the times spent before reaching the platform decreased progressively, to become constant on Days 4 and 5. The groups which could not employ the allothetic extramaze component (extramaze cues not available; changing of the quadrant of platform location; scopolamine administration) showed a progressively better performance even though their delays on the last 2 days were longer than those of the ‘‘standard MWM’’ and ‘‘visible platform’’ groups. The slightly less efficient performance is attributable to the rat’s search strategy, a ‘‘subcircular’’ swimming pattern within the geometric limits of the central areas of the quadrants, where the platform was constantly placed. That no extramaze allothetic information was employed is shown by the finding that on Day 6 (probe test: 90 s in the tank without platform) no animals exhibited preference for any quadrant, while the ‘‘standard MWM’’ group did show such a preference. It can be concluded that rats under conditions of constant relationship of the goal to the contours of the pool employ search strategies based on general procedural components. D 2003 Elsevier Science Inc. All rights resrved. Keywords: Spatial learning; Procedural strategies; Allothetic strategies; Water maze navigation; Intramaze cues; Extramaze cues
1. Introduction How animals orient themselves in their surroundings is a very interesting topic on which much work has been done, since orientation depends on memory and the learning of many types of information. To study spatial learning and memory in the rat, the Morris water maze (MWM) [1,2] is possibly more useful than other apparatus (T-maze, radialarm maze) [3,4]. It is indeed a very simple test, which can be learnt in very few trials, and learning it does not require strong motivating agents or conditions (punishments, food or water shortage). The MWM is easy to employ also because * Corresponding author. Tel.: +39-055-4237329; fax: +39-0554379506. E-mail address: [email protected] (C. Bucherelli).
although rats are good swimmers, they prefer to stay out of the water [2,5]. The test requires the rats, starting from different points along the border of a circular tank, to find a nonvisible submerged platform, which occupies a fixed location (usually at the center of one of the quadrants in which the surface area of the tank is ideally divided). To reach the goal, i.e. the submerged platform, the animal employs what is called a ‘‘place strategy,’’ i.e. it builds a spatial map correlating context information (extramaze cues) and platform location. Once the map has been built, the rat swims directly towards the platform from any point of the circumference of the tank. This behavior shows the existence of efficient ‘‘allothetic (allocentric) spatial mapping’’ [2,6,7]. On the other hand, it has been shown that rats can orient themselves within the tank when no extramaze information is available, i.e. when allothetic extramaze cues cannot be
0031-9384/03/$ – see front matter D 2003 Elsevier Science Inc. All rights resrved. doi:10.1016/S0031-9384(03)00064-7
786
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
used [8 – 11]. The nonmapping strategies by which the animals may locate the goal (platform) in these far more difficult conditions are thought to be based on a more general learning component, the so-called procedural component [1,2,7,10 –14], defined as ‘‘the representation of stimulus – response habits necessary to guide the animal to correct location’’ [12]. It has also been suggested that even in the standard paradigm (extramaze cues available) the goal is reached by employing both place navigation and stimulus – response strategies [15]. The relative importance of the two strategies has yet to be assessed [9,12]. The hippocampus is involved in a broad range of memory types [16]. In particular, it appears to play a special role in spatial navigation and spatial memory [17 – 19]. These conclusions are based on electrophysiological recordings of hippocampal cells (place cells) which become active when the animal is in a given location in an open field [17]. Results of hippocampal lesions, both in humans and in other mammalian species [17,20], support the hypothesis that the hippocampus is the main neural structure involved in spatial information elaboration. On the other hand, spatial navigation and memory may not depend exclusively on the hippocampus. Rats with hippocampal lesions can still solve landmark-based navigation tasks, for instance, when training is prolonged [21] or when task difficulty progressively increases [22,23]. Moreover, if it appears that hippocampal integrity is necessary for building cognitive maps [17], it has also been shown that navigation is also impaired, although less consistently, by lesions of the neocortex (parietal, frontal, and temporal) [24], the thalamic nuclei [25], the basal ganglia [26], and the cerebellum [27,28]. Even more interesting is the conclusion that there is a dissociation between the neural systems required for spatial navigation based on allothetic spatial learning and those subserving spatial learning based on procedural mechanisms. The hippocampus seems to play a role both in allothetic [17,20] and in procedural [22] spatial learning. The posterior parietal cortex appears to be responsible for allothetic spatial learning, while the prefrontal cortex, possibly together with the caudate nucleus [29] and the cerebellum [10,28], for procedural spatial learning. It is reasonable to think that in order to assess the role of the diverse cerebral sites in spatial navigation it would be helpful to define the behavioral mechanisms subserving it. The question does not appear to be an easy one. It is possible to eliminate all the extramaze information, and in these experimental conditions it is clear that the rats can employ only ‘‘procedural’’ strategies. Unfortunately, the converse condition is not feasible, since it is not possible to retain the allothetic information and exclude the ‘‘procedural’’ one (or ones). Nevertheless, the present work seeks to define as well as possible the ‘‘efficiency’’ of the ‘‘procedural strategies,’’ by means of several paradigms, in all of which the rat starts at random with respect to the hidden platform, which is placed at a constant distance from the wall of the tank. This disposition will allow the subjects to
develop procedural strategies, and not those based on the extramaze allothetic cues, because of: (i) absence of extramaze information, (ii) presence of extramaze information which cannot be utilized, and (iii) pretraining scopolamine administration. By comparing the results obtained in these conditions with those obtained using (i) the standard MWM protocol, (ii) a protocol in which distance of the platform from the wall is not constant, and (iii) a protocol in which the platform is visible, it may be possible to assess the importance of the procedural mechanisms in goal reaching.
2. Materials and methods Seventy-day-old male Long– Evans hooded rats (average body weight 270 g) (Morini, San Polo D’Enza, Italy) were used. The animals were individually housed in stainless-steel cages in a room with a natural light/dark cycle and constant temperature of 22 ± 1 °C. The rats had ad libitum access to food and water throughout the experiment. All animal care and experimental procedures were conducted in accordance with Italian law and European Communities Council recommendations on use of laboratory animals (Directive of November 24, 1986; 86/609/ECC). 2.1. Water maze Spatial learning was assessed in an open-field water maze [1,2] consisting of a circular tank (diameter 1.5 m; depth 0.6 m) containing water at 24 ± 1 °C to a depth of 0.3 m. The rats’ task was to escape from the water by locating an escape platform (diameter 13 cm) submerged 1.5 – 2 cm below or elevated 1.5– 2 cm above the surface of the water. The water was made opaque by adding 3 l of partially skimmed milk in order to prevent the animals from seeing the submerged platform. The pool was located in the center of an acoustically insulated room (3.5 3.6 2.1 (h) m) containing various prominent cues (e.g. wall posters, electrical fittings on the wall) and kept at a constant temperature of 22 ± 1 °C. Illumination was 60 lux and was provided by six incandescent lamps suspended 150 cm above the apparatus, and shielded by a white Plexiglas sheet (5 mm thickness) acting as light diffuser (to avoid sharp lighting and shadows). A low-intensity white noise was emitted from a centrally located loudspeaker to cover sounds that might occur during behavioral training. The swim paths of the animals in the tank were always monitored with a video camera mounted on the ceiling. The video signal was relayed to a video recorder. For all groups, except the NER group (see below), the escape platform was placed in the middle of one quadrant, the center of the platform located 37.5 cm from the side wall. In the NER condition, the escape platform was randomly located at different distances from the side wall in each trial. Each animal underwent a block of 10 trials every day for 5 days. Ten starting point positions were randomly used in a session (i.e. 10 points from 0° to 324° at 36° intervals). From these the rat was gently released into the water, always facing the tank wall. The animal
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
was allowed to swim freely until it found the platform. On reaching the platform, the rat was allowed to remain on it for 20 s before being placed in a cage in a corner of the room covered by a light cloth, so as to exclude extramaze cues. It remained in the cage for 20 s and was then returned to the water for the next trial. If a rat failed to locate the platform within 60 s, it was guided there by the experimenter and allowed to stay there for 20 s. Between one trial and the next, the water was stirred in order to erase olfactory traces of previous swim patterns [30,31]. On completion of the behavioral training (10 trials), the rats were returned to their home cages, where they were briefly warmed under a heat lamp. The procedure was repeated for five consecutive days. On each trial, the time needed to reach the platform was measured. The mean of the escape latencies of every daily block of trials was elaborated statistically. Rats’swimming speed was measured during acquisition training on the fifth trial of every daily block of 10. The time spent to swim over 70 – 100 cm was measured. To define swimming patterns exhibited during acquisition, patterns in which at least 50% of swimming time was spent continuously following a path within a 25-cm annulus (inner diameter: 50 cm, outer diameter: 100 cm) centered on the constant platform – tank wall distance, going through more than two quadrants, were labelled ‘‘subcircular’’ swimming. On the sixth day, the probe test was performed. The platform was not placed in the tank and the rats were allowed to swim freely for 90 s. The total time that subjects spent in each quadrant of the tank was recorded. In these conditions for the ‘‘subcircular’’ swimming pattern, the criterion was that at least 50% of the available time was spent within the 25-cm annulus, crossing more than two quadrants. All time measurements were performed by means of a stopwatch by personnel blinded to which experimental group each animal belonged. 2.2. Experimental groups A total of 52 rats were employed, randomly divided in seven groups of seven to eight animals. The groups were: (1) extramaze fixed (EF) which underwent the standard MWM paradigm, with available extramaze cues and constant platform location, (2) extramaze moved (EM) with available extramaze cues and nonconstant platform location, (3) nonextramaze fixed (NEF) with no extramaze cues available (thick uniform opaque curtains surrounding the tank) and constant platform location, (4) non-extramaze moved (NEM) no extramaze cues (thick uniform opaque curtains surrounding the tank) and nonconstant platform location, (5) nonextramaze random (NER) with no extramaze cues (thick uniform opaque curtains surrounding the tank), platform location and distance from the side wall are at random in each trial, (6) visible platform (VP) with a visible platform, no extramaze cues (thick uniform opaque curtains surrounding the tank), and nonconstant platform location, (7) extramaze fixed-scopolamine (EF-S) with experimental conditions identical to those of group EF, but 20 min before the
787
daily block of trials scopolamine hydrobromide (0.8 mg/kg ip) was administered to the rats during the 5 days of training. 2.3. Statistical analysis Mixed ANOVAs with repeated measures on trials as the seven different procedures as a between-subjects variable and the daily blocks of trials as a within-subjects variable were used to analyze acquisition. The same analysis was employed for the probe tests conducted at the end of water maze training [seven procedures (between subjects) four quadrants (within subjects)]. Newman– Keuls multiple comparisons were used as post hoc test.
