HIPPOCAMPUS, VOL. 2, NO. 1, PAGES 73-80, JANUARY 1992

Bilateral Knife Cuts to the Perforant Path Disrupt Spatial Learning in the Morris Water Maze Ronald W. Skelton and Robert K. McNamara Department of Psychology, University of Victoria, Victoria, British Columbia, Canada, V8W 3P5

ABSTRACT Both the hippocampus and the entorhinal cortex are known to be crucial for spatial learning, but the contribution of the pathway linking the two structures, the perforant path (PP), has never been tested in a spatial learning paradigm. The present study examined the role of the PP in spatial learning using the Morris water maze. Seven days after bilateral transection of the PP with a fine-bladed knife, rats were habituated to the pool, then trained to swim from varying start locations to a platform submerged in a fixed location. After 28 training trials over 5 days, probe trials (without any platform present) were given to assess spatial memory for the location. Compared to sham-operated controls, lesioned rats showed slower learning and poorer asymptotic performance in terms of both swim path distance and escape latency, and less preference for the correct quadrant during probe trials. When the platform location was “reversed” to the opposite quadrant, the lesioned rats again showed poorer learning, poorer asymptotic performance, and reduced preference for the correct quadrant on the probe trial. When tested with a visible platform whose position varied from trial to trial, lesioned rats performed as well as controls. These results are congruent with previous analyses of the contributions of the entorhinal cortex and hippocampus to spatial learning and suggest that for spatial learning, the PP is a critical functional link between these two structures. Key words: hippocampus, memory, rats, lesions, spatial memory

Numerous studies have shown that lesions of the hippocampal formation produce severe and lasting deficits in spatial learning and memory (for reviews see O’Keefe and Nadel, 1978; Sutherland and Rudy, 1989). A number of studies have shown that lesions of the entorhinal cortex, a primary source of cortical afferent fibers, also produce deficits in spatial learning (Thompson, 1976; Olton et al., 1978; Jarrard et al., 1984; Ramirez and Stein, 1984: Reeves and Smith, 1987: Ramirez et al., 1988: Rasmussen et al., I989), though there have been exceptions (Bouffard and Jarrard, 1988). Perhaps the clearest and most detailed analysis of the spatial memory deficits produced by lesions of entorhinal cortex was made by Schenk and Morris (1985), who studied the effects of retrohippocampal lesions on acquisition and retention of spatial navigation in the Morris water maze. They were able to show that the performance improvements displayed by rats with lesions of the entorhinal cortex (with or without damage to the subiculum) were actually due to acquisition of relatively efficient procedures for finding the platform each trial, rather than to any residual place learning ability. Swim patterns during probe trials, in which no platform was present in the pool, and a positional discrimination task clearly showed that the Correspondence and reprint requests to R. W. Skelton, Department of Psychology, Box 3050, Victoria, B.C., Canada, V8W 3P5

lesioned rats were unable to learn or remember spatial locations. In this regard, deficits following lesions of the retrohippocampal area were essentially the same as those produced by lesions of the hippocampus itself (Morris et al., 1982; Sutherland et al., 1982). Surprisingly, very few studies have measured the effects of lesions of the pathway connecting the entorhinal cortex to the hippocampus, the perforant path (PP). The PP is the main cortical afferent pathway to the hippocampus and projects from the entorhinal cortex to the dentate gyrus (DG), through the angular bundle just ventral to the splenium of the corpus callosum, near the dorsal hippocampal commissure (HjorthSimonsen and Jeune, 1972). The few studies that have assessed the behavioral effects of PP lesions examined only the effects on avoidance learning (Myhrer, 1975), exploratory behavior, and reactions to novelty (Myhrer, 1988a) and acquisition or retention of a brightness-oddity discrimination (Myhrer, 1988b). No systematic investigation of PP lesions of spatial navigation has yet been made. The PP is probably the best (or at least the most extensively studied) anatomical structure for determining the relation between changes in synaptic efficacy and learning. The PP-DG synapses are well known for their ability to undergo longterm potentiation (LTP; see Teyler and DiScenna, 1987, for a review), and a number of studies have examined the relation

73

74 HIPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992 between LTP-like changes in behavior and learning (Ruthrich et al., 1982; Berger, 1984; Weisz et al., 1984; Sharp et al., 1985; 1989; Skelton et al., 1985; 1987). There even have been studies that have purportedly produced a functional lesion of the PP by exhausting the synaptic plasticity in the PP-DG through high-frequency stimulation to the point at which the ability to undergo LTP (or long-term enhancement; LTE) is saturated (McNaughton et al., 1986; Castro et al., 1989). However, the interpretation that the behavioral deficits result from PP dysfunction due to depletion of synaptic plasticity in PP-DG synapses could be questioned on the grounds that functional deficits in spatial tasks following PP lesions never have been directly demonstrated. The present study examined the role of the PP on learning and memory by testing the effects of bilateral PP lesions on spatial learning in the Morris water maze. The Morris water maze was chosen for its sensitivity to lesions of the hippocampus and entorhinal cortex and for its exceptional ability to reveal and dissociate deficits in spatial learning from deficits in the procedural components of the task (e.g., Schenk and Morris, 1985). Knife cuts, rather than electrolytic lesions, were used to transect the PP to minimize electrophysiological disturbances in the hippocampus. In most respects, the behavioral methods used here replicated those of Schenk and Morris (1985) except that the present study examined only acquisition, not retention, of preoperatively learned, submerged platform locations. In addition, the effects of PP lesions were tested on acquisition of a second platform position, after acquisition to one platform position was complete, to help determine whether performance improvements were due to acquisition of procedural as opposed to spatial mapping strategies. Finally, acquisition of escape to a visible platform was tested to determine whether the cuts produced deficits in sensory, motor, or motivational processes, or in the formation of simple associations.

