- Email: [email protected]
Proximal versus distal cue utilization in preweanling spatial localization: the influence of cue number and location
Physiology & Behavior 79 (2003) 157 – 165
Proximal versus distal cue utilization in preweanling spatial localization: the influence of cue number and location Heidi M. Carmana,*, Rosemarie M. Boozeb, Diane M. Snowa, Charles F. Mactutusb a
Department of Anatomy and Neurobiology, College of Medicine, University of Kentucky, Chandler Medical Center, 800 Rose Street, Lexington, KY 40536-0298, USA b Department of Psychology, Barnwell College, University of South Carolina, Columbia, SC 29208, USA Received 19 November 2002; received in revised form 20 February 2003; accepted 26 February 2003
Abstract The present study was designed to examine the role of cue location and number in spatial navigation of the preweanling Fischer-344N rat in the Morris water maze using a protocol consistent with the pups’ response repertoire. The proximal (visible platform) versus distal (hidden platform) cue strategy was used, and spatial cues within the extramaze environment were configured such that the arrangement presented either a double cue or null cull condition relative to the platform location. All pups’ performance improved with training; however, probe trial performance, defined by quadrant time and platform crossings, revealed distal-double cue pups demonstrated spatial navigational ability superior to the remaining groups. This experimental dissociation suggests that a pup’s ability to spatially navigate a hidden platform is dependent on not only its response repertoire and task parameters but also its visual acuity, as determined by the number of extramaze cues and the location of these cues within the testing environment. The hidden versus visible platform dissociation may not be a satisfactory strategy for the control of potential sensorimotor deficits. D 2003 Elsevier Science Inc. All rights reserved. Keywords: Spatial navigation; Visual acuity; Proximal; Distal; Development; Ontogeny; Immature rats; Morris water maze
1. Introduction The strategy employing hidden versus visible platforms is putatively held to dissociate proximal and distal cue use in the Morris water maze. In this approach to spatial navigational study originally used by Morris [1,2] and Morris et al. [3], a local cue or cues are concomitant with a goal object such as a visible platform and comprise the proximal cue situation; whereas no cues are directly associated with the goal object and a hidden platform comprise the distal cue situation. In the distal cue situation, the configuration of several cues within the extramaze environment permits spatial localization. Because the hippocampus is believed to be the primary mediator in spatial navigation tasks using a hidden platform and distal cues, the visible platform/ proximal cue condition is characteristically used as a sensorimotor control to indicate that animals are otherwise capable of performing the task.
* Corresponding author. Tel.: +1-859-323-6112; fax: +1-859-323-5946. E-mail address: [email protected] (H.M. Carman).
The majority of the ontogenetic Morris water maze studies have used a similar approach to examine the development of proximal cue learning versus distal cue learning in preweanling spatial navigation (e.g., Refs. [4– 7]). With the exception of Brown and Whishaw [4], the general finding has been that proximal cue use develops prior to distal cue use suggesting that the deficit in preweanling spatial navigation results, in part, from an underdeveloped hippocampal formation [6,7]. However, the ages at which pups develop the ability to use proximal and distal information are conflicting [4 –7]. For example, the emergence of spatial information processes was initially reported at 42 days of age [7]. Rudy et al. [5] reported weanling rats must be 20– 23 days of age to demonstrate spatial memory. More recent work [8], however, has shown that spatial processing emerges as early as 17 days of age. The differences reported among this body of work are most likely the result of the various parameters employed. The specific training procedures differed considerably from Morris’ [1] original training regimen with one notable exception, the size of the pool. All of the ontogenetic studies, excluding Schenk [7], used a pool comparable to
0031-9384/03/$ – see front matter D 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0031-9384(03)00089-1
158
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
the size Morris [1] employed for adult spatial navigational studies. Despite the purported navigational success claimed in the various ontogenetic Morris water maze studies [5,6,7,9,10], there is little compelling data supporting the ability of the immature rat to spatially locate a hidden platform. For example, Rudy et al. [5] demonstrated that pups learned to find a hidden platform with a single illuminated cue located distally to the tank. Kraemer and Randall [10] also showed that preweanling pups located a hidden platform using a directional light gradient. While these tasks have spatial components, they do not demonstrate ‘‘true’’ spatial navigation requiring the use of a multiple cue configuration to localize the platform. These studies are better described as ‘‘beacon homing’’ or ‘‘landmark learning’’ where directional light cues indicate the platform location (cf. Ref. [11]). Recent work from our laboratory [8,11] suggests, however, that when the training regimen is consistent with that Morris [1] employed and the role of response requirements are sufficiently addressed without making the task easier, the preweanling rat can successfully use multiple distal cues to spatially navigate the location of a hidden platform. An implicit assumption among the extant ontogenetic water maze studies [4 – 10,12] is that the acuity of a preweanling’s visual system is sufficient to allow robust spatial navigation. Perhaps visual acuity in the preweanling rat was not previously investigated in the Morris water maze because of the presumption in the adult literature that the visual platform/proximal cue condition is sufficient to account for any potential sensory deficits present in subjects. Furthermore, when visual acuity in adult rats was explored, no support for deficits in visual acuity negatively impacting spatial navigation beyond initial acquisition was found (e.g., Ref. [13]). In stark contrast to the lack of effect of visual acuity on spatial navigation in adults and postweanling rats, recent work from our laboratory [14] suggests that visual acuity plays a significant role in the preweanling rat. Given the importance of this issue, it was imperative to replicate our previous work [14] to gain a better understanding of the ontogeny of spatial navigation and visual acuity in the spatial navigation capability of preweanling rats by experimentally manipulating the number and location of extramaze cues in the Morris water maze.
2. Methods 2.1. Subjects Eight litters of Fischer-344N rats, bred in the National Institutes of Health/National Institute of Environmental Health Sciences (NIH/NIEHS) vivarium, were used as subjects. The animals were maintained according to NIH guidelines in facilities accredited by the Association for
Assessment and Accreditation of Laboratory Animal Care (AAALAC). The vivarium was maintained at 21 ± 2 C, 50 ± 10% relative humidity and had a 12-h light – dark cycle with lights on at 07:00 h (EST). The pups were housed with their dams throughout the experiment. Food (NIH Diet #31) and water were available ad libitum. The animals’ day of birth, defined as postnatal day (PND) 0, is accurate within ± 1/2 day. 2.2. Apparatus The apparatus used during the experiment was a seamless, aluminum tank measuring 0.40 m in diameter by 0.25 m high. The water level was maintained at 0.20 m and the temperature was 29 ± 1 C. The escape platform measured 2.5 2.5 cm and during training was submerged 1.0 cm below the surface of the water for animals in the ‘‘distal’’ condition (hidden platform) but elevated by the addition of a 2.0-cm wet sponge for animals in the ‘‘proximal’’ condition (visible platform). Nondairy coffee creamer was added to the water to render it opaque. As previously reported [8] with this protocol, the platform was undetectable by sight based on human observations. The tank, in a 1.83 1.83 m room, was surrounded by two sets of cues. Four distinct background curtains comprised the walls of the environment: vertical navy and white stripes (7.5 cm wide), horizontal navy and white stripes (7.5 cm wide), solid navy, and solid white. The four curtains were moved forward to create a smaller, false room measuring 1.22 m 1.22 m. In addition, four objects were suspended from the ceiling outside the tank perimeter: a white plastic lid (6.4 cm diameter, 4.2 cm high) at the NW corner (single cue), a black plastic lid (6.4 cm diameter, 4.2 cm high) at the SE corner (single cue), an orange tennis ball (6.5 cm diameter), and a yellow sponge (8.3 3.8 15.2 cm) at the SW corner (double cues). No objects were suspended from the NE corner (null cue). See Fig. 1 for a schematic representation of the cue and platform configurations. Please note that the hanging objects are not drawn in proportion to the size of the tank, although the letters representing the platform are. For animals assigned to the distal condition, a paper cutout of a pointing hand (7.6 cm 15.25 cm at its widest and longest points) was hung in the center of the pool 6.0 cm above the surface of the water. This radially symmetric landmark was at a comparable distance to the hidden platform as the two closest extramaze cues. For animals assigned to the proximal condition, the pointing-hand cutout was 4.0 cm directly above the elevated platform; although hung from the same height in the ceiling, the addition of the 2.0-cm sponge to create a visible platform reduced the distance between the pointing-hand cutout and the platform. The hand was pointing to the correct quadrant in the proximal condition only; the hand was centered above the tank in the distal condition. The mean ambient light level across the water surface was 26.61 lx (S.D. = 0.65). A closed circuit video system, recessed mounted in the ceiling above
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
Fig. 1. Schematic representation of the pool with its respective cues for the null cues condition (Panel A) and double cues condition (Panel B). The compass point letters (N, E, S, W) represent the location and relative size of the platform (compared to the tank) for the proximal and distal conditions. The circle in the center represents the location of the paper cutout of a pointing hand, 6.0 cm above the surface of the water. For animals in the proximal-null cue and proximal-double cue conditions, the platform location was made visible by the addition of a 2.0-cm elevated wet sponge, and the paper cutout of a pointing hand was placed 4.0 cm directly above the visible platform. Critically, external cue configuration was identical for all animals regardless of condition. Note the absence of cues in the NE quadrant.
the center of the pool, recorded the swimming behavior of the animals. 2.3. Procedure 2.3.1. Design A hierarchical design was employed with litter nested within treatments. Specifically, litters were randomly assigned to one of two groups, ‘‘double cue’’ or ‘‘null cue,’’ designating the location of the platform. Within these groups, one male and one female from each litter were randomly assigned to each of two conditions, ‘‘proximal’’ and ‘‘distal’’ designating, relative to the external cue configuration whether the platform was visible (proximal) or hidden (distal). Thus, the independent variables were external cue configuration relative to platform location (double cue or null cue), platform condition (proximal or distal), and gender of the subjects. The animals began training on PND 17, the first day of fully opened eyes in this strain of rat [8,12].