3. Results 3.1. Swimming speed In the 260 samples, swimming speed was within 33 and 36 cm/s. There were no significant between-groups differences, nor between days. Scopolamine administration did not influence swimming speed. Mixed ANOVA conducted for swimming speeds during 5 days of water maze acquisition revealed neither significant effects of days [ F(4,180) = 0.86, N.S.] nor significant effects of groups [ F(6,45) = 0.74, N.S.]. 3.2. Escape latency during acquisition Fig. 1 shows that NER escape latency durations remained high during training. In contrast, all other groups of rats exhibited a shortening of escape latency duration during training, eventually showing an asymptotic trend. The shortest latencies are those of the VP group, during the earlier days. At the end of the training, EF latency durations became equal to those of the VP group. All other groups, after an initial shortening of escape latency duration on Days 2 and 3, did not exhibit further shortening of escape latency duration. The above statements are based on statistical analysis results. Mixed ANOVA conducted for escape latencies during 5 days of water maze acquisition revealed a significant main effect of days [ F(4,180) = 6.56, P < .001] and a significant main effect of groups [ F(6,45) = 4.26, P < .01]. There was a significant Groups Days interaction [ F(24,180) = 1.94, P < .01]. Newman– Keuls post hoc comparisons revealed: (i) On Day 1, VP escape latency durations were significantly shorter than those of all other groups ( P < .01 in all instances); there were no differences between the other groups; (ii) On Day 2, escape latency durations of the VP group were again significantly shorter than those of all other groups ( P < .01 in all instances). Escape latency durations of the NER group were higher than those of EF-S, EF, NEF, and NEM ( P < .05 in all instances). Escape latency durations of groups EF-S, EF, EM, NEF, NEM were significantly shorter than those of the same groups on Day 1 ( P < .05 in all
788
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
Fig. 1. Escape latency duratios (mean and S.E.M.) of the seven experimental groups in the MWM during acquisition training (5 daily blocks of 10 trials). VP (visible platform); EF (extramaze cues, platform constantly in the center of the same quadrant); EM (extramaze cues, platform placed in the center of a random quadrant); NEF (no extramaze cues, platform constantly in the center of the same quadrant); NEM (no extramaze cues, platform placed in the center of a random quadrant); EF-S (extramaze cues, platform constantly in the center of the same quadrant, intraperitoneal scopolamine administration before training); NER (no extramaze cues, platform randomly placed in the tank). For explanations, see Materials and methods section.
instances); (iii) On Day 3, escape latency durations of the VP group were significantly shorter than those of groups NER ( P < .01) and EF-S, EM, NEF, NEM ( P < .05 in all instances) but not those of the EF group. Escape latency durations of the NER group were higher than those of EF ( P < .01), EF-S, EM, NEF, and NEM ( P < .05 in all instances). Only the EF group escape latency durations of Day 3 were significantly shorter than those of the same group on Day 2 ( P < .05); (iv) On Day 4, escape latency durations of groups VP and EF were significantly shorter than those of the other groups (NER, EF-S, EM, NEF, and NEM) (VP and EF vs. NER and EM, P < .01 in all four instances; P < .05 in all other instances); escape latency durations of the NER group were higher than those of EF-S, EM, NEF, and NEM ( P < .05 in all instances); no group showed significant differences between escape latency durations of Days 3 and 4; (v) On Day 5, escape latency durations of groups VP and EF were significantly shorter than those of the other groups (VP and EF vs. NER, P < .01 in both instances; P < .05 in all other instances); escape latency durations of the NER group were higher than those of EF-S, EM, NEF, and NEM ( P < .05 in all instances); no group showed significant differences between the escape latency durations of Days 4 and 5.
pattern is shown. On Day 1, the swimming pattern of the EF group was quite similar to those of the groups NER, EM, NEF, NEM, and EF-S: the rats after exhibiting a fairly high thigmotaxis on the first trials happened to reach the platform
3.3. Swimming search strategies Fig. 2 shows the typical swimming patterns of the experimental groups on Days 1, 3, and 5. The patterns were chosen according to the criterion that escape duration of a given rat should be as close as possible to the mean group value on that day. The swimming pattern is used as an indication of the search strategy employed by the rats in the different paradigms. The swimming patterns of groups EM, NEF, NEM, and EF-S were coincident, so that only one
Fig. 2. Typical swim patterns on Days 1, 3, and 5. From Day 1, VP group rats swim a shorter route which becomes almost rectilinear on Day 3. EF group rats follow progressively shorter routes by using an allothetic strategy. EM, NEF, NEM, and EF-S groups rats shorten their escape latency by means of procedural strategy ‘‘subcircular’’ swimming. NER group swimming patterns remain always at random, even on Days 3 – 5. (For explanations, see Fig. 1).
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793 Table 1 Subcircular swimming pattern sampling Days
1
2
3
4
5
6
Trial
1
10
1
10
1
10
1
10
1
10
Probe
VP EF NER EM NEF NEM EF-S
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 28 24 14 12
0 0 0 14 24 14 24
0 28 0 71 87 85 75
0 0 0 43 50 43 37
0 0 0 85 75 71 87
0 0 0 57 50 43 50
0 0 0 85 87 85 75
0 0 0 57 62 71 50
The table shows the percent values of rats of each group exhibiting the ‘‘subcircular’’ swimming pattern (according to criterion) on the first and tenth trial of each daily block and on the probe test. (For explanation, see Fig. 1 and Materials and methods section).
after having swum in the tank following random patterns, while the VP group exhibited a much more direct pattern pointing towards the visible platform. At the 10th trial of Day 1, none of the swimming patterns of all the rats reached the ‘‘subcircular’’ swimming pattern criterion (Table 1). On Day 3: (i) the VP group followed an almost rectilinear swimming pattern towards the platform (Fig. 2) and none of the animals showed ‘‘subcircular’’ swimming (Table 1), (ii) the EF group swimming pattern was considerably shortened, oriented towards the quadrant where the platform was located (Fig. 2); at the 10th trial two rats out of eight exhibited ‘‘subcircular’’ swimming (Table 1), (iii) the NER group continued to show a random pattern (Fig. 2) and none of the rats exhibited ‘‘subcircular’’ swimming (Table 1), (iiii) groups EM, NEF, NEM, and EF-S no longer showed a random swimming pattern, but reached the platform after swimming longer than the EF group. In fact, at the 10th trial of Day 3 between 71% and 87% of the rats of all these groups swam following the subcircular pattern within a 25-cm annulus centered on the known platform location (Table 1).
789
In these groups this behavior begins to appear at the end of the Day 2 block of trials, and is present at the beginning of the Day 3 training (Table 1). The same behavior was observed on Days 4 and 5, becoming very evident at the 10th trial (Table 1). This homogeneous behavior of EM, NEF, NEM, and EF-S groups fits quite well with escape durations of the subsequent days (Fig. 1). The escape latencies became shorter from Day 1 to Day 3, to remain constant up to Day 5. S.E.M. values show that escape duration variance in these four groups is always quite similar, decreasing conspicuously from Day 1 to Day 3 to remain constant on Days 4 and 5. On Day 5, the EF group exhibited a further improvement of performance. Their swimming patterns were almost rectilinear, pointing towards the location of the underwater platform. These swimming patterns were almost identical to those of the VP group (Fig. 2). None of the rats of these two groups ever showed a ‘‘subcircular’’ swimming pattern (Table 1). The NER group did not exhibit performance improvements after Day 3 (Figs. 1 and 2) and none of the rats of this group ever showed a ‘‘subcircular’’ swimming pattern (Table 1). 3.4. Quadrant bias Fig. 3 shows that on Day 6 (probe test) only the EF group showed bias towards one of the quadrants. During the 90 s of free swimming rats of all other groups visited all quadrants equally. ANOVA of the 90-s probe trial revealed a significant main effect of quadrants [ F(2,142) = 3.27, P < .05; numerator degrees of freedom reduced by 1 as the scores for individual animals necessarily add to 90 s] and a significant interaction between Groups Quadrants [ F(12,142) = 1.97, P < .05]. The Newman – Keuls multiple comparisons test showed that the total time spent by the EF rats in the quadrant
Fig. 3. Water maze probe test performed on Day 6. Time spent in each of the four quadrants (mean and S.E.M.). Free exploration for 90 s without platform. For all groups, quadrants are numbered clockwise starting from the quadrant in which, in the cases of constant placement, the platform was located. The symbol * shows a statistically significant difference ( P < .01) of time spent by the EF group in Quadrant 1 against time spent in the other three quadrants and against time spent in all four quadrants by all other groups. For explanations, see Fig. 1 and Materials and methods section.