MATERIALS A N D METHODS Subjects Eighteen male Long-Evans rats (Charles River, Canada), weighing 320-720 g, were housed individually in shoe box cages with food and water ad libitum. All tests were conducted during the light portion of the 12: 12-hour light-dark cycle.

Surgery The rats were deprived of food overnight prior to surgery, then anesthetized with Nembutal (60 mg/kg) and placed in a Kopf stereotaxic frame with skull level. A wide slot was drilled in the skull 1 .O-2.4 mm anterior to the intraaural line, extending bilaterally from 0.5 mm to 5.0 mm lateral to the midline. The bone over the sagittal sinus was partially removed as well. The PP was cut using a disposable arteriotomy knife (Moira, from Fine Science Tools), which had a triangular blade 7 mm long, 2 mm wide at the base, and 0.2 mm thick, with an edge angled 15” to the handle. The knife was mounted on a manipulator angled 20” from vertical in the coronal plane, so that the knife edge was angled a total of 35”

n’

Kn’fe Cur

Fig. 1 . The intended path and targets of the bilateral knife cuts, depicted on a coronal section 1.7 mm anterior to the intraaural line. Fiber bundles within the knife cut area are shown as dashed rather than solid lines; cell areas are shown as dotted rather than dashed lines. Redrawn from Paxinos, G . , and C. Watson (1988) The Rat Brain in Stereotaxic Coordinates. Academic Press, Orlando, FL. With permission.

from vertical. The knife was positioned 1.7 mm anterior to the intraaural line, 4.0 mm lateral to the midline on the left side, and then inserted 5.0 mm (as measured by the “vertical” scale of the manipulator), penetrating at a 20” angle. The blade was then moved 2.5 mm medially (as measured by the “horizontal” scale of the manipulatorj, then withdrawn using the “vertical” adjustment of the manipulator. The procedure was repeated on the right side. Figure 1 shows the intended paths of the knife cuts. After both cuts were complete, the slot in the skull was sealed with sterile bone wax and the wound was sutured. Sham-operated rats were treated identically, except that the knife was not inserted. All rats were returned to a recovery cage with a heat lamp overnight. The rats were given at least 7 days to recover from surgery prior to any testing.

Apparatus The Morris water maze consisted of a circular pool (150 cm diameter, 45 cm high) with a featureless white inner surface. The pool was filled to a depth of 25 cm with 26°C ( k 1°C) water, rendered opaque by the addition of 1,500 mL skimmilk powder. The hidden escape platform was a clear Plexiglas stand with a 13 X 13-cm top submerged 2.5 cm below the surface of the water so as to be invisible at water level. The visible escape platform was black, and its top (13 x 13 cm) protruded 5 cm above the surface. The maze was housed in a colony-type room that had a light directly over the pool, a variety of visual cues on the surrounding walls, and a radio providing background noise.

Procedures After recovery from surgery, all rats were habituated to the pool using a single 60-second pretraining probe trial. In this and all subsequent probe trials, the rat was placed into the pool facing the wall at the “north” pole and allowed to swim

PERFORANT PATH CUTS AND SPATIAL LEARNING / Skelton and McNarnara

for 60 seconds without any platform in the pool. Testing began on the following day, and consisted of three phases: acquisition, reversal, and cue. The acquisition phase consisted of 5 days of training with the submerged platform maintained in the center of the northwest (NW) quadrant with 4, 4, 8, 8, and 4 trials on days 1-5, respectively, as per Schenk and Morris (1985). A posttraining probe trial was given after the final trial of the fifth day. The reversal phase began the next day and repeated these procedures exactly, except that the platform was “reversed” to the center of the diagonally opposite (i.e., southwest; SW) quadrant. It concluded with a postreversal probe trial. The cue phase was the following day and consisted of eight trials with the visible platform located in the center of a different quadrant on each trial; the location varied in pseudorandom order (e.g., NW, SE, SW, NE). Each trial was begun by placing the rat gently into the water facing the pool wall at one of four randomly determined starting positions (north, east, south, or west pole). During each trial, one experimenter (the handler) would place the rat into the pool and record its escape latency, while a second experimenter (the recorder) drew the path taken by the rat on a map of the pool. The handler would remain near the (varying) start location, and the recorder would stand on the opposite side of the pool (180” away); both were visible, stationary, and silent throughout each trial. (The varying of the positions of both experimenters was designed to prevent the rats from using them as navigational cues). If the rat did not locate the platform within 60 seconds, it was gently guided to the platform and allowed to climb on. Upon reaching the platform, the rat was left there for 20 seconds and then removed to a holding cage located 90 cm below a 250 W brooding lamp for a 5-minute intertrial interval. Core body temperature was recorded before the first and after the last trial of each day to ensure hypothermia did not occur. Swim path distances were measured using a map-reading device (curvimeter or planimeter; Morris, 1981). Probe trials were scored for total distance in each quadrant and for the number of times the rat crossed a platform-sized square area (13 x 13 cm) in the center of each quadrant.