159
2.3.2. Training All animals received 20 training trials across 3 days: eight trials on Days 1 and 2 followed by four training trials on Day 3. A probe test, described below, followed the last training trial on Day 3. At the beginning of each training trial, the rat was briefly held (3 – 5 s) over the water, facing outward, to view the particular testing area in which it was being placed for that trial such as the NW quadrant of the room. The pup was then gently placed into the water to avoid immersing its head, facing the wall of the pool, at position NE, NW, SE, or SW. The latency to find and climb onto the platform was recorded. If the animal failed to find the platform within 60 s, the trial ended and the rat was directed to the platform. The animal remained on the platform for 30 s and was then returned to a holding cage. To guard against potential hypothermia, animals were patted dry with towels between trials and placed in a clear, plastic holding cage over a heating pad with conspecifics in squads of four. One half of the cage was placed over a waterproof, hospital-grade heating pad (Casco, type WP-H) while the other half was not. This arrangement allowed pups to choose which side of the cage they preferred. Temperature directly over the heating pad was 34 C while the ambient temperature within the holding cage was 23 C. Because rats were trained four at a time, the intertrial interval was approximately 5 min. A rat started from one of two locations over the course of eight trials per day, and the sequence of start positions was counter balanced as Morris [1] originally employed. Platform location was balanced between the S and W quadrants for animals in the distaldouble cue condition and the N and E quadrants for animals in the distal-null cue condition as a result of the fixed orientation of the extramaze cues. However, platform location was balanced across all quadrants (N, E, S, W), and platform location changed on each trial for animals in the proximal condition. 2.3.3. Probe test Immediately following the last training trial on Day 3, the platform was removed from the pool; the visual– spatial environment was otherwise unaltered. The probe test was conducted as a training trial, and the animals’ swimming behavior was recorded for 60 s. The dependent variables measured include quadrant preference, i.e., the relative distribution of swimming time in each of the quadrants, and platform crossings, i.e., the number of crossings through the conceptual location of the platform in each of the quadrants. The comparisons of the target and opposite platform locations in the probe measures were closely examined because they provide an index of the animals’ spatial discriminability. Those means are represented in the figures against the means ± S.E.M. for the adjacent (left and right) quadrant preference and platform crossings. Additional comparisons contribute to the spatial discriminability index such as the comparison between the target and adjacent platform locations, which yields a preference
160
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
measure, and the comparison between the opposite and adjacent platform locations, which yields an evasion measure. Collectively, these three measures indicate the degree of spatial discriminability animals have developed within the testing environment. Typically, animals learn to discriminate, which is indicated by spending more time or making more platform crossings in the target quadrant relative to the opposite quadrant. In other words, the pups learn where to go in the pool. A preference for the target develops if pups spend the majority of test time in the target quadrant or cross the platform area more, relative to the adjacent quadrants. Lastly, pups that learn to avoid the opposite quadrant relative to the adjacent quadrants are demonstrating an evasion effect. In this example, pups have learned where not to go in the pool. 2.4. Data analysis An observer blind to the conditions of the task scored the videotapes of the trial and probe test sessions. All data were analyzed by analysis of variance (ANOVA) using BMDP Statistical Software, Release 7, Los Angeles, CA [15]. The average of the left and right quadrants was used for all analyses and figures related to quadrant time and platform crossings. The conservative Greenhouse – Geisser degrees of freedom correction factor [16] was used for violations of compound symmetry. Specific planned contrasts [17] were employed to evaluate potential group and trial-dependent effects, specifically discrimination, preference, and evasion for measures of spatial navigation. Additional planned comparisons between the groups included double cue versus null cue at the distal condition, double cue versus null cue at the proximal condition, and proximal versus distal at the double cue condition. An a level of P < 0.05 was considered significant for all statistical tests employed. The selection of the most appropriate error term for the hierarchical design was determined by conducting a preliminary test. No differences among litters were detected for any of the dependent variables ( P’s>.25); therefore MSsubjects was justifiably used as the error term to provide the most powerful statistical test for all remaining analyses [18].
3. Results 3.1. Acquisition Fig. 2 represents the mean escape latencies for the preweanling rats during acquisition, as a function of training trials. No significant gender differences were detected ( P > .05); thus, data were analyzed collapsed across this factor. Main effects of platform condition [proximal vs. distal, F(1,28) = 15.97, P < .0004] and external cue configuration were revealed [double vs. null cues, F(1,28) = 10.73, P <.0028]; these factors did not interact when collapsed across trials [ F(1,28) < 1.0]. Laten-
Fig. 2. Mean escape response latencies by group as a function of training trials for the 3 days of training. Error bars represent standard errors of the means ( ± S.E.M.).
cies to reach the platform decreased for all groups across the 20 training trials [ F(19,532) = 6.12, PGG < .0001; with a significant linear component F(1,28) = 50.76, P < .0001] indicating that the pups were learning to locate and escape from the water to the platform. A significant linear component was detected for each group of pups [thus, there was no interaction among trials, external cue configuration, and platform condition, F(19,532) < 1.0]; however, a significant quadratic component [ F(1,28) = 6.38, P < .0395] was also present in the proximal-double condition suggesting that they displayed the most rapid learning. A Trial External Cue Configuration interaction was revealed [ F(19,532) = 2.43, PGG < .0076] with animals in the double cue configuration learning to escape more rapidly than the rats trained with the null cue configuration. Specifically for the distal conditions, a Trial External Cue Condition interaction [ F(19,532) =1.95, PGG < .0366] was noted. For the proximal conditions, there was a consistent advantage throughout training to learning with the double cues [ F(1, 28) = 7.21, P < .0120]. An analysis by days demonstrated a significant main effect of external cue configuration (double vs. null) on Day 1 [ F(1,28) = 16.84, P < .0003], Day 2 [approached significance, F(1,28) = 3.42, P < .0752], and Day 3 [ F(1,28) = 8.07, P < .0083] of training, which suggests that extramaze cue placement was important for locating and escaping to the platform. The analysis by days also indicated a significant main effect of platform condition (proximal vs. distal) on Day 1 [ F(1,28) = 11.45, P < .0021] and Day 2 [ F(1,28) =12.68, P < .0013]; the visible platform was easier to locate than the hidden platform. On Day 3, the visible platform was easier to locate than the hidden platform for the double cue groups [ F(1,28) = 3.71, P < .0642] but not the null cue groups [ F(1,28) < 1.0]. Perhaps more importantly, the external cue configuration (double cue vs. null cue) had little effect on rats in the proximal condition [ F(1,28) =1.73, P < .20], but a
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
significant effect on rats in the distal condition [ F(1,28) = 7.30, P < .0116]. Looking more closely at performance on Day 3, the ANOVA for trial latencies showed that the terminal acquisition data were best described by a significant Trial External Configuration interaction [ F(3,84) = 5.40, P < .0027] with a cubic component for the Trial Cue Configuration in the proximal conditions [ F(1,28) = 10.01, P < .0037] and a quadratic component for the distal conditions [ F(1,28) = 7.01, P <.0132]. These results suggest that in the distal conditions, rats clearly relied upon the extramaze cues to spatially locate the platform; and in the proximal conditions, rats used the extramaze cues to locate the visible platform provided the cues were readily visible, i.e., near the location of the platform. 3.2. Probe test 3.2.1. Quadrant dwell time Fig. 3 depicts the mean time animals spent in the platform and opposite quadrants during the probe test. The ANOVA indicated a main effect of quadrant [ F(2,56) = 9.98, PGG < .0009] suggesting that the pups were treating the quadrants of the tank differently based on training. A
161
main effect of platform condition [ F(1,28) =4.42, P <.0446], but not external cue configuration [ F(1,28 < 1.0], was also revealed. While the External Cue Configuration Platform Condition [ F (1,28) = 1.86, P < .18] and the Quadrant External Cue Platform Condition [ F(2,56) = 1.26, PGG <.29] interactions were not significant, the Quadrant Platform Condition interaction was significant [ F(2,56) = 10.66, PGG < .0006]. Fig. 3 shows and the results confirm that the significant interaction was the result of animals trained in the distal conditions. Simple effects revealed a Quadrant Platform Condition interaction at the double cue condition: F(2,56) = 9.15, PGG < .0014; the Quadrant Platform Condition interaction at the null cue condition approached significance: F(2,56) =2.77, PGG < .0881. Pups trained in the proximal condition spent equivalent amounts of time in all quadrants, and the extramaze cue configuration had little to no effect on the proximal rats’ swim pattern [external cue at proximal: F(2,56) < 1.0]. The spatial discriminability index further characterized the animals’ behavior during the probe test. An overall significant discrimination effect was revealed [ F(1,28) = 15.08, P < .0006] indicating that particular animals were capable of distinguishing the target quadrant from the
Fig. 3. Mean time spent in the platform and opposite quadrants by group during the probe test. Error bars represent standard errors of the means. The mean ( ± S.E.M.) time spent in the quadrants adjacent to the platform is represented by the solid and dashed horizontal lines.