790
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
where the platform was previously located was significantly higher than that spent in the other three quadrants ( P < .01 in all instances) and higher than the time spent by the rats of the other six groups in each quadrant ( P < .05 in all instances). In EF rats there were no significant differences between the time spent in the three previously unrewarding quadrants. In the other six groups there never were differences between the time spent in the four quadrants. The rats of groups VP, NER, and EF never showed ‘‘subcircular’’ swimming patterns. In contrast, a high number of rats (between 50% and 71%) of groups EF-S (four out of eight), EM (four out of seven), NEF (five out of eight), and NEM (five out of seven) exhibit a well-developed ‘‘subcircular’’ swimming pattern, as defined in the Materials and methods section (see Table 1).
4. Discussion The aim of the present work was to ascertain how efficiently the rats could reach the hidden platform of the MWM when deprived of the extramaze visual cues, i.e. when the animals could not build the standard ‘‘allothetic place representation.’’ In particular, the experimental design aimed at clarifying the efficiency of procedural learning based on tank shape and dimensions, and on platform placement in the tank (distance from tank wall). The present findings show that rats which have to reach the target (the underwater not visible platform) in a MWM when it has been made impossible for them to build an allothetic frame of spatial reference (NEF, NEM, EM, and EF-S groups) still learn to reach their goal, and not by chance. In other words, even under such severe conditions, the rats succeed in reaching the escape platform, only twice slower than with the standard procedure. This may be taken as indication of the good efficiency of their procedural learning. On the technical side, a few remarks may be made. The performance of the animals was quantitatively measured as escape latency, i.e. the time lapsed between the immersion of the animal in the tank and its reaching the platform. This parameter was chosen on the basis of the measured uniform swimming speed of all animals over all trials. In this way, it became unnecessary to measure the length of the swimming pattern, since there is a satisfactory proportionality between times and distances. The measured values of 33 – 36 cm/s confirm previous findings [32]. Scopolamine was administered at dosages known to disrupt spatial memory in the MWM paradigm [3,4,33,34]. Starting locations were randomly chosen on each trial, so as to minimize the possibility that the animals could profit from constant topographical relationships between starting site and hidden platform. Training was ended after Day 5, because in none of the seven experimental groups were there significant differences between escape latencies between Day 4 and Day 5 (Fig. 1). Probe test results on Day 6 showed that only rats of the EF (i.e. standard MWM paradigm) group show an allothetic orientation by their preference for one quadrant of the tank
(Fig. 3). It may be recalled that there are some reports on relatively satisfactory task solving in the MWM by rats variously deprived of extramaze visual cues [8 –11,35,36]. The present results both confirm and extend the previous ones by means of the analysis of a set of diverse experimental designs all proven to be effective in excluding visual extramaze cues from behavioral utilization. When deprived of visual cues, the rats may locate the goal on the basis of auditory [9,11,36] or olfactory [31] cues. Some have even proposed a hierarchical sensorial scale in decreasing order: visual, olfactory, and stimuli generated by the movements of the animal (proprioceptive or kinesthetic) [30]. In the present work, the extramaze visual cues were excluded by the curtains, the olfactory ones by the repeated stirring of the contents of the tank, and the auditory ones by the imposed white noise. Thus, to solve the task the rats had access only to the information originating from their motility and those deriving from the shape and dimensions of the tank and particularly on the constant distance of the platform from the wall of the tank (groups EM, NEF, NEM, and EF-S), i.e. they could employ only the general procedural mechanisms. As recalled in the Introduction, according to several authors, task solving in the standard MWM may be due not only to allothetic mechanisms but also to ‘‘procedural memory’’ [1,2,7,10 – 14]. In fact, when the animals are allowed to employ both mapping and nonmapping (allothetic and general procedural) mechanisms, they learn quite efficiently to reach their goal, the hidden platform (see the EF group performance, Figs. 1 and 2). Indeed, their performance is almost as good as that of the animals that have to reach a visible platform (VP group, Figs. 1 and 2). The results show that the EF group (standard MWM task) exhibited the best performance. EM, NEM, and NEF groups exhibited a lesser performance in terms of escape delay duration but still adequately reached the goal (Figs. 1 and 2). In other words, in extramaze allothetic deprivation conditions, task solving on purely general procedural mechanisms was more time consuming, but not conducive to an inferior performance as far as strictly finding the goal is concerned. As expected, when also the constant platform – tank wall distance information was denied (group NER), the rats could not progressively improve their performance. The performance of EM, NEM, and NEF groups is clearly due to a learning process, since over the days of training there is a progressive decrease of escape latency duration down to a constant value. In more detail, it can be underlined that the NEM and NEF groups reach the goal after swimming progressively less (escape latency duration), in both instances, in the absence of extramaze visual cues and without significant differences between groups. The absence of significant differences in escape latency durations between these groups shows that the deprivation of extramaze cues was satisfactorily efficient. In fact, when hidden platform location was kept constant, a better performance was not measured. On the other hand, for the EM group, extramaze visual cues were available but of no use in reaching the
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
hidden platform, since platform location was not constant, being varied on each trial. Thus, in these experimental conditions, the extramaze allothetic information was not sufficient for task solving [8,10,13,37]. The animals must rely only on general procedural mechanisms to reach the goal, which, as stated above, they do satisfactorily. The swimming patterns of these groups (EM, NEF, NEM) show that in these conditions the animals swim keeping themselves at a fairly constant distance from the wall of the tank, following an approximately circular path (Fig. 2, Table 1). This ‘‘subcircular’’ swimming pattern is almost universally present already on Day 3 at the end of the daily block of training (Table 1). In all these groups, swimming patterns and escape durations appear to be almost identical. The swimming behavior shows that the rats have learnt that the platform is located at a constant distance from the wall of the tank (the ‘‘where’’) and that the most efficient way to find it is to swim following a ‘‘subcircular pattern’’ (the ‘‘how’’). The finding that the ‘‘subcircular’’ swimming pattern is exhibited on the first trials of Days 4 and 5, and on the probe test (Day 6) shows that this behavior is not only maintained by working memory mechanisms developing during the daily block of 10 trials, but indeed, is a learned and consolidated behavior that can be recalled at 24-h intervals. The finding that NER group rats do not develop the ‘‘subcircular’’ swimming pattern shows that the constancy of the platform –tank wall relationship is necessary for developing the search strategy. These findings support the hypotheses advanced by some researchers [38] according to whom the animals find the goal using both procedural ‘‘spatial’’ and ‘‘nonspatial’’ cues and confirm and extend data reported for other species of rodents [39]. At this point, we must recall that allothetic is not an antonym to procedural. In fact, the procedural strategy is based also on an intramaze allothesis. It is true that a general procedure solution was used only when the training conditions did not offer to the rats any extramaze allothetic cues, making it possible to enter the position of the goal into their cognitive maps. On the other hand, it is not entirely true that procedural training was totally devoid of allothetic memories. In fact, swimming in a circular annulus from the center of the pool can only be accomplished by continuously monitoring the distance from the wall of the pool, i.e. by an intramaze allothetically guided behavior. Thus, ‘‘general procedure goal finding strategies’’ must not be considered completely independent from allothetic mechanisms. The swimming behavior of the EF-S group was quite similar to that of the other groups to which extramaze allothetic information was denied (EM, NEF, NEM). Thus, animals of the EF-S group reach the platform using a strategy based on general procedural mechanisms (Fig. 2, Table 1). Scopolamine disrupts spatial memory [3,4,33,34], and it is accepted that the central cholinergic systems are those that support spatial allothetic mapping (platform location), but not other superior nervous functions, such as those necessary
791
for nonextramaze visual searching strategies (how to find the platform) [3,40]. The behavior of the EF-S group confirms this hypothesis. The present findings show that the general procedure goal finding strategies can be evolved and employed in complete independence from extramaze allothetic information. This is shown both by the recorded swimming strategies, and by the finding that during the probe test none of these four experimental groups (NEM, NEF, EM, EF-S) exhibited any quadrant preference, i.e. they did not exhibit any behavior based on classically defined allothetic memories. Thus, under conditions of constant relationship of the goal to the contours of the pool, even without extramaze allothetic but with intramaze information, rats are able to elaborate a search strategy probably based both on praxis and intramaze allothetic solutions of the task. The presently employed paradigm does not allow us to assess the relative importance of either component in solving the task (NEM, NEF, EM, and EF-S groups). If the position of the goal was rotated by the same angle for every new position of the start (e.g. start/goal positions 0/90° on Swim 1 to 72/162° on Swim 2), probably in this case, the animals (employing a praxis strategy) would be able to find the platform with a shorter latency than in the present cases. On the other hand, even a direct goal approach from any start may be supported by intramaze allothesis alone when the goal place is uniquely determined by the contours of the pool, as when the goal is in the center of a circular pool. In general, we must recall that since it does not appear to be possible to devise an experimental paradigm by means of which the general procedure mechanisms can be excluded so that rats must solve the task of finding the goal using only the allothetic ones, it is not possible to know in the standard EF paradigm: (i) if only extramaze allothetic information or both are concurrently employed; (ii) if both, whether the two mechanisms concur equally to goal-finding or not; and (iii) if not, if there is a hierarchical relationship, or bias, between the employment of the two mechanisms. Nonetheless, some hypotheses have been formulated on these points. According to several authors [7,11,14,38,41] in the standard MWM paradigm, procedural learning is followed by allothetic learning. Within limits, the present results support this hypothesis. In fact, significant statistical differences between delay durations of the EF group and all the allotheticdeprived ones (NEF, NEM, EM, EF-S) are found only from Day 4 on, i.e. after general procedural mechanisms have been completely exploited. Moreover, it can be observed that at the 10th trial of Day 3 the ‘‘subcircular’’ swimming pattern was temporarily exhibited by some (25%) of the rats of the EF group (Table 1). On the following days, this pattern no longer occurred, all animals exhibiting the more efficient ‘‘allothetic goal-directed swim.’’ In fact, from Day 4 onwards, the allothetic-deprived groups did not exhibit a further shortening of escape delay, which was instead exhibited by the EF group, down to the same durations of the group which had to reach a visible platform (VP).