162
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
opposite quadrant. A significant Discrimination Platform Condition interaction [ F(1,28) = 14.86, P < .0006] suggested and individual analyses revealed that the effect was the result of animals from the distal conditions [distal-double cue: F(1,7) = 13.40, P < .0081; distal-null cue: F(1,7) = 9.34, P < .0184]. A similar pattern of results was revealed by the examination of the preference effect (target vs. adjacent quadrants): a significant effect of preference [ F(1,28) = 5.56, P < .0256], and a significant interaction between preference and platform condition [ F(1,28) =9.08, P < .0054]. This result was driven by animals in the distal-double cue condition [ F(1,7) = 8.15, P < .0245]. Although a significant overall evasion effect (adjacent vs. opposite quadrants) was revealed [ F(1,28) = 7.50, P < .0106], evasion did not interact with the remaining factors [external cue configuration: F(1,28) < 1.0; platform condition: F(1,28) =2.40, P < .13; or External Cue Configuration Platform Condition: F(1, 28) < 1.0]. Only animals in the distal-double condition displayed an evasion effect [ F(1,7) = 7.17, P <.0316]. 3.2.2. Platform crossings The mean platform crossing data by group are illustrated in Fig. 4. The ANOVA showed a significant main effect of
platform crossings [ F(2,56) = 6.57, PGG < .0039] but no main effects of cue configuration [ F(1,28) = 1.88, P < .18] or platform condition [ F(1,28) = 0.39, P < .54]. Platform crossings also failed to interact with cue configuration [ F(2,56) = 1.09, PGG < .34], platform condition [ F(2,56) = 1.47, PGG < .2], and Cue Configuration Platform Condition [ F(2,56) = 1.01, PGG < .3]. Nonetheless, specificplanned comparisons revealed a platform crossing effect at the distal [ F(2,56) = 7.11, PGG < .0026] but not at the proximal [ F(2,56) = 0.55, PGG < .56] cue conditions. This result indicates that pups in the distal conditions were learning to use the extramaze cues to navigate the platform location; however, the extramaze cues had no influence on navigation for pups in the proximal conditions. A platform crossing effect was also revealed at the double cue [ F(2,56) = 6.46, PGG < .0042] but not at the null cue [ F(2,56) = 1.19, PGG < .31] condition. Fig. 4 clearly shows that animals in the distal-double condition were more accurate in their search and made more crossings in the platform quadrant relative to the remaining quadrants compared to the other groups, implicating the animals’ ability to use the extramaze cues if the cues were near the location of the platform.
Fig. 4. Mean platform crossings within each quadrant by group during the probe test. Error bars represent standard errors of the means. The solid and dashed horizontal lines represent the mean platform crossings ± S.E.M. in the quadrants adjacent to the platform.
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
The spatial discriminability index indicated an overall significant discrimination [target vs. opposite, F(1,28) =9.57 P < .0045] and preference [target vs. adjacent, F(1,28) = 6.47, P < .0168] effect but not an evasion effect [opposite vs. adjacent, F(1,28) = 1.29 P < .26]. The data in Fig. 4 indicate and individual analyses confirmed that these results are from animals trained in the distal-double cue condition [discrimination: F(1,7) = 10.18, P < .0153; preference: F(1,7) = 8.76, P < .0211]. Cue configuration and platform condition failed to interact with discrimination, preference, or evasion ( P>.05).
4. Discussion The results of the present experiment support our previous work [14] and suggest that cue location and number are important in the preweanling rats’ ability to spatially navigate the location of a platform. This conclusion is derived from clear dissociations as a function of extramaze cue condition in probe test performance, namely the differential periods of time spent in the quadrants, and the respective platform crossing profiles. The more sensitive test measures, time in quadrant and platform crossings, indicated that the rats in the four groups were learning differently within their respective spatial environments. Rats in the proximal-null and proximal-double cue conditions spent equivalent amounts of time in all the quadrants, indicating they failed to discriminate the platform location among the quadrants, did not develop a preference for the target quadrant nor learned to avoid the opposite quadrant. The lack of spatial discriminability suggests that the pups learned little regarding the surrounding environment and its relationship to the platform’s location. This finding contrasts with our previous work [14], which indicated that rats trained in the proximal-double cue condition were using the extramaze distal cues to search for the visible platform on the probe test. However, the proximal-double cue pups’ behavior during acquisition indicated that the extramaze cues were influencing their search for the platform, a finding consistent with our previous work [14]. Moreover, all groups demonstrated improved performance across training trials as indicated by the decrease in escape latencies. Rats trained in the proximal conditions, particularly the double cue condition, demonstrated the fastest rates of escape compared to the distal conditions suggesting that the visibility of the platform played a significant role in localization. One factor that may have contributed to the different results between experiments is procedural in nature and applies to animals in the proximal conditions. In the previous experiment [14], the platform location remained the same during every trial whereas in the current experiment, the platform location changed every other trial. Perhaps maintaining the same location allowed animals to develop turn preferences when learning to locate the platform, giving the appearance that the rats were learning to
163
use the extramaze cues in the proximal-double cue condition. Whereas changing platform location from trial to trial eliminated the possibility of developing any turn preferences and the possibility of using the configuration of extramaze cues, it nonetheless created a ‘‘truer’’ test of platform visibility. Rats in the distal-null and distal-double cue conditions behaved differently than rats in the proximal conditions. These pups spent significantly more time in the target quadrant relative to the opposite quadrant indicating discrimination. Moreover, animals in the distal-double cue group demonstrated more proficient use of the cues by spending more time in the target quadrant relative to the adjacent quadrants indicating a preference. Neither distal condition demonstrated evasion, which characterizes the performance of adult and postweanling rats [14]. Because the distal-double cue pups, but not the distal-null cue pups, developed a preference of location suggests that adept use of the extramaze cues was constrained by the pups’ visual acuity. Nonetheless, the cues were sufficiently visible for the rats to discriminate, and thus learn the spatial location of the platform in relationship to the configuration of multiple cues provided. The platform crossing data, the probe measure for navigational accuracy, revealed more specifically the manner in which animals learned about their spatial environment. Although rats in the distal-null cue condition crossed the original platform location slightly more frequently than the conceptual locations of the adjacent and opposite platforms, they were less accurate and failed to spatially discriminate among these areas. For rats in the proximal-null cue condition, the frequency of crossings was equivalent regardless of location indicating external cues had little to no influence on their navigational behavior. Rats in the proximal-double cue condition made more overall crossings, but surprisingly more crossings were over the original platform location relative to the opposite platform location. This finding is consistent with our previous results [14] and suggests that the proximal-double cue rats may have been using the external spatial cues to aid in the spatial navigation of the visible platform. While rats in the distalnull cue condition made more crossings over the original platform location relative to the opposite quadrant, rats in the distal-double cue condition crossed the platform’s original location significantly more relative to the opposite quadrant location, demonstrating the most proficient use of the extramaze cues. This outcome suggests that the rat’s ability to use the distal cues in the extramaze environment was enhanced by presenting them in the double cue configuration, presumably a more visible arrangement of the cues. It is our view that the different measures, i.e., latency to find the platform, time in quadrant, and number of platform crossings, are tapping into the different learning and memory processes associated with spatial navigation. For example, training trial latencies reflect the animals’ ability to learn the task, i.e., to find and climb onto the platform
164
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
whether it is hidden or visible. However, marked changes in escape latencies are possible under conditions that support no true spatial learning, i.e., when there are no consistent relationships between the escape platform and extramaze cues. This has been demonstrated previously in adult animals (e.g., Ref. [8]) and currently in preweanlings (proximal condition). Nonetheless, the reduction in trial latencies across time appears to be necessary but not sufficient for spatial navigation. The probe trial measures, dwell time (time in quadrant/quadrant preference), and platform crossings allow one to examine the animals’ memory for the platform’s previous location at different levels of spatial accuracy. By examining the relative distribution of swimming time in each of the quadrants, the animal is indicating its preference for a particular area of the pool based on its training history. In addition, the number of platform crossings through the location previously containing the platform indicates how accurately (or not) the animal learned the prior platform location and its relationship with the configuration of extramaze cues. Comparisons made among the target, opposite, and adjacent platform locations provide a statistical index of the animals’ ability to spatially discriminate among pool locations. These comparisons demonstrate what the animal recalled during the probe test. The comparison between the target and opposite platform locations yields a discrimination measure; the comparison between the target and adjacent platform locations yields a preference measure; and the comparison between the opposite and adjacent platform locations yields an evasion measure (the same statistical comparisons were made between the various quadrants with swim time measures). Collectively, these three measures— discrimination, preference, and evasion—indicate the degree of spatial discriminability animals have developed within the testing environment. Typically, animals learn to discriminate as indicated by spending more time or making more platform crossings in the target quadrant relative to the opposite quadrant, i.e., the pups learn where to go in the pool. A preference for the target develops if pups spend the majority of test time in the target quadrant or cross the platform area more, relative to the adjacent quadrants. Pups that learn to avoid the opposite quadrant relative to the adjacent quadrants are demonstrating the evasion effect, i.e., pups learn where not to go in the pool. The acuity of the preweanling’s visual system and the role it plays in spatial navigation is an issue that only recently has been addressed. An implicit assumption of previous ontogenetic water maze studies [4 – 10,12] has been that the animal’s visual system is sufficiently formed to permit navigation, perhaps a result of the adult literature, which has also presumed that the visual platform/proximal cue condition is a sufficient control for potential sensory deficits. However, previously [14] and here, we show that visual acuity in the preweanling rat plays a significant role in its ability to spatially navigate the location of a platform. Thus, the standard use of the visible platform/proximal cue
is not a sufficient control for visual acuity in the preweanling rat. We also know from previous work that visual pattern experience contributes to the young animal’s performance in the Morris water maze [8,12]. The degree of eye opening was assessed in rats 16 days of age and correlated with the respective mean escape latencies on the first day of training (PND 17). A significant inverse relationship was revealed: as the degree of eye opening increased, escape latencies decreased across animals [12]. Though individual visual acuity was not directly assessed in the experiment reported here, it is clear from the manipulation of cues within the extramaze environment that acuity played a role in the animal’s probe trial performance. The experimental manipulation of visual acuity has been previously studied in adult rats in the Morris water maze. Prusky et al. [13] reported that binocularly deprived rats were significantly worse than controls on the visual acuity measure. Importantly, the sight-deprived rats were significantly slower in acquiring the Morris water maze task. Surprisingly, however, the critical probe test measure of quadrant time indicated equivalent performance between groups suggesting that animals in either condition learned the location of the platform equally well despite significant differences in visual acuity. Visual acuity in potentially compromised subpopulations, e.g., aged rats, has also been examined. Because the age-related decline in visual acuity of rats is well documented (e.g., Refs. [19,20]), Lindner and Gribkoff [21] assessed visual acuity in aged rats after spatial training by adding a high contrast, proximal (visual) cue to the hidden platform. The rats were then trained to asymptotic performance. Despite reporting deficits in visual acuity among aged rats, this study [20] failed to find a correlation between visual acuity and spatial learning, which was defined by swimming distance (length of the path from its initial release to the platform). However, Spencer et al. [22], assessed the relationship between retinal degeneration and spatial ability in aged rats more directly. Following an extensive spatial training regiment, rats were classified either as ‘‘learners’’ or ‘‘nonlearners’’ based on performance. Histological and morphometric measures of the retinas were conducted. Among aged rats with the least amount of retinal degradation, 40 of 41 were classified as spatial learners, whereas among aged rats with the most retinal degeneration, 20 of 27 were classified as nonlearners. These results [22] suggest that cognitive performance measures that rely heavily upon visual cues, such as the water maze, may be confounded by visual impairment in aged animals. The results reported here replicate and confirm our previous findings [14] and suggest that the typical use of a visible platform or proximal cues is not sufficient for visual acuity control in the preweanling animal. Although individual visual acuity was not directly assessed, the manipulation of cues within the extramaze test environment shows a clear experimental dissociation indicating that visual acuity played a significant role in the preweanling’s
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
ability to spatially navigate the water maze. The roles of response requirements [8,12] and visual acuity represent critical factors in elucidating the ontogeny of spatial navigation in the rat.
Acknowledgements Heidi M. Carman was supported by NIDA training grant, DA07304. This work was supported by grants from the National Institute of Drug Abuse (DA09160, DA012719, DA12719) and the National Institute of Environmental Health Sciences (ES06259). The authors would like to thank Debra Murray for excellent technical assistance.
References [1] Morris RGM. Spatial localization does not require the presence of local cues. Learn Motiv 1981;12:239 – 60. [2] Morris RGM. An attempt to dissociate ‘‘spatial-mapping’’ and ‘‘working memory’’ theories of hippocampal function. In: Seifert W, editor. Neurobiology of the hippocampus. New York: Academic Press; 1983. p. 405 – 32. [3] Morris RGM, Garrud P, Rawlins JNP, O’Keefe J. Place navigation in rats with hippocampal lesions. Nature 1982;297:681 – 3. [4] Brown RW, Whishaw IQ. Similarities in the development of place and cue navigation by rats in a swimming pool. Dev Psychobiol 2000;37: 238 – 45. [5] Rudy JW, Stadler-Morris S, Albert P. Ontogeny of spatial navigation behaviors in the rat: dissociation of ‘‘proximal-’’ and ‘‘distal-’’ cuebased behaviors. Behav Neurosci 1987;101:62 – 73. [6] Rudy JW, Paylor R. Reducing the temporal demands of the Morris place-learning task fails to ameliorate the place-learning impairment of preweanling rats. Psychobiology 1988;16:152 – 6. [7] Schenk F. Development of place navigation in rats from weaning to puberty. Behav Neural Biol 1985;43:69 – 85.
165
[8] Carman HM, Mactutus CF. Ontogeny of spatial navigation in rats: a role for response requirements? Behav Neurosci 2001;115:870 – 9. [9] Brown RW, Kraemer PJ. Ontogenetic differences in retention of spatial learning tested with the Morris water maze. Dev Psychobiol 1997;30:329 – 41. [10] Kraemer PJ, Randall CK. Spatial learning in preweanling rats trained in a Morris water maze. Psychobiology 1995;23:144 – 52. [11] Roberts WA. Principles of animal cognition. Boston (MA): McGrawHill; 1998. [12] Carman HM, Booze RM, Mactutus CF. Long-term retention of spatial navigation by preweanling rats. Dev Psychobiol 2002;40:68 – 77. [13] Prusky GT, West PWR, Douglas RM. Reduced visual acuity impairs place but not cued learning in the Morris water task. Behav Brain Res 2000;116:135 – 40. [14] Carman HM, Mactutus CF. Proximal versus distal cue utilization in spatial navigation: the role of visual acuity? Neurobiol Learn Mem 2002;78:332 – 46. [15] BMDP Statistical Software. Release 7 [Computer software]. Los Angeles (CA): University of California Press; 1993. [16] Greenhouse SW, Geisser S. On methods in the analysis of profile data. Psychometrika 1959;24:95 – 112. [17] Winer BJ. Statistical principles in experimental design. 2nd ed. New York: McGraw-Hill; 1971. [18] Denenberg VH. Statistics and experimental design for behavioral and biological researchers. New York: Wiley; 1976. [19] O’Steen WK, Sweatt AJ, Brodish A. Effects of acute and chronic stress on the neural retina of young, mid-age, and aged Fischer-344 rats. Brain Res 1987;426:37 – 46. [20] Lindner MD, Plone MA, Schallert T, Emerich WF. Blind rats are not profoundly impaired in the reference memory Morris water maze and cannot be clearly discriminated from rats with cognitive deficits in the cued platform task. Cogn Brain Res 1997;5:329 – 33. [21] Lindner MD, Gribkoff VK. Relationship between performance in the Morris water task, visual acuity and thermoregulatory function in aged F-344 rats. Behav Brain Res 1991;45:45 – 55. [22] Spencer RL, O’Steen WK, McEwen BS. Water maze performance of aged Sprague – Dawley rats in relation to retinal morphologic measures. Behav Brain Res 1995;68:139 – 50.