792
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
In conclusion, the present findings show that in the MWM rats are able to evolve an efficient search strategy for goal reaching, even when they are denied the possibility of building an allothetic strategy based on extramaze visual cues. On the other hand, when sufficient intramaze information is available (constancy of platform – tank wall relationship), the rats evolve an adequate search strategy, in which a ‘‘subcircular’’ swimming pattern is used. This pattern, which is based on general procedural mechanisms, is evolved by most animals in these experimental conditions, and is remembered for a long time. As recalled in the Introduction, besides the primary role of the hippocampus in spatial learning, several other neural sites are thought to play a role in this type of learning. One question that is still to be addressed concerns which structures, sites, or systems of the central nervous system are more or less specifically involved in the distinct elaboration of allothetic and procedural learning [3,4]. It could be of some interest, on the basis of the present findings and by employing similar protocols, to try to assess the involvement of several subcortical and cortical sites at least in general procedures learning and utilization.
Acknowledgements The authors thank Mr. A. Aiazzi, Mr. S. Cammarata, Mr. M. Dolfi, and Mr. A. Vannucchi for their technical assistance.
References [1] Morris RGM. Spatial localization does not require the presence of local cues. Learn Motiv 1981;12:239 – 60. [2] Morris R. Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 1984;11:47 – 60. [3] Brandeis R, Brandys Y, Yehuda S. The use of the Morris water maze in the study of memory and learning. Int J Neurosci 1989;48:29 – 69. [4] D’Hooge R, De Deyn PP. Applications of the Morris water maze in the study of learning and memory. Brain Res Rev 2001;36:60 – 90. [5] Whishaw IQ. Place learning in hippocampal rats and the path integration hypothesis. Neurosci Biobehav Rev 1998;22:209 – 20. [6] DiMattia BD, Kesner RP. Spatial cognitive maps: differential role of parietal cortex and hippocampal formation. Behav Neurosci 1988;102: 471 – 80. [7] Fenton AA, Bures J. Place navigation in rats with unilateral tetrodotoxin inactivation of the dorsal hippocampus: place but not procedural learning can be lateralized to one hippocampus. Behav Neurosci 1993; 107:552 – 64. [8] Bannerman DM, Good MA, Butcher SP, Ramsay M, Morris RGM. Distinct components of spatial learning revealed by prior training and NMDA receptor blockade. Nature 1995;378:182 – 6. [9] Buresova O, Homuta L, Krekule I, Bures J. Does nondirectional signalization of target distance contribute to navigation in the Morris water maze? Behav Neural Biol 1988;49:240 – 8. [10] Leggio MG, Neri P, Graziano A, Mandolesi L, Molinari M, Petrosini L. Cerebellar contribution to spatial event processing: characterization of procedural learning. Exp Brain Res 1999;127:1 – 11. [11] Sutherland RJ, Dyck RH. Place navigation by rats in a swimming pool. Can J Psychol 1984;38:322 – 47.
[12] Schenk F, Morris RGM. Dissociation between components of spatial memory in rats after recovery from the effects of retrohippocampal lesions. Exp Brain Res 1985;58:11 – 28. [13] Whishaw IQ. Dissociating performance and learning deficits on spatial navigation tasks in rats subjected to cholinergic muscarinic blockade. Brain Res Bull 1989;23:347 – 58. [14] Whishaw IQ, Mittleman G. Visits to starts, routes and places by rats (Rattus norvegicus) in swimming pool navigation tasks. J Comp Psychol 1986;100:422 – 32. [15] Morris RGM, Frey U. Hippocampal synaptic plasticity: role in spatial learning or the automatic recording of attended experience? Philos Trans R Soc Lond B 1997;352:1489 – 503. [16] Squire LR. Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychol Rev 1992;99:195 – 231. [17] O’Keefe J, Nadel L. The hippocampus as a cognitive map. Oxford: Clarendon; 1978. [18] Nadel L. The hippocampus and space revisited. Hippocampus 1991;1: 221 – 9. [19] Steckler T, Drinkenburg WH, Sahgal A, Aggleton JP. Recognition memory in rats: II. Neuroanatomical substrates. Prog Neurobiol 1998;54:313 – 32. [20] Morris RGM, Garrud P, Rawlins JNP, O’Keefe J. Place navigation impaired in rats with hippocampal lesions. Nature 1982;297: 681 – 3. [21] Morris RGM, Schenk F, Tweedie F, Jarrard LE. Ibotenate lesions of hippocampus and/or subiculum: dissociating components of allocentric spatial learning. Eur J Neurosci 1990;2:1016 – 28. [22] Whishaw IQ, Cassel JC, Jarrard LE. Rats with fimbria – fornix lesions display a place response in a swimming pool: a dissociation between getting there and knowing where. J Neurosci 1995;15:5779 – 88. [23] Whishaw IQ, Jarrad LE. Evidence for extrahippocampal involvement in place learning and hippocampal involvement in path integration. Hippocampus 1996;6:513 – 24. [24] Kolb B, Sutherland RJ, Whishaw IQ. A comparison of the contributions of the frontal and parietal association cortex to spatial localizations in rats. Behav Neurosci 1983;97:13 – 27. [25] Mizumori SJ, Miya DY, Ward KE. Reversible inactivation of the lateral dorsal thalamus disrupts hippocampal place representation and impairs spatial learning. Brain Res 1994;644:168 – 74. [26] Whishaw IQ, Mittleman G, Bunch ST, Dunnett SB. Impairments in the acquisition, retention and selection of spatial navigation strategies after medial caudate – putamen lesions in rats. Behav Brain Res 1987; 24:125 – 38. [27] Lalonde R. Cerebellar contributions to instrumental learning. Neurosci Biobehav Rev 1994;18:161 – 70. [28] Petrosini L, Leggio MG, Molinari M. The cerebellum in the spatial problem solving: a co-star or a guest star? Prog Neurobiol 1998;56: 191 – 210. [29] Packard MG, McGaugh JL. Inactivation of hippocampus or caudate nucleus with lidocaine differentially affects expression of place and response learning. Neurobiol Learn Mem 1996;65:65 – 72. [30] Maaswinkel H, Whishaw IQ. Homing with locale, taxon, and dead reckoning strategies by foraging rats: sensory hierarchy in spatial navigation. Behav Brain Res 1999;99:143 – 52. [31] Means LW, Alexander SR, O’Neal MF. Those cheating rats: male and female rats use odor trails in a water-escape ‘‘working memory’’ task. Behav Neural Biol 1992;58:144 – 51. [32] Bannerman DM, Yee BK, Lemaire M, Wilbrecht L, Jarrard L, Iversen SD, et al. The role of the entorhinal cortex in two forms of spatial learning and memory. Exp Brain Res 2001;141:281 – 303. [33] Buresova O, Bolhuis JJ, Bures J. Differential effects of cholinergic blockade on performance of rats in the water tank navigation task and in a radial water maze. Behav Neurosci 1986;100:476 – 82. [34] McNamara RK, Skelton RW. The neuropharmacological and neurochemical basis of place learning in the Morris water maze. Brain Res Rev 1993;18:33 – 49. [35] Liu Z, Francis Turner L, Bures J. Impairment of place navigation of
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793 rats in the Morris water maze by intermittent light is inversely related to the duration of the flash. Neurosci Lett 1994;180:59 – 62. [36] Moghaddam M, Bures J. Contribution of egocentric spatial memory to place navigation of rats in the Morris water maze. Behav Brain Res 1996;78:121 – 9. [37] De Bruin JPC, Swinkels WAM, De Brabander JM. Response learning of rats in a Morris water maze: involvement of the medial prefrontal cortex. Behav Brain Res 1997;85:47 – 55. [38] Carman HM, Mactutus CF. Ontogeny of spatial navigation in rats: a role for response requirements? Behav Neurosci 2001;115:870 – 9.
793
[39] Dalm S, Grootendorst J, De Koet ER, Oitzl MS. Quantification of swim patterns in the Morris water maze. Behav Res Methods Instrum Comput 2000;32:134 – 9. [40] Janis SL, Glasier MM, Fulop Z, Stein DG. Intraseptal injections of 192 IgG saporin produce deficits for strategy selection in spatialmemory tasks. Behav Brain Res 1998;90:23 – 34. [41] Whishaw IQ, Petrie BF. Cholinergic blockade in the rat impairs strategy selection but not learning and retention of nonspatial visual discrimination problems in a swimming pool. Behav Neurosci 1988;102: 662 – 77.
Task solving by procedural strategies in the Morris water maze Elisabetta Baldi, Carlo Ambrogi Lorenzini, Corrado Bucherelli* Department of Physiological Sciences, University of Florence, Viale G.B. Morgani 63, I-50134, Florence, Italy Received 7 May 2002; received in revised form 17 February 2003; accepted 3 March 2003
Abstract The aim of the present work was to assess the importance of the ‘‘general procedural’’ components, when for rats it was impossible to employ extramaze allothetic information to reach the goal in the Morris water maze (MWM). Groups of Long – Evans rats (males, 70 days old) were trained (10 trials per day, over five consecutive days) following seven paradigms. Four paradigms differed in context (extramaze cues available; extramaze cues not available) and in platform location (constantly at the center of one quadrant of the water maze; at random at the center of any one of the quadrants). In the fifth paradigm, there were no extramaze cues available, and the platform was located at random distances from the maze wall. In the sixth paradigm, rats underwent the standard MWM training (extramaze cues available, invisible platform constantly placed in the center of one quadrant) but they were administered with scopolamine before the daily trials. In a seventh paradigm, the platform was visible. In all paradigms, the starting point was randomized with respect to the goal. When platform distance from the wall was random, there was no significant better performance after the trials. In all the six paradigms in which platform location was at a constant distance from the wall the times spent before reaching the platform decreased progressively, to become constant on Days 4 and 5. The groups which could not employ the allothetic extramaze component (extramaze cues not available; changing of the quadrant of platform location; scopolamine administration) showed a progressively better performance even though their delays on the last 2 days were longer than those of the ‘‘standard MWM’’ and ‘‘visible platform’’ groups. The slightly less efficient performance is attributable to the rat’s search strategy, a ‘‘subcircular’’ swimming pattern within the geometric limits of the central areas of the quadrants, where the platform was constantly placed. That no extramaze allothetic information was employed is shown by the finding that on Day 6 (probe test: 90 s in the tank without platform) no animals exhibited preference for any quadrant, while the ‘‘standard MWM’’ group did show such a preference. It can be concluded that rats under conditions of constant relationship of the goal to the contours of the pool employ search strategies based on general procedural components. D 2003 Elsevier Science Inc. All rights resrved. Keywords: Spatial learning; Procedural strategies; Allothetic strategies; Water maze navigation; Intramaze cues; Extramaze cues
1. Introduction How animals orient themselves in their surroundings is a very interesting topic on which much work has been done, since orientation depends on memory and the learning of many types of information. To study spatial learning and memory in the rat, the Morris water maze (MWM) [1,2] is possibly more useful than other apparatus (T-maze, radialarm maze) [3,4]. It is indeed a very simple test, which can be learnt in very few trials, and learning it does not require strong motivating agents or conditions (punishments, food or water shortage). The MWM is easy to employ also because * Corresponding author. Tel.: +39-055-4237329; fax: +39-0554379506. E-mail address: [email protected] (C. Bucherelli).
although rats are good swimmers, they prefer to stay out of the water [2,5]. The test requires the rats, starting from different points along the border of a circular tank, to find a nonvisible submerged platform, which occupies a fixed location (usually at the center of one of the quadrants in which the surface area of the tank is ideally divided). To reach the goal, i.e. the submerged platform, the animal employs what is called a ‘‘place strategy,’’ i.e. it builds a spatial map correlating context information (extramaze cues) and platform location. Once the map has been built, the rat swims directly towards the platform from any point of the circumference of the tank. This behavior shows the existence of efficient ‘‘allothetic (allocentric) spatial mapping’’ [2,6,7]. On the other hand, it has been shown that rats can orient themselves within the tank when no extramaze information is available, i.e. when allothetic extramaze cues cannot be
0031-9384/03/$ – see front matter D 2003 Elsevier Science Inc. All rights resrved. doi:10.1016/S0031-9384(03)00064-7
786
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
used [8 – 11]. The nonmapping strategies by which the animals may locate the goal (platform) in these far more difficult conditions are thought to be based on a more general learning component, the so-called procedural component [1,2,7,10 –14], defined as ‘‘the representation of stimulus – response habits necessary to guide the animal to correct location’’ [12]. It has also been suggested that even in the standard paradigm (extramaze cues available) the goal is reached by employing both place navigation and stimulus – response strategies [15]. The relative importance of the two strategies has yet to be assessed [9,12]. The hippocampus is involved in a broad range of memory types [16]. In particular, it appears to play a special role in spatial navigation and spatial memory [17 – 19]. These conclusions are based on electrophysiological recordings of hippocampal cells (place cells) which become active when the animal is in a given location in an open field [17]. Results of hippocampal lesions, both in humans and in other mammalian species [17,20], support the hypothesis that the hippocampus is the main neural structure involved in spatial information elaboration. On the other hand, spatial navigation and memory may not depend exclusively on the hippocampus. Rats with hippocampal lesions can still solve landmark-based navigation tasks, for instance, when training is prolonged [21] or when task difficulty progressively increases [22,23]. Moreover, if it appears that hippocampal integrity is necessary for building cognitive maps [17], it has also been shown that navigation is also impaired, although less consistently, by lesions of the neocortex (parietal, frontal, and temporal) [24], the thalamic nuclei [25], the basal ganglia [26], and the cerebellum [27,28]. Even more interesting is the conclusion that there is a dissociation between the neural systems required for spatial navigation based on allothetic spatial learning and those subserving spatial learning based on procedural mechanisms. The hippocampus seems to play a role both in allothetic [17,20] and in procedural [22] spatial learning. The posterior parietal cortex appears to be responsible for allothetic spatial learning, while the prefrontal cortex, possibly together with the caudate nucleus [29] and the cerebellum [10,28], for procedural spatial learning. It is reasonable to think that in order to assess the role of the diverse cerebral sites in spatial navigation it would be helpful to define the behavioral mechanisms subserving it. The question does not appear to be an easy one. It is possible to eliminate all the extramaze information, and in these experimental conditions it is clear that the rats can employ only ‘‘procedural’’ strategies. Unfortunately, the converse condition is not feasible, since it is not possible to retain the allothetic information and exclude the ‘‘procedural’’ one (or ones). Nevertheless, the present work seeks to define as well as possible the ‘‘efficiency’’ of the ‘‘procedural strategies,’’ by means of several paradigms, in all of which the rat starts at random with respect to the hidden platform, which is placed at a constant distance from the wall of the tank. This disposition will allow the subjects to
develop procedural strategies, and not those based on the extramaze allothetic cues, because of: (i) absence of extramaze information, (ii) presence of extramaze information which cannot be utilized, and (iii) pretraining scopolamine administration. By comparing the results obtained in these conditions with those obtained using (i) the standard MWM protocol, (ii) a protocol in which distance of the platform from the wall is not constant, and (iii) a protocol in which the platform is visible, it may be possible to assess the importance of the procedural mechanisms in goal reaching.
2. Materials and methods Seventy-day-old male Long– Evans hooded rats (average body weight 270 g) (Morini, San Polo D’Enza, Italy) were used. The animals were individually housed in stainless-steel cages in a room with a natural light/dark cycle and constant temperature of 22 ± 1 °C. The rats had ad libitum access to food and water throughout the experiment. All animal care and experimental procedures were conducted in accordance with Italian law and European Communities Council recommendations on use of laboratory animals (Directive of November 24, 1986; 86/609/ECC). 2.1. Water maze Spatial learning was assessed in an open-field water maze [1,2] consisting of a circular tank (diameter 1.5 m; depth 0.6 m) containing water at 24 ± 1 °C to a depth of 0.3 m. The rats’ task was to escape from the water by locating an escape platform (diameter 13 cm) submerged 1.5 – 2 cm below or elevated 1.5– 2 cm above the surface of the water. The water was made opaque by adding 3 l of partially skimmed milk in order to prevent the animals from seeing the submerged platform. The pool was located in the center of an acoustically insulated room (3.5 3.6 2.1 (h) m) containing various prominent cues (e.g. wall posters, electrical fittings on the wall) and kept at a constant temperature of 22 ± 1 °C. Illumination was 60 lux and was provided by six incandescent lamps suspended 150 cm above the apparatus, and shielded by a white Plexiglas sheet (5 mm thickness) acting as light diffuser (to avoid sharp lighting and shadows). A low-intensity white noise was emitted from a centrally located loudspeaker to cover sounds that might occur during behavioral training. The swim paths of the animals in the tank were always monitored with a video camera mounted on the ceiling. The video signal was relayed to a video recorder. For all groups, except the NER group (see below), the escape platform was placed in the middle of one quadrant, the center of the platform located 37.5 cm from the side wall. In the NER condition, the escape platform was randomly located at different distances from the side wall in each trial. Each animal underwent a block of 10 trials every day for 5 days. Ten starting point positions were randomly used in a session (i.e. 10 points from 0° to 324° at 36° intervals). From these the rat was gently released into the water, always facing the tank wall. The animal
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
was allowed to swim freely until it found the platform. On reaching the platform, the rat was allowed to remain on it for 20 s before being placed in a cage in a corner of the room covered by a light cloth, so as to exclude extramaze cues. It remained in the cage for 20 s and was then returned to the water for the next trial. If a rat failed to locate the platform within 60 s, it was guided there by the experimenter and allowed to stay there for 20 s. Between one trial and the next, the water was stirred in order to erase olfactory traces of previous swim patterns [30,31]. On completion of the behavioral training (10 trials), the rats were returned to their home cages, where they were briefly warmed under a heat lamp. The procedure was repeated for five consecutive days. On each trial, the time needed to reach the platform was measured. The mean of the escape latencies of every daily block of trials was elaborated statistically. Rats’swimming speed was measured during acquisition training on the fifth trial of every daily block of 10. The time spent to swim over 70 – 100 cm was measured. To define swimming patterns exhibited during acquisition, patterns in which at least 50% of swimming time was spent continuously following a path within a 25-cm annulus (inner diameter: 50 cm, outer diameter: 100 cm) centered on the constant platform – tank wall distance, going through more than two quadrants, were labelled ‘‘subcircular’’ swimming. On the sixth day, the probe test was performed. The platform was not placed in the tank and the rats were allowed to swim freely for 90 s. The total time that subjects spent in each quadrant of the tank was recorded. In these conditions for the ‘‘subcircular’’ swimming pattern, the criterion was that at least 50% of the available time was spent within the 25-cm annulus, crossing more than two quadrants. All time measurements were performed by means of a stopwatch by personnel blinded to which experimental group each animal belonged. 2.2. Experimental groups A total of 52 rats were employed, randomly divided in seven groups of seven to eight animals. The groups were: (1) extramaze fixed (EF) which underwent the standard MWM paradigm, with available extramaze cues and constant platform location, (2) extramaze moved (EM) with available extramaze cues and nonconstant platform location, (3) nonextramaze fixed (NEF) with no extramaze cues available (thick uniform opaque curtains surrounding the tank) and constant platform location, (4) non-extramaze moved (NEM) no extramaze cues (thick uniform opaque curtains surrounding the tank) and nonconstant platform location, (5) nonextramaze random (NER) with no extramaze cues (thick uniform opaque curtains surrounding the tank), platform location and distance from the side wall are at random in each trial, (6) visible platform (VP) with a visible platform, no extramaze cues (thick uniform opaque curtains surrounding the tank), and nonconstant platform location, (7) extramaze fixed-scopolamine (EF-S) with experimental conditions identical to those of group EF, but 20 min before the
787
daily block of trials scopolamine hydrobromide (0.8 mg/kg ip) was administered to the rats during the 5 days of training. 2.3. Statistical analysis Mixed ANOVAs with repeated measures on trials as the seven different procedures as a between-subjects variable and the daily blocks of trials as a within-subjects variable were used to analyze acquisition. The same analysis was employed for the probe tests conducted at the end of water maze training [seven procedures (between subjects) four quadrants (within subjects)]. Newman– Keuls multiple comparisons were used as post hoc test.