Proximal versus distal cue utilization in preweanling spatial localization: the influence of cue number and location Heidi M. Carmana,*, Rosemarie M. Boozeb, Diane M. Snowa, Charles F. Mactutusb a
Department of Anatomy and Neurobiology, College of Medicine, University of Kentucky, Chandler Medical Center, 800 Rose Street, Lexington, KY 40536-0298, USA b Department of Psychology, Barnwell College, University of South Carolina, Columbia, SC 29208, USA Received 19 November 2002; received in revised form 20 February 2003; accepted 26 February 2003
Abstract The present study was designed to examine the role of cue location and number in spatial navigation of the preweanling Fischer-344N rat in the Morris water maze using a protocol consistent with the pups’ response repertoire. The proximal (visible platform) versus distal (hidden platform) cue strategy was used, and spatial cues within the extramaze environment were configured such that the arrangement presented either a double cue or null cull condition relative to the platform location. All pups’ performance improved with training; however, probe trial performance, defined by quadrant time and platform crossings, revealed distal-double cue pups demonstrated spatial navigational ability superior to the remaining groups. This experimental dissociation suggests that a pup’s ability to spatially navigate a hidden platform is dependent on not only its response repertoire and task parameters but also its visual acuity, as determined by the number of extramaze cues and the location of these cues within the testing environment. The hidden versus visible platform dissociation may not be a satisfactory strategy for the control of potential sensorimotor deficits. D 2003 Elsevier Science Inc. All rights reserved. Keywords: Spatial navigation; Visual acuity; Proximal; Distal; Development; Ontogeny; Immature rats; Morris water maze
1. Introduction The strategy employing hidden versus visible platforms is putatively held to dissociate proximal and distal cue use in the Morris water maze. In this approach to spatial navigational study originally used by Morris [1,2] and Morris et al. [3], a local cue or cues are concomitant with a goal object such as a visible platform and comprise the proximal cue situation; whereas no cues are directly associated with the goal object and a hidden platform comprise the distal cue situation. In the distal cue situation, the configuration of several cues within the extramaze environment permits spatial localization. Because the hippocampus is believed to be the primary mediator in spatial navigation tasks using a hidden platform and distal cues, the visible platform/ proximal cue condition is characteristically used as a sensorimotor control to indicate that animals are otherwise capable of performing the task.
* Corresponding author. Tel.: +1-859-323-6112; fax: +1-859-323-5946. E-mail address: [email protected] (H.M. Carman).
The majority of the ontogenetic Morris water maze studies have used a similar approach to examine the development of proximal cue learning versus distal cue learning in preweanling spatial navigation (e.g., Refs. [4– 7]). With the exception of Brown and Whishaw [4], the general finding has been that proximal cue use develops prior to distal cue use suggesting that the deficit in preweanling spatial navigation results, in part, from an underdeveloped hippocampal formation [6,7]. However, the ages at which pups develop the ability to use proximal and distal information are conflicting [4 –7]. For example, the emergence of spatial information processes was initially reported at 42 days of age [7]. Rudy et al. [5] reported weanling rats must be 20– 23 days of age to demonstrate spatial memory. More recent work [8], however, has shown that spatial processing emerges as early as 17 days of age. The differences reported among this body of work are most likely the result of the various parameters employed. The specific training procedures differed considerably from Morris’ [1] original training regimen with one notable exception, the size of the pool. All of the ontogenetic studies, excluding Schenk [7], used a pool comparable to
0031-9384/03/$ – see front matter D 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0031-9384(03)00089-1
158
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
the size Morris [1] employed for adult spatial navigational studies. Despite the purported navigational success claimed in the various ontogenetic Morris water maze studies [5,6,7,9,10], there is little compelling data supporting the ability of the immature rat to spatially locate a hidden platform. For example, Rudy et al. [5] demonstrated that pups learned to find a hidden platform with a single illuminated cue located distally to the tank. Kraemer and Randall [10] also showed that preweanling pups located a hidden platform using a directional light gradient. While these tasks have spatial components, they do not demonstrate ‘‘true’’ spatial navigation requiring the use of a multiple cue configuration to localize the platform. These studies are better described as ‘‘beacon homing’’ or ‘‘landmark learning’’ where directional light cues indicate the platform location (cf. Ref. [11]). Recent work from our laboratory [8,11] suggests, however, that when the training regimen is consistent with that Morris [1] employed and the role of response requirements are sufficiently addressed without making the task easier, the preweanling rat can successfully use multiple distal cues to spatially navigate the location of a hidden platform. An implicit assumption among the extant ontogenetic water maze studies [4 – 10,12] is that the acuity of a preweanling’s visual system is sufficient to allow robust spatial navigation. Perhaps visual acuity in the preweanling rat was not previously investigated in the Morris water maze because of the presumption in the adult literature that the visual platform/proximal cue condition is sufficient to account for any potential sensory deficits present in subjects. Furthermore, when visual acuity in adult rats was explored, no support for deficits in visual acuity negatively impacting spatial navigation beyond initial acquisition was found (e.g., Ref. [13]). In stark contrast to the lack of effect of visual acuity on spatial navigation in adults and postweanling rats, recent work from our laboratory [14] suggests that visual acuity plays a significant role in the preweanling rat. Given the importance of this issue, it was imperative to replicate our previous work [14] to gain a better understanding of the ontogeny of spatial navigation and visual acuity in the spatial navigation capability of preweanling rats by experimentally manipulating the number and location of extramaze cues in the Morris water maze.
2. Methods 2.1. Subjects Eight litters of Fischer-344N rats, bred in the National Institutes of Health/National Institute of Environmental Health Sciences (NIH/NIEHS) vivarium, were used as subjects. The animals were maintained according to NIH guidelines in facilities accredited by the Association for
Assessment and Accreditation of Laboratory Animal Care (AAALAC). The vivarium was maintained at 21 ± 2 C, 50 ± 10% relative humidity and had a 12-h light – dark cycle with lights on at 07:00 h (EST). The pups were housed with their dams throughout the experiment. Food (NIH Diet #31) and water were available ad libitum. The animals’ day of birth, defined as postnatal day (PND) 0, is accurate within ± 1/2 day. 2.2. Apparatus The apparatus used during the experiment was a seamless, aluminum tank measuring 0.40 m in diameter by 0.25 m high. The water level was maintained at 0.20 m and the temperature was 29 ± 1 C. The escape platform measured 2.5 2.5 cm and during training was submerged 1.0 cm below the surface of the water for animals in the ‘‘distal’’ condition (hidden platform) but elevated by the addition of a 2.0-cm wet sponge for animals in the ‘‘proximal’’ condition (visible platform). Nondairy coffee creamer was added to the water to render it opaque. As previously reported [8] with this protocol, the platform was undetectable by sight based on human observations. The tank, in a 1.83 1.83 m room, was surrounded by two sets of cues. Four distinct background curtains comprised the walls of the environment: vertical navy and white stripes (7.5 cm wide), horizontal navy and white stripes (7.5 cm wide), solid navy, and solid white. The four curtains were moved forward to create a smaller, false room measuring 1.22 m 1.22 m. In addition, four objects were suspended from the ceiling outside the tank perimeter: a white plastic lid (6.4 cm diameter, 4.2 cm high) at the NW corner (single cue), a black plastic lid (6.4 cm diameter, 4.2 cm high) at the SE corner (single cue), an orange tennis ball (6.5 cm diameter), and a yellow sponge (8.3 3.8 15.2 cm) at the SW corner (double cues). No objects were suspended from the NE corner (null cue). See Fig. 1 for a schematic representation of the cue and platform configurations. Please note that the hanging objects are not drawn in proportion to the size of the tank, although the letters representing the platform are. For animals assigned to the distal condition, a paper cutout of a pointing hand (7.6 cm 15.25 cm at its widest and longest points) was hung in the center of the pool 6.0 cm above the surface of the water. This radially symmetric landmark was at a comparable distance to the hidden platform as the two closest extramaze cues. For animals assigned to the proximal condition, the pointing-hand cutout was 4.0 cm directly above the elevated platform; although hung from the same height in the ceiling, the addition of the 2.0-cm sponge to create a visible platform reduced the distance between the pointing-hand cutout and the platform. The hand was pointing to the correct quadrant in the proximal condition only; the hand was centered above the tank in the distal condition. The mean ambient light level across the water surface was 26.61 lx (S.D. = 0.65). A closed circuit video system, recessed mounted in the ceiling above
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
Fig. 1. Schematic representation of the pool with its respective cues for the null cues condition (Panel A) and double cues condition (Panel B). The compass point letters (N, E, S, W) represent the location and relative size of the platform (compared to the tank) for the proximal and distal conditions. The circle in the center represents the location of the paper cutout of a pointing hand, 6.0 cm above the surface of the water. For animals in the proximal-null cue and proximal-double cue conditions, the platform location was made visible by the addition of a 2.0-cm elevated wet sponge, and the paper cutout of a pointing hand was placed 4.0 cm directly above the visible platform. Critically, external cue configuration was identical for all animals regardless of condition. Note the absence of cues in the NE quadrant.
the center of the pool, recorded the swimming behavior of the animals. 2.3. Procedure 2.3.1. Design A hierarchical design was employed with litter nested within treatments. Specifically, litters were randomly assigned to one of two groups, ‘‘double cue’’ or ‘‘null cue,’’ designating the location of the platform. Within these groups, one male and one female from each litter were randomly assigned to each of two conditions, ‘‘proximal’’ and ‘‘distal’’ designating, relative to the external cue configuration whether the platform was visible (proximal) or hidden (distal). Thus, the independent variables were external cue configuration relative to platform location (double cue or null cue), platform condition (proximal or distal), and gender of the subjects. The animals began training on PND 17, the first day of fully opened eyes in this strain of rat [8,12].