3. Results 3.1. Swimming speed In the 260 samples, swimming speed was within 33 and 36 cm/s. There were no significant between-groups differences, nor between days. Scopolamine administration did not influence swimming speed. Mixed ANOVA conducted for swimming speeds during 5 days of water maze acquisition revealed neither significant effects of days [ F(4,180) = 0.86, N.S.] nor significant effects of groups [ F(6,45) = 0.74, N.S.]. 3.2. Escape latency during acquisition Fig. 1 shows that NER escape latency durations remained high during training. In contrast, all other groups of rats exhibited a shortening of escape latency duration during training, eventually showing an asymptotic trend. The shortest latencies are those of the VP group, during the earlier days. At the end of the training, EF latency durations became equal to those of the VP group. All other groups, after an initial shortening of escape latency duration on Days 2 and 3, did not exhibit further shortening of escape latency duration. The above statements are based on statistical analysis results. Mixed ANOVA conducted for escape latencies during 5 days of water maze acquisition revealed a significant main effect of days [ F(4,180) = 6.56, P < .001] and a significant main effect of groups [ F(6,45) = 4.26, P < .01]. There was a significant Groups Days interaction [ F(24,180) = 1.94, P < .01]. Newman– Keuls post hoc comparisons revealed: (i) On Day 1, VP escape latency durations were significantly shorter than those of all other groups ( P < .01 in all instances); there were no differences between the other groups; (ii) On Day 2, escape latency durations of the VP group were again significantly shorter than those of all other groups ( P < .01 in all instances). Escape latency durations of the NER group were higher than those of EF-S, EF, NEF, and NEM ( P < .05 in all instances). Escape latency durations of groups EF-S, EF, EM, NEF, NEM were significantly shorter than those of the same groups on Day 1 ( P < .05 in all
788
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
Fig. 1. Escape latency duratios (mean and S.E.M.) of the seven experimental groups in the MWM during acquisition training (5 daily blocks of 10 trials). VP (visible platform); EF (extramaze cues, platform constantly in the center of the same quadrant); EM (extramaze cues, platform placed in the center of a random quadrant); NEF (no extramaze cues, platform constantly in the center of the same quadrant); NEM (no extramaze cues, platform placed in the center of a random quadrant); EF-S (extramaze cues, platform constantly in the center of the same quadrant, intraperitoneal scopolamine administration before training); NER (no extramaze cues, platform randomly placed in the tank). For explanations, see Materials and methods section.
instances); (iii) On Day 3, escape latency durations of the VP group were significantly shorter than those of groups NER ( P < .01) and EF-S, EM, NEF, NEM ( P < .05 in all instances) but not those of the EF group. Escape latency durations of the NER group were higher than those of EF ( P < .01), EF-S, EM, NEF, and NEM ( P < .05 in all instances). Only the EF group escape latency durations of Day 3 were significantly shorter than those of the same group on Day 2 ( P < .05); (iv) On Day 4, escape latency durations of groups VP and EF were significantly shorter than those of the other groups (NER, EF-S, EM, NEF, and NEM) (VP and EF vs. NER and EM, P < .01 in all four instances; P < .05 in all other instances); escape latency durations of the NER group were higher than those of EF-S, EM, NEF, and NEM ( P < .05 in all instances); no group showed significant differences between escape latency durations of Days 3 and 4; (v) On Day 5, escape latency durations of groups VP and EF were significantly shorter than those of the other groups (VP and EF vs. NER, P < .01 in both instances; P < .05 in all other instances); escape latency durations of the NER group were higher than those of EF-S, EM, NEF, and NEM ( P < .05 in all instances); no group showed significant differences between the escape latency durations of Days 4 and 5.
pattern is shown. On Day 1, the swimming pattern of the EF group was quite similar to those of the groups NER, EM, NEF, NEM, and EF-S: the rats after exhibiting a fairly high thigmotaxis on the first trials happened to reach the platform
3.3. Swimming search strategies Fig. 2 shows the typical swimming patterns of the experimental groups on Days 1, 3, and 5. The patterns were chosen according to the criterion that escape duration of a given rat should be as close as possible to the mean group value on that day. The swimming pattern is used as an indication of the search strategy employed by the rats in the different paradigms. The swimming patterns of groups EM, NEF, NEM, and EF-S were coincident, so that only one
Fig. 2. Typical swim patterns on Days 1, 3, and 5. From Day 1, VP group rats swim a shorter route which becomes almost rectilinear on Day 3. EF group rats follow progressively shorter routes by using an allothetic strategy. EM, NEF, NEM, and EF-S groups rats shorten their escape latency by means of procedural strategy ‘‘subcircular’’ swimming. NER group swimming patterns remain always at random, even on Days 3 – 5. (For explanations, see Fig. 1).
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793 Table 1 Subcircular swimming pattern sampling Days
1
2
3
4
5
6
Trial
1
10
1
10
1
10
1
10
1
10
Probe
VP EF NER EM NEF NEM EF-S
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 28 24 14 12
0 0 0 14 24 14 24
0 28 0 71 87 85 75
0 0 0 43 50 43 37
0 0 0 85 75 71 87
0 0 0 57 50 43 50
0 0 0 85 87 85 75
0 0 0 57 62 71 50
The table shows the percent values of rats of each group exhibiting the ‘‘subcircular’’ swimming pattern (according to criterion) on the first and tenth trial of each daily block and on the probe test. (For explanation, see Fig. 1 and Materials and methods section).
after having swum in the tank following random patterns, while the VP group exhibited a much more direct pattern pointing towards the visible platform. At the 10th trial of Day 1, none of the swimming patterns of all the rats reached the ‘‘subcircular’’ swimming pattern criterion (Table 1). On Day 3: (i) the VP group followed an almost rectilinear swimming pattern towards the platform (Fig. 2) and none of the animals showed ‘‘subcircular’’ swimming (Table 1), (ii) the EF group swimming pattern was considerably shortened, oriented towards the quadrant where the platform was located (Fig. 2); at the 10th trial two rats out of eight exhibited ‘‘subcircular’’ swimming (Table 1), (iii) the NER group continued to show a random pattern (Fig. 2) and none of the rats exhibited ‘‘subcircular’’ swimming (Table 1), (iiii) groups EM, NEF, NEM, and EF-S no longer showed a random swimming pattern, but reached the platform after swimming longer than the EF group. In fact, at the 10th trial of Day 3 between 71% and 87% of the rats of all these groups swam following the subcircular pattern within a 25-cm annulus centered on the known platform location (Table 1).
789
In these groups this behavior begins to appear at the end of the Day 2 block of trials, and is present at the beginning of the Day 3 training (Table 1). The same behavior was observed on Days 4 and 5, becoming very evident at the 10th trial (Table 1). This homogeneous behavior of EM, NEF, NEM, and EF-S groups fits quite well with escape durations of the subsequent days (Fig. 1). The escape latencies became shorter from Day 1 to Day 3, to remain constant up to Day 5. S.E.M. values show that escape duration variance in these four groups is always quite similar, decreasing conspicuously from Day 1 to Day 3 to remain constant on Days 4 and 5. On Day 5, the EF group exhibited a further improvement of performance. Their swimming patterns were almost rectilinear, pointing towards the location of the underwater platform. These swimming patterns were almost identical to those of the VP group (Fig. 2). None of the rats of these two groups ever showed a ‘‘subcircular’’ swimming pattern (Table 1). The NER group did not exhibit performance improvements after Day 3 (Figs. 1 and 2) and none of the rats of this group ever showed a ‘‘subcircular’’ swimming pattern (Table 1). 3.4. Quadrant bias Fig. 3 shows that on Day 6 (probe test) only the EF group showed bias towards one of the quadrants. During the 90 s of free swimming rats of all other groups visited all quadrants equally. ANOVA of the 90-s probe trial revealed a significant main effect of quadrants [ F(2,142) = 3.27, P < .05; numerator degrees of freedom reduced by 1 as the scores for individual animals necessarily add to 90 s] and a significant interaction between Groups Quadrants [ F(12,142) = 1.97, P < .05]. The Newman – Keuls multiple comparisons test showed that the total time spent by the EF rats in the quadrant
Fig. 3. Water maze probe test performed on Day 6. Time spent in each of the four quadrants (mean and S.E.M.). Free exploration for 90 s without platform. For all groups, quadrants are numbered clockwise starting from the quadrant in which, in the cases of constant placement, the platform was located. The symbol * shows a statistically significant difference ( P < .01) of time spent by the EF group in Quadrant 1 against time spent in the other three quadrants and against time spent in all four quadrants by all other groups. For explanations, see Fig. 1 and Materials and methods section.