159
2.3.2. Training All animals received 20 training trials across 3 days: eight trials on Days 1 and 2 followed by four training trials on Day 3. A probe test, described below, followed the last training trial on Day 3. At the beginning of each training trial, the rat was briefly held (3 – 5 s) over the water, facing outward, to view the particular testing area in which it was being placed for that trial such as the NW quadrant of the room. The pup was then gently placed into the water to avoid immersing its head, facing the wall of the pool, at position NE, NW, SE, or SW. The latency to find and climb onto the platform was recorded. If the animal failed to find the platform within 60 s, the trial ended and the rat was directed to the platform. The animal remained on the platform for 30 s and was then returned to a holding cage. To guard against potential hypothermia, animals were patted dry with towels between trials and placed in a clear, plastic holding cage over a heating pad with conspecifics in squads of four. One half of the cage was placed over a waterproof, hospital-grade heating pad (Casco, type WP-H) while the other half was not. This arrangement allowed pups to choose which side of the cage they preferred. Temperature directly over the heating pad was 34 C while the ambient temperature within the holding cage was 23 C. Because rats were trained four at a time, the intertrial interval was approximately 5 min. A rat started from one of two locations over the course of eight trials per day, and the sequence of start positions was counter balanced as Morris [1] originally employed. Platform location was balanced between the S and W quadrants for animals in the distaldouble cue condition and the N and E quadrants for animals in the distal-null cue condition as a result of the fixed orientation of the extramaze cues. However, platform location was balanced across all quadrants (N, E, S, W), and platform location changed on each trial for animals in the proximal condition. 2.3.3. Probe test Immediately following the last training trial on Day 3, the platform was removed from the pool; the visual– spatial environment was otherwise unaltered. The probe test was conducted as a training trial, and the animals’ swimming behavior was recorded for 60 s. The dependent variables measured include quadrant preference, i.e., the relative distribution of swimming time in each of the quadrants, and platform crossings, i.e., the number of crossings through the conceptual location of the platform in each of the quadrants. The comparisons of the target and opposite platform locations in the probe measures were closely examined because they provide an index of the animals’ spatial discriminability. Those means are represented in the figures against the means ± S.E.M. for the adjacent (left and right) quadrant preference and platform crossings. Additional comparisons contribute to the spatial discriminability index such as the comparison between the target and adjacent platform locations, which yields a preference
160
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
measure, and the comparison between the opposite and adjacent platform locations, which yields an evasion measure. Collectively, these three measures indicate the degree of spatial discriminability animals have developed within the testing environment. Typically, animals learn to discriminate, which is indicated by spending more time or making more platform crossings in the target quadrant relative to the opposite quadrant. In other words, the pups learn where to go in the pool. A preference for the target develops if pups spend the majority of test time in the target quadrant or cross the platform area more, relative to the adjacent quadrants. Lastly, pups that learn to avoid the opposite quadrant relative to the adjacent quadrants are demonstrating an evasion effect. In this example, pups have learned where not to go in the pool. 2.4. Data analysis An observer blind to the conditions of the task scored the videotapes of the trial and probe test sessions. All data were analyzed by analysis of variance (ANOVA) using BMDP Statistical Software, Release 7, Los Angeles, CA [15]. The average of the left and right quadrants was used for all analyses and figures related to quadrant time and platform crossings. The conservative Greenhouse – Geisser degrees of freedom correction factor [16] was used for violations of compound symmetry. Specific planned contrasts [17] were employed to evaluate potential group and trial-dependent effects, specifically discrimination, preference, and evasion for measures of spatial navigation. Additional planned comparisons between the groups included double cue versus null cue at the distal condition, double cue versus null cue at the proximal condition, and proximal versus distal at the double cue condition. An a level of P < 0.05 was considered significant for all statistical tests employed. The selection of the most appropriate error term for the hierarchical design was determined by conducting a preliminary test. No differences among litters were detected for any of the dependent variables ( P’s>.25); therefore MSsubjects was justifiably used as the error term to provide the most powerful statistical test for all remaining analyses [18].
3. Results 3.1. Acquisition Fig. 2 represents the mean escape latencies for the preweanling rats during acquisition, as a function of training trials. No significant gender differences were detected ( P > .05); thus, data were analyzed collapsed across this factor. Main effects of platform condition [proximal vs. distal, F(1,28) = 15.97, P < .0004] and external cue configuration were revealed [double vs. null cues, F(1,28) = 10.73, P <.0028]; these factors did not interact when collapsed across trials [ F(1,28) < 1.0]. Laten-
Fig. 2. Mean escape response latencies by group as a function of training trials for the 3 days of training. Error bars represent standard errors of the means ( ± S.E.M.).
cies to reach the platform decreased for all groups across the 20 training trials [ F(19,532) = 6.12, PGG < .0001; with a significant linear component F(1,28) = 50.76, P < .0001] indicating that the pups were learning to locate and escape from the water to the platform. A significant linear component was detected for each group of pups [thus, there was no interaction among trials, external cue configuration, and platform condition, F(19,532) < 1.0]; however, a significant quadratic component [ F(1,28) = 6.38, P < .0395] was also present in the proximal-double condition suggesting that they displayed the most rapid learning. A Trial External Cue Configuration interaction was revealed [ F(19,532) = 2.43, PGG < .0076] with animals in the double cue configuration learning to escape more rapidly than the rats trained with the null cue configuration. Specifically for the distal conditions, a Trial External Cue Condition interaction [ F(19,532) =1.95, PGG < .0366] was noted. For the proximal conditions, there was a consistent advantage throughout training to learning with the double cues [ F(1, 28) = 7.21, P < .0120]. An analysis by days demonstrated a significant main effect of external cue configuration (double vs. null) on Day 1 [ F(1,28) = 16.84, P < .0003], Day 2 [approached significance, F(1,28) = 3.42, P < .0752], and Day 3 [ F(1,28) = 8.07, P < .0083] of training, which suggests that extramaze cue placement was important for locating and escaping to the platform. The analysis by days also indicated a significant main effect of platform condition (proximal vs. distal) on Day 1 [ F(1,28) = 11.45, P < .0021] and Day 2 [ F(1,28) =12.68, P < .0013]; the visible platform was easier to locate than the hidden platform. On Day 3, the visible platform was easier to locate than the hidden platform for the double cue groups [ F(1,28) = 3.71, P < .0642] but not the null cue groups [ F(1,28) < 1.0]. Perhaps more importantly, the external cue configuration (double cue vs. null cue) had little effect on rats in the proximal condition [ F(1,28) =1.73, P < .20], but a
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
significant effect on rats in the distal condition [ F(1,28) = 7.30, P < .0116]. Looking more closely at performance on Day 3, the ANOVA for trial latencies showed that the terminal acquisition data were best described by a significant Trial External Configuration interaction [ F(3,84) = 5.40, P < .0027] with a cubic component for the Trial Cue Configuration in the proximal conditions [ F(1,28) = 10.01, P < .0037] and a quadratic component for the distal conditions [ F(1,28) = 7.01, P <.0132]. These results suggest that in the distal conditions, rats clearly relied upon the extramaze cues to spatially locate the platform; and in the proximal conditions, rats used the extramaze cues to locate the visible platform provided the cues were readily visible, i.e., near the location of the platform. 3.2. Probe test 3.2.1. Quadrant dwell time Fig. 3 depicts the mean time animals spent in the platform and opposite quadrants during the probe test. The ANOVA indicated a main effect of quadrant [ F(2,56) = 9.98, PGG < .0009] suggesting that the pups were treating the quadrants of the tank differently based on training. A
161
main effect of platform condition [ F(1,28) =4.42, P <.0446], but not external cue configuration [ F(1,28 < 1.0], was also revealed. While the External Cue Configuration Platform Condition [ F (1,28) = 1.86, P < .18] and the Quadrant External Cue Platform Condition [ F(2,56) = 1.26, PGG <.29] interactions were not significant, the Quadrant Platform Condition interaction was significant [ F(2,56) = 10.66, PGG < .0006]. Fig. 3 shows and the results confirm that the significant interaction was the result of animals trained in the distal conditions. Simple effects revealed a Quadrant Platform Condition interaction at the double cue condition: F(2,56) = 9.15, PGG < .0014; the Quadrant Platform Condition interaction at the null cue condition approached significance: F(2,56) =2.77, PGG < .0881. Pups trained in the proximal condition spent equivalent amounts of time in all quadrants, and the extramaze cue configuration had little to no effect on the proximal rats’ swim pattern [external cue at proximal: F(2,56) < 1.0]. The spatial discriminability index further characterized the animals’ behavior during the probe test. An overall significant discrimination effect was revealed [ F(1,28) = 15.08, P < .0006] indicating that particular animals were capable of distinguishing the target quadrant from the
Fig. 3. Mean time spent in the platform and opposite quadrants by group during the probe test. Error bars represent standard errors of the means. The mean ( ± S.E.M.) time spent in the quadrants adjacent to the platform is represented by the solid and dashed horizontal lines.