790
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
where the platform was previously located was significantly higher than that spent in the other three quadrants ( P < .01 in all instances) and higher than the time spent by the rats of the other six groups in each quadrant ( P < .05 in all instances). In EF rats there were no significant differences between the time spent in the three previously unrewarding quadrants. In the other six groups there never were differences between the time spent in the four quadrants. The rats of groups VP, NER, and EF never showed ‘‘subcircular’’ swimming patterns. In contrast, a high number of rats (between 50% and 71%) of groups EF-S (four out of eight), EM (four out of seven), NEF (five out of eight), and NEM (five out of seven) exhibit a well-developed ‘‘subcircular’’ swimming pattern, as defined in the Materials and methods section (see Table 1).
4. Discussion The aim of the present work was to ascertain how efficiently the rats could reach the hidden platform of the MWM when deprived of the extramaze visual cues, i.e. when the animals could not build the standard ‘‘allothetic place representation.’’ In particular, the experimental design aimed at clarifying the efficiency of procedural learning based on tank shape and dimensions, and on platform placement in the tank (distance from tank wall). The present findings show that rats which have to reach the target (the underwater not visible platform) in a MWM when it has been made impossible for them to build an allothetic frame of spatial reference (NEF, NEM, EM, and EF-S groups) still learn to reach their goal, and not by chance. In other words, even under such severe conditions, the rats succeed in reaching the escape platform, only twice slower than with the standard procedure. This may be taken as indication of the good efficiency of their procedural learning. On the technical side, a few remarks may be made. The performance of the animals was quantitatively measured as escape latency, i.e. the time lapsed between the immersion of the animal in the tank and its reaching the platform. This parameter was chosen on the basis of the measured uniform swimming speed of all animals over all trials. In this way, it became unnecessary to measure the length of the swimming pattern, since there is a satisfactory proportionality between times and distances. The measured values of 33 – 36 cm/s confirm previous findings [32]. Scopolamine was administered at dosages known to disrupt spatial memory in the MWM paradigm [3,4,33,34]. Starting locations were randomly chosen on each trial, so as to minimize the possibility that the animals could profit from constant topographical relationships between starting site and hidden platform. Training was ended after Day 5, because in none of the seven experimental groups were there significant differences between escape latencies between Day 4 and Day 5 (Fig. 1). Probe test results on Day 6 showed that only rats of the EF (i.e. standard MWM paradigm) group show an allothetic orientation by their preference for one quadrant of the tank
(Fig. 3). It may be recalled that there are some reports on relatively satisfactory task solving in the MWM by rats variously deprived of extramaze visual cues [8 –11,35,36]. The present results both confirm and extend the previous ones by means of the analysis of a set of diverse experimental designs all proven to be effective in excluding visual extramaze cues from behavioral utilization. When deprived of visual cues, the rats may locate the goal on the basis of auditory [9,11,36] or olfactory [31] cues. Some have even proposed a hierarchical sensorial scale in decreasing order: visual, olfactory, and stimuli generated by the movements of the animal (proprioceptive or kinesthetic) [30]. In the present work, the extramaze visual cues were excluded by the curtains, the olfactory ones by the repeated stirring of the contents of the tank, and the auditory ones by the imposed white noise. Thus, to solve the task the rats had access only to the information originating from their motility and those deriving from the shape and dimensions of the tank and particularly on the constant distance of the platform from the wall of the tank (groups EM, NEF, NEM, and EF-S), i.e. they could employ only the general procedural mechanisms. As recalled in the Introduction, according to several authors, task solving in the standard MWM may be due not only to allothetic mechanisms but also to ‘‘procedural memory’’ [1,2,7,10 – 14]. In fact, when the animals are allowed to employ both mapping and nonmapping (allothetic and general procedural) mechanisms, they learn quite efficiently to reach their goal, the hidden platform (see the EF group performance, Figs. 1 and 2). Indeed, their performance is almost as good as that of the animals that have to reach a visible platform (VP group, Figs. 1 and 2). The results show that the EF group (standard MWM task) exhibited the best performance. EM, NEM, and NEF groups exhibited a lesser performance in terms of escape delay duration but still adequately reached the goal (Figs. 1 and 2). In other words, in extramaze allothetic deprivation conditions, task solving on purely general procedural mechanisms was more time consuming, but not conducive to an inferior performance as far as strictly finding the goal is concerned. As expected, when also the constant platform – tank wall distance information was denied (group NER), the rats could not progressively improve their performance. The performance of EM, NEM, and NEF groups is clearly due to a learning process, since over the days of training there is a progressive decrease of escape latency duration down to a constant value. In more detail, it can be underlined that the NEM and NEF groups reach the goal after swimming progressively less (escape latency duration), in both instances, in the absence of extramaze visual cues and without significant differences between groups. The absence of significant differences in escape latency durations between these groups shows that the deprivation of extramaze cues was satisfactorily efficient. In fact, when hidden platform location was kept constant, a better performance was not measured. On the other hand, for the EM group, extramaze visual cues were available but of no use in reaching the
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
hidden platform, since platform location was not constant, being varied on each trial. Thus, in these experimental conditions, the extramaze allothetic information was not sufficient for task solving [8,10,13,37]. The animals must rely only on general procedural mechanisms to reach the goal, which, as stated above, they do satisfactorily. The swimming patterns of these groups (EM, NEF, NEM) show that in these conditions the animals swim keeping themselves at a fairly constant distance from the wall of the tank, following an approximately circular path (Fig. 2, Table 1). This ‘‘subcircular’’ swimming pattern is almost universally present already on Day 3 at the end of the daily block of training (Table 1). In all these groups, swimming patterns and escape durations appear to be almost identical. The swimming behavior shows that the rats have learnt that the platform is located at a constant distance from the wall of the tank (the ‘‘where’’) and that the most efficient way to find it is to swim following a ‘‘subcircular pattern’’ (the ‘‘how’’). The finding that the ‘‘subcircular’’ swimming pattern is exhibited on the first trials of Days 4 and 5, and on the probe test (Day 6) shows that this behavior is not only maintained by working memory mechanisms developing during the daily block of 10 trials, but indeed, is a learned and consolidated behavior that can be recalled at 24-h intervals. The finding that NER group rats do not develop the ‘‘subcircular’’ swimming pattern shows that the constancy of the platform –tank wall relationship is necessary for developing the search strategy. These findings support the hypotheses advanced by some researchers [38] according to whom the animals find the goal using both procedural ‘‘spatial’’ and ‘‘nonspatial’’ cues and confirm and extend data reported for other species of rodents [39]. At this point, we must recall that allothetic is not an antonym to procedural. In fact, the procedural strategy is based also on an intramaze allothesis. It is true that a general procedure solution was used only when the training conditions did not offer to the rats any extramaze allothetic cues, making it possible to enter the position of the goal into their cognitive maps. On the other hand, it is not entirely true that procedural training was totally devoid of allothetic memories. In fact, swimming in a circular annulus from the center of the pool can only be accomplished by continuously monitoring the distance from the wall of the pool, i.e. by an intramaze allothetically guided behavior. Thus, ‘‘general procedure goal finding strategies’’ must not be considered completely independent from allothetic mechanisms. The swimming behavior of the EF-S group was quite similar to that of the other groups to which extramaze allothetic information was denied (EM, NEF, NEM). Thus, animals of the EF-S group reach the platform using a strategy based on general procedural mechanisms (Fig. 2, Table 1). Scopolamine disrupts spatial memory [3,4,33,34], and it is accepted that the central cholinergic systems are those that support spatial allothetic mapping (platform location), but not other superior nervous functions, such as those necessary
791
for nonextramaze visual searching strategies (how to find the platform) [3,40]. The behavior of the EF-S group confirms this hypothesis. The present findings show that the general procedure goal finding strategies can be evolved and employed in complete independence from extramaze allothetic information. This is shown both by the recorded swimming strategies, and by the finding that during the probe test none of these four experimental groups (NEM, NEF, EM, EF-S) exhibited any quadrant preference, i.e. they did not exhibit any behavior based on classically defined allothetic memories. Thus, under conditions of constant relationship of the goal to the contours of the pool, even without extramaze allothetic but with intramaze information, rats are able to elaborate a search strategy probably based both on praxis and intramaze allothetic solutions of the task. The presently employed paradigm does not allow us to assess the relative importance of either component in solving the task (NEM, NEF, EM, and EF-S groups). If the position of the goal was rotated by the same angle for every new position of the start (e.g. start/goal positions 0/90° on Swim 1 to 72/162° on Swim 2), probably in this case, the animals (employing a praxis strategy) would be able to find the platform with a shorter latency than in the present cases. On the other hand, even a direct goal approach from any start may be supported by intramaze allothesis alone when the goal place is uniquely determined by the contours of the pool, as when the goal is in the center of a circular pool. In general, we must recall that since it does not appear to be possible to devise an experimental paradigm by means of which the general procedure mechanisms can be excluded so that rats must solve the task of finding the goal using only the allothetic ones, it is not possible to know in the standard EF paradigm: (i) if only extramaze allothetic information or both are concurrently employed; (ii) if both, whether the two mechanisms concur equally to goal-finding or not; and (iii) if not, if there is a hierarchical relationship, or bias, between the employment of the two mechanisms. Nonetheless, some hypotheses have been formulated on these points. According to several authors [7,11,14,38,41] in the standard MWM paradigm, procedural learning is followed by allothetic learning. Within limits, the present results support this hypothesis. In fact, significant statistical differences between delay durations of the EF group and all the allotheticdeprived ones (NEF, NEM, EM, EF-S) are found only from Day 4 on, i.e. after general procedural mechanisms have been completely exploited. Moreover, it can be observed that at the 10th trial of Day 3 the ‘‘subcircular’’ swimming pattern was temporarily exhibited by some (25%) of the rats of the EF group (Table 1). On the following days, this pattern no longer occurred, all animals exhibiting the more efficient ‘‘allothetic goal-directed swim.’’ In fact, from Day 4 onwards, the allothetic-deprived groups did not exhibit a further shortening of escape delay, which was instead exhibited by the EF group, down to the same durations of the group which had to reach a visible platform (VP).
792
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793
In conclusion, the present findings show that in the MWM rats are able to evolve an efficient search strategy for goal reaching, even when they are denied the possibility of building an allothetic strategy based on extramaze visual cues. On the other hand, when sufficient intramaze information is available (constancy of platform – tank wall relationship), the rats evolve an adequate search strategy, in which a ‘‘subcircular’’ swimming pattern is used. This pattern, which is based on general procedural mechanisms, is evolved by most animals in these experimental conditions, and is remembered for a long time. As recalled in the Introduction, besides the primary role of the hippocampus in spatial learning, several other neural sites are thought to play a role in this type of learning. One question that is still to be addressed concerns which structures, sites, or systems of the central nervous system are more or less specifically involved in the distinct elaboration of allothetic and procedural learning [3,4]. It could be of some interest, on the basis of the present findings and by employing similar protocols, to try to assess the involvement of several subcortical and cortical sites at least in general procedures learning and utilization.
Acknowledgements The authors thank Mr. A. Aiazzi, Mr. S. Cammarata, Mr. M. Dolfi, and Mr. A. Vannucchi for their technical assistance.
References [1] Morris RGM. Spatial localization does not require the presence of local cues. Learn Motiv 1981;12:239 – 60. [2] Morris R. Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 1984;11:47 – 60. [3] Brandeis R, Brandys Y, Yehuda S. The use of the Morris water maze in the study of memory and learning. Int J Neurosci 1989;48:29 – 69. [4] D’Hooge R, De Deyn PP. Applications of the Morris water maze in the study of learning and memory. Brain Res Rev 2001;36:60 – 90. [5] Whishaw IQ. Place learning in hippocampal rats and the path integration hypothesis. Neurosci Biobehav Rev 1998;22:209 – 20. [6] DiMattia BD, Kesner RP. Spatial cognitive maps: differential role of parietal cortex and hippocampal formation. Behav Neurosci 1988;102: 471 – 80. [7] Fenton AA, Bures J. Place navigation in rats with unilateral tetrodotoxin inactivation of the dorsal hippocampus: place but not procedural learning can be lateralized to one hippocampus. Behav Neurosci 1993; 107:552 – 64. [8] Bannerman DM, Good MA, Butcher SP, Ramsay M, Morris RGM. Distinct components of spatial learning revealed by prior training and NMDA receptor blockade. Nature 1995;378:182 – 6. [9] Buresova O, Homuta L, Krekule I, Bures J. Does nondirectional signalization of target distance contribute to navigation in the Morris water maze? Behav Neural Biol 1988;49:240 – 8. [10] Leggio MG, Neri P, Graziano A, Mandolesi L, Molinari M, Petrosini L. Cerebellar contribution to spatial event processing: characterization of procedural learning. Exp Brain Res 1999;127:1 – 11. [11] Sutherland RJ, Dyck RH. Place navigation by rats in a swimming pool. Can J Psychol 1984;38:322 – 47.
[12] Schenk F, Morris RGM. Dissociation between components of spatial memory in rats after recovery from the effects of retrohippocampal lesions. Exp Brain Res 1985;58:11 – 28. [13] Whishaw IQ. Dissociating performance and learning deficits on spatial navigation tasks in rats subjected to cholinergic muscarinic blockade. Brain Res Bull 1989;23:347 – 58. [14] Whishaw IQ, Mittleman G. Visits to starts, routes and places by rats (Rattus norvegicus) in swimming pool navigation tasks. J Comp Psychol 1986;100:422 – 32. [15] Morris RGM, Frey U. Hippocampal synaptic plasticity: role in spatial learning or the automatic recording of attended experience? Philos Trans R Soc Lond B 1997;352:1489 – 503. [16] Squire LR. Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychol Rev 1992;99:195 – 231. [17] O’Keefe J, Nadel L. The hippocampus as a cognitive map. Oxford: Clarendon; 1978. [18] Nadel L. The hippocampus and space revisited. Hippocampus 1991;1: 221 – 9. [19] Steckler T, Drinkenburg WH, Sahgal A, Aggleton JP. Recognition memory in rats: II. Neuroanatomical substrates. Prog Neurobiol 1998;54:313 – 32. [20] Morris RGM, Garrud P, Rawlins JNP, O’Keefe J. Place navigation impaired in rats with hippocampal lesions. Nature 1982;297: 681 – 3. [21] Morris RGM, Schenk F, Tweedie F, Jarrard LE. Ibotenate lesions of hippocampus and/or subiculum: dissociating components of allocentric spatial learning. Eur J Neurosci 1990;2:1016 – 28. [22] Whishaw IQ, Cassel JC, Jarrard LE. Rats with fimbria – fornix lesions display a place response in a swimming pool: a dissociation between getting there and knowing where. J Neurosci 1995;15:5779 – 88. [23] Whishaw IQ, Jarrad LE. Evidence for extrahippocampal involvement in place learning and hippocampal involvement in path integration. Hippocampus 1996;6:513 – 24. [24] Kolb B, Sutherland RJ, Whishaw IQ. A comparison of the contributions of the frontal and parietal association cortex to spatial localizations in rats. Behav Neurosci 1983;97:13 – 27. [25] Mizumori SJ, Miya DY, Ward KE. Reversible inactivation of the lateral dorsal thalamus disrupts hippocampal place representation and impairs spatial learning. Brain Res 1994;644:168 – 74. [26] Whishaw IQ, Mittleman G, Bunch ST, Dunnett SB. Impairments in the acquisition, retention and selection of spatial navigation strategies after medial caudate – putamen lesions in rats. Behav Brain Res 1987; 24:125 – 38. [27] Lalonde R. Cerebellar contributions to instrumental learning. Neurosci Biobehav Rev 1994;18:161 – 70. [28] Petrosini L, Leggio MG, Molinari M. The cerebellum in the spatial problem solving: a co-star or a guest star? Prog Neurobiol 1998;56: 191 – 210. [29] Packard MG, McGaugh JL. Inactivation of hippocampus or caudate nucleus with lidocaine differentially affects expression of place and response learning. Neurobiol Learn Mem 1996;65:65 – 72. [30] Maaswinkel H, Whishaw IQ. Homing with locale, taxon, and dead reckoning strategies by foraging rats: sensory hierarchy in spatial navigation. Behav Brain Res 1999;99:143 – 52. [31] Means LW, Alexander SR, O’Neal MF. Those cheating rats: male and female rats use odor trails in a water-escape ‘‘working memory’’ task. Behav Neural Biol 1992;58:144 – 51. [32] Bannerman DM, Yee BK, Lemaire M, Wilbrecht L, Jarrard L, Iversen SD, et al. The role of the entorhinal cortex in two forms of spatial learning and memory. Exp Brain Res 2001;141:281 – 303. [33] Buresova O, Bolhuis JJ, Bures J. Differential effects of cholinergic blockade on performance of rats in the water tank navigation task and in a radial water maze. Behav Neurosci 1986;100:476 – 82. [34] McNamara RK, Skelton RW. The neuropharmacological and neurochemical basis of place learning in the Morris water maze. Brain Res Rev 1993;18:33 – 49. [35] Liu Z, Francis Turner L, Bures J. Impairment of place navigation of
E. Baldi et al. / Physiology & Behavior 78 (2003) 785–793 rats in the Morris water maze by intermittent light is inversely related to the duration of the flash. Neurosci Lett 1994;180:59 – 62. [36] Moghaddam M, Bures J. Contribution of egocentric spatial memory to place navigation of rats in the Morris water maze. Behav Brain Res 1996;78:121 – 9. [37] De Bruin JPC, Swinkels WAM, De Brabander JM. Response learning of rats in a Morris water maze: involvement of the medial prefrontal cortex. Behav Brain Res 1997;85:47 – 55. [38] Carman HM, Mactutus CF. Ontogeny of spatial navigation in rats: a role for response requirements? Behav Neurosci 2001;115:870 – 9.
793
[39] Dalm S, Grootendorst J, De Koet ER, Oitzl MS. Quantification of swim patterns in the Morris water maze. Behav Res Methods Instrum Comput 2000;32:134 – 9. [40] Janis SL, Glasier MM, Fulop Z, Stein DG. Intraseptal injections of 192 IgG saporin produce deficits for strategy selection in spatialmemory tasks. Behav Brain Res 1998;90:23 – 34. [41] Whishaw IQ, Petrie BF. Cholinergic blockade in the rat impairs strategy selection but not learning and retention of nonspatial visual discrimination problems in a swimming pool. Behav Neurosci 1988;102: 662 – 77.