162
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
opposite quadrant. A significant Discrimination Platform Condition interaction [ F(1,28) = 14.86, P < .0006] suggested and individual analyses revealed that the effect was the result of animals from the distal conditions [distal-double cue: F(1,7) = 13.40, P < .0081; distal-null cue: F(1,7) = 9.34, P < .0184]. A similar pattern of results was revealed by the examination of the preference effect (target vs. adjacent quadrants): a significant effect of preference [ F(1,28) = 5.56, P < .0256], and a significant interaction between preference and platform condition [ F(1,28) =9.08, P < .0054]. This result was driven by animals in the distal-double cue condition [ F(1,7) = 8.15, P < .0245]. Although a significant overall evasion effect (adjacent vs. opposite quadrants) was revealed [ F(1,28) = 7.50, P < .0106], evasion did not interact with the remaining factors [external cue configuration: F(1,28) < 1.0; platform condition: F(1,28) =2.40, P < .13; or External Cue Configuration Platform Condition: F(1, 28) < 1.0]. Only animals in the distal-double condition displayed an evasion effect [ F(1,7) = 7.17, P <.0316]. 3.2.2. Platform crossings The mean platform crossing data by group are illustrated in Fig. 4. The ANOVA showed a significant main effect of
platform crossings [ F(2,56) = 6.57, PGG < .0039] but no main effects of cue configuration [ F(1,28) = 1.88, P < .18] or platform condition [ F(1,28) = 0.39, P < .54]. Platform crossings also failed to interact with cue configuration [ F(2,56) = 1.09, PGG < .34], platform condition [ F(2,56) = 1.47, PGG < .2], and Cue Configuration Platform Condition [ F(2,56) = 1.01, PGG < .3]. Nonetheless, specificplanned comparisons revealed a platform crossing effect at the distal [ F(2,56) = 7.11, PGG < .0026] but not at the proximal [ F(2,56) = 0.55, PGG < .56] cue conditions. This result indicates that pups in the distal conditions were learning to use the extramaze cues to navigate the platform location; however, the extramaze cues had no influence on navigation for pups in the proximal conditions. A platform crossing effect was also revealed at the double cue [ F(2,56) = 6.46, PGG < .0042] but not at the null cue [ F(2,56) = 1.19, PGG < .31] condition. Fig. 4 clearly shows that animals in the distal-double condition were more accurate in their search and made more crossings in the platform quadrant relative to the remaining quadrants compared to the other groups, implicating the animals’ ability to use the extramaze cues if the cues were near the location of the platform.
Fig. 4. Mean platform crossings within each quadrant by group during the probe test. Error bars represent standard errors of the means. The solid and dashed horizontal lines represent the mean platform crossings ± S.E.M. in the quadrants adjacent to the platform.
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
The spatial discriminability index indicated an overall significant discrimination [target vs. opposite, F(1,28) =9.57 P < .0045] and preference [target vs. adjacent, F(1,28) = 6.47, P < .0168] effect but not an evasion effect [opposite vs. adjacent, F(1,28) = 1.29 P < .26]. The data in Fig. 4 indicate and individual analyses confirmed that these results are from animals trained in the distal-double cue condition [discrimination: F(1,7) = 10.18, P < .0153; preference: F(1,7) = 8.76, P < .0211]. Cue configuration and platform condition failed to interact with discrimination, preference, or evasion ( P>.05).
4. Discussion The results of the present experiment support our previous work [14] and suggest that cue location and number are important in the preweanling rats’ ability to spatially navigate the location of a platform. This conclusion is derived from clear dissociations as a function of extramaze cue condition in probe test performance, namely the differential periods of time spent in the quadrants, and the respective platform crossing profiles. The more sensitive test measures, time in quadrant and platform crossings, indicated that the rats in the four groups were learning differently within their respective spatial environments. Rats in the proximal-null and proximal-double cue conditions spent equivalent amounts of time in all the quadrants, indicating they failed to discriminate the platform location among the quadrants, did not develop a preference for the target quadrant nor learned to avoid the opposite quadrant. The lack of spatial discriminability suggests that the pups learned little regarding the surrounding environment and its relationship to the platform’s location. This finding contrasts with our previous work [14], which indicated that rats trained in the proximal-double cue condition were using the extramaze distal cues to search for the visible platform on the probe test. However, the proximal-double cue pups’ behavior during acquisition indicated that the extramaze cues were influencing their search for the platform, a finding consistent with our previous work [14]. Moreover, all groups demonstrated improved performance across training trials as indicated by the decrease in escape latencies. Rats trained in the proximal conditions, particularly the double cue condition, demonstrated the fastest rates of escape compared to the distal conditions suggesting that the visibility of the platform played a significant role in localization. One factor that may have contributed to the different results between experiments is procedural in nature and applies to animals in the proximal conditions. In the previous experiment [14], the platform location remained the same during every trial whereas in the current experiment, the platform location changed every other trial. Perhaps maintaining the same location allowed animals to develop turn preferences when learning to locate the platform, giving the appearance that the rats were learning to
163
use the extramaze cues in the proximal-double cue condition. Whereas changing platform location from trial to trial eliminated the possibility of developing any turn preferences and the possibility of using the configuration of extramaze cues, it nonetheless created a ‘‘truer’’ test of platform visibility. Rats in the distal-null and distal-double cue conditions behaved differently than rats in the proximal conditions. These pups spent significantly more time in the target quadrant relative to the opposite quadrant indicating discrimination. Moreover, animals in the distal-double cue group demonstrated more proficient use of the cues by spending more time in the target quadrant relative to the adjacent quadrants indicating a preference. Neither distal condition demonstrated evasion, which characterizes the performance of adult and postweanling rats [14]. Because the distal-double cue pups, but not the distal-null cue pups, developed a preference of location suggests that adept use of the extramaze cues was constrained by the pups’ visual acuity. Nonetheless, the cues were sufficiently visible for the rats to discriminate, and thus learn the spatial location of the platform in relationship to the configuration of multiple cues provided. The platform crossing data, the probe measure for navigational accuracy, revealed more specifically the manner in which animals learned about their spatial environment. Although rats in the distal-null cue condition crossed the original platform location slightly more frequently than the conceptual locations of the adjacent and opposite platforms, they were less accurate and failed to spatially discriminate among these areas. For rats in the proximal-null cue condition, the frequency of crossings was equivalent regardless of location indicating external cues had little to no influence on their navigational behavior. Rats in the proximal-double cue condition made more overall crossings, but surprisingly more crossings were over the original platform location relative to the opposite platform location. This finding is consistent with our previous results [14] and suggests that the proximal-double cue rats may have been using the external spatial cues to aid in the spatial navigation of the visible platform. While rats in the distalnull cue condition made more crossings over the original platform location relative to the opposite quadrant, rats in the distal-double cue condition crossed the platform’s original location significantly more relative to the opposite quadrant location, demonstrating the most proficient use of the extramaze cues. This outcome suggests that the rat’s ability to use the distal cues in the extramaze environment was enhanced by presenting them in the double cue configuration, presumably a more visible arrangement of the cues. It is our view that the different measures, i.e., latency to find the platform, time in quadrant, and number of platform crossings, are tapping into the different learning and memory processes associated with spatial navigation. For example, training trial latencies reflect the animals’ ability to learn the task, i.e., to find and climb onto the platform
164
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
whether it is hidden or visible. However, marked changes in escape latencies are possible under conditions that support no true spatial learning, i.e., when there are no consistent relationships between the escape platform and extramaze cues. This has been demonstrated previously in adult animals (e.g., Ref. [8]) and currently in preweanlings (proximal condition). Nonetheless, the reduction in trial latencies across time appears to be necessary but not sufficient for spatial navigation. The probe trial measures, dwell time (time in quadrant/quadrant preference), and platform crossings allow one to examine the animals’ memory for the platform’s previous location at different levels of spatial accuracy. By examining the relative distribution of swimming time in each of the quadrants, the animal is indicating its preference for a particular area of the pool based on its training history. In addition, the number of platform crossings through the location previously containing the platform indicates how accurately (or not) the animal learned the prior platform location and its relationship with the configuration of extramaze cues. Comparisons made among the target, opposite, and adjacent platform locations provide a statistical index of the animals’ ability to spatially discriminate among pool locations. These comparisons demonstrate what the animal recalled during the probe test. The comparison between the target and opposite platform locations yields a discrimination measure; the comparison between the target and adjacent platform locations yields a preference measure; and the comparison between the opposite and adjacent platform locations yields an evasion measure (the same statistical comparisons were made between the various quadrants with swim time measures). Collectively, these three measures— discrimination, preference, and evasion—indicate the degree of spatial discriminability animals have developed within the testing environment. Typically, animals learn to discriminate as indicated by spending more time or making more platform crossings in the target quadrant relative to the opposite quadrant, i.e., the pups learn where to go in the pool. A preference for the target develops if pups spend the majority of test time in the target quadrant or cross the platform area more, relative to the adjacent quadrants. Pups that learn to avoid the opposite quadrant relative to the adjacent quadrants are demonstrating the evasion effect, i.e., pups learn where not to go in the pool. The acuity of the preweanling’s visual system and the role it plays in spatial navigation is an issue that only recently has been addressed. An implicit assumption of previous ontogenetic water maze studies [4 – 10,12] has been that the animal’s visual system is sufficiently formed to permit navigation, perhaps a result of the adult literature, which has also presumed that the visual platform/proximal cue condition is a sufficient control for potential sensory deficits. However, previously [14] and here, we show that visual acuity in the preweanling rat plays a significant role in its ability to spatially navigate the location of a platform. Thus, the standard use of the visible platform/proximal cue
is not a sufficient control for visual acuity in the preweanling rat. We also know from previous work that visual pattern experience contributes to the young animal’s performance in the Morris water maze [8,12]. The degree of eye opening was assessed in rats 16 days of age and correlated with the respective mean escape latencies on the first day of training (PND 17). A significant inverse relationship was revealed: as the degree of eye opening increased, escape latencies decreased across animals [12]. Though individual visual acuity was not directly assessed in the experiment reported here, it is clear from the manipulation of cues within the extramaze environment that acuity played a role in the animal’s probe trial performance. The experimental manipulation of visual acuity has been previously studied in adult rats in the Morris water maze. Prusky et al. [13] reported that binocularly deprived rats were significantly worse than controls on the visual acuity measure. Importantly, the sight-deprived rats were significantly slower in acquiring the Morris water maze task. Surprisingly, however, the critical probe test measure of quadrant time indicated equivalent performance between groups suggesting that animals in either condition learned the location of the platform equally well despite significant differences in visual acuity. Visual acuity in potentially compromised subpopulations, e.g., aged rats, has also been examined. Because the age-related decline in visual acuity of rats is well documented (e.g., Refs. [19,20]), Lindner and Gribkoff [21] assessed visual acuity in aged rats after spatial training by adding a high contrast, proximal (visual) cue to the hidden platform. The rats were then trained to asymptotic performance. Despite reporting deficits in visual acuity among aged rats, this study [20] failed to find a correlation between visual acuity and spatial learning, which was defined by swimming distance (length of the path from its initial release to the platform). However, Spencer et al. [22], assessed the relationship between retinal degeneration and spatial ability in aged rats more directly. Following an extensive spatial training regiment, rats were classified either as ‘‘learners’’ or ‘‘nonlearners’’ based on performance. Histological and morphometric measures of the retinas were conducted. Among aged rats with the least amount of retinal degradation, 40 of 41 were classified as spatial learners, whereas among aged rats with the most retinal degeneration, 20 of 27 were classified as nonlearners. These results [22] suggest that cognitive performance measures that rely heavily upon visual cues, such as the water maze, may be confounded by visual impairment in aged animals. The results reported here replicate and confirm our previous findings [14] and suggest that the typical use of a visible platform or proximal cues is not sufficient for visual acuity control in the preweanling animal. Although individual visual acuity was not directly assessed, the manipulation of cues within the extramaze test environment shows a clear experimental dissociation indicating that visual acuity played a significant role in the preweanling’s
H.M. Carman et al. / Physiology & Behavior 79 (2003) 157–165
ability to spatially navigate the water maze. The roles of response requirements [8,12] and visual acuity represent critical factors in elucidating the ontogeny of spatial navigation in the rat.
Acknowledgements Heidi M. Carman was supported by NIDA training grant, DA07304. This work was supported by grants from the National Institute of Drug Abuse (DA09160, DA012719, DA12719) and the National Institute of Environmental Health Sciences (ES06259). The authors would like to thank Debra Murray for excellent technical assistance.
References [1] Morris RGM. Spatial localization does not require the presence of local cues. Learn Motiv 1981;12:239 – 60. [2] Morris RGM. An attempt to dissociate ‘‘spatial-mapping’’ and ‘‘working memory’’ theories of hippocampal function. In: Seifert W, editor. Neurobiology of the hippocampus. New York: Academic Press; 1983. p. 405 – 32. [3] Morris RGM, Garrud P, Rawlins JNP, O’Keefe J. Place navigation in rats with hippocampal lesions. Nature 1982;297:681 – 3. [4] Brown RW, Whishaw IQ. Similarities in the development of place and cue navigation by rats in a swimming pool. Dev Psychobiol 2000;37: 238 – 45. [5] Rudy JW, Stadler-Morris S, Albert P. Ontogeny of spatial navigation behaviors in the rat: dissociation of ‘‘proximal-’’ and ‘‘distal-’’ cuebased behaviors. Behav Neurosci 1987;101:62 – 73. [6] Rudy JW, Paylor R. Reducing the temporal demands of the Morris place-learning task fails to ameliorate the place-learning impairment of preweanling rats. Psychobiology 1988;16:152 – 6. [7] Schenk F. Development of place navigation in rats from weaning to puberty. Behav Neural Biol 1985;43:69 – 85.
165
[8] Carman HM, Mactutus CF. Ontogeny of spatial navigation in rats: a role for response requirements? Behav Neurosci 2001;115:870 – 9. [9] Brown RW, Kraemer PJ. Ontogenetic differences in retention of spatial learning tested with the Morris water maze. Dev Psychobiol 1997;30:329 – 41. [10] Kraemer PJ, Randall CK. Spatial learning in preweanling rats trained in a Morris water maze. Psychobiology 1995;23:144 – 52. [11] Roberts WA. Principles of animal cognition. Boston (MA): McGrawHill; 1998. [12] Carman HM, Booze RM, Mactutus CF. Long-term retention of spatial navigation by preweanling rats. Dev Psychobiol 2002;40:68 – 77. [13] Prusky GT, West PWR, Douglas RM. Reduced visual acuity impairs place but not cued learning in the Morris water task. Behav Brain Res 2000;116:135 – 40. [14] Carman HM, Mactutus CF. Proximal versus distal cue utilization in spatial navigation: the role of visual acuity? Neurobiol Learn Mem 2002;78:332 – 46. [15] BMDP Statistical Software. Release 7 [Computer software]. Los Angeles (CA): University of California Press; 1993. [16] Greenhouse SW, Geisser S. On methods in the analysis of profile data. Psychometrika 1959;24:95 – 112. [17] Winer BJ. Statistical principles in experimental design. 2nd ed. New York: McGraw-Hill; 1971. [18] Denenberg VH. Statistics and experimental design for behavioral and biological researchers. New York: Wiley; 1976. [19] O’Steen WK, Sweatt AJ, Brodish A. Effects of acute and chronic stress on the neural retina of young, mid-age, and aged Fischer-344 rats. Brain Res 1987;426:37 – 46. [20] Lindner MD, Plone MA, Schallert T, Emerich WF. Blind rats are not profoundly impaired in the reference memory Morris water maze and cannot be clearly discriminated from rats with cognitive deficits in the cued platform task. Cogn Brain Res 1997;5:329 – 33. [21] Lindner MD, Gribkoff VK. Relationship between performance in the Morris water task, visual acuity and thermoregulatory function in aged F-344 rats. Behav Brain Res 1991;45:45 – 55. [22] Spencer RL, O’Steen WK, McEwen BS. Water maze performance of aged Sprague – Dawley rats in relation to retinal morphologic measures. Behav Brain Res 1995;68:139 – 50.