Histology After the conclusion of all testing, the rats were sacrificed with an overdose of sodium pentobarbital and perfused with saline and formalin. The brains were frozen and sectioned horizontally, taking serial 80-ym sections through the entire hippocampus. Knife-cut damage to cell and fiber areas of each rat was documented by matching Thionine-stained serial sections to 1 of 12 horizontal plates from Paxinos and Watson (1988) and drawing the direct damage on a reproduction of the plate. Cellular degeneration in pyramidal cell layers I1 and 111 of the entorhinal cortex and the granule and pyramidal cell fields of the hippocampus was evaluated by visually comparing the number and density of cells in sections at 0.5-mm intervals through the entire dorsal-ventral extent of the entorhinal cortex of each lesioned rat with equivalent sections from control rats, and rating cell losses in both the medial and lateral entorhinal cortices as minimal (or nonexistent), moderate, or severe (or complete).

75

RESULTS

Histology The knife cuts transected the posterior neocortex at the anteroposterior (AP) level of the subiculum, separating the entorhinal cortex and posterior perirhinal cortex from the more rostra1 parts of the brain (see Fig. 2 ) , directly damaging the PP as well as the subicular areas and overlying cortices (occipital, perirhinal, and retrosplenial). In most rats (7 of 9), the knife cuts passed through the angular bundle and transected the corpus callosum anterior to the forceps major in both hemispheres. In one rat the angular bundle was damaged only unilaterally, and in the other it was almost completely spared. In the latter rat, the knife passed through only the most caudal portion of the forceps major and into the medial entorhinal cortex. On behavioral measures, this rat was as proficient as the best control rat in all stages of the experiment, and under histologic examination, cell loss in the superficial layers of the entorhinal cortex of this rat was confined to a small caudal “shadow” of the knife cut in the medial entorhinal cortex. The other eight rats had substantial cell loss in the superficial layers of both the medial and lateral entorhinal cortices, through most of the dorso-ventral extent,

I

- 3 1 mm

-5. I

Fig. 2. Reconstruction of knife cuts depicted on partial horizontal sections at 3.1, 4.1, and 5.1 mm ventral to Bregma. The heavy border shows the range of damage from all knife cuts, while the thin lines within these borders show location of knife cuts as estimated from maximal medial-lateral extent of damage in each rat. (All nine bilateral knife cuts are shown on all three horizontal sections.) Redrawn from Paxinos, G . , and C. Watson (1988) The Rat Brain in Stereofaxic Coordinates. Academic Press, Orlando, FL. With permission.

76

HIPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992

bilaterally. In contrast, there was no visible damage or cell A loss in the dentate gyrus or pyramidal cell fields of the hippocampus. Sham-operated (control) rats sustained only minor damage to the overlying cortices.

Behavior Rats with PP cuts were severely impaired in their ability to learn the location of the submerged platform in both its original and reversed locations, but were unimpaired at escaping to the visible platform. Figure 3A shows mean distances (lengths of swim paths) required to find the submerged platform during acquisition, reversal, and cue training (visible platform). Lesioned rats showed slower learning and worse asymptotic performance than sham-operated rats, taking paths that were on average twice as long at the end (last block) of both the acquisition phase (507 t 151 cm and 246 & 73 cm, mean +- SEM for lesioned and control groups, respectively) and the reversal phase (400 ? 109 cm and 201 ? 39 cm). Both groups showed longer swim distances at the start of the reversal phase, and on the first block, had almost identical values (Fig. 3A). However, for the lesioned rats, this represented only a minor increase in swim distance (34%), whereas for the controls, this was a substantial change (161%), suggesting that the change in platform position was more apparent to the controls than to the lesioned rats. Repeated measures multivariate analysis of variance (MANOVA) on block average data revealed significant differences between groups during both acquisition (F(1,16) = 4.76, P < .04) and reversal (F(1,16) = 5.68, P < .03). Changes over blocks were significant (P < .05) during both phases, but the interaction between group and blocks was not (P> .lo). Both groups quickly learned to swim directly to the visible platform, and swim distances were virtually indistinguishable (Fig. 3A). Lesioned rats also took more time to find the submerged platform (Fig. 3B). Learning was slower, and asymptotic performance was worse. Latencies at the end (last block) of the 5.4 seconds and 8.5 ? 2.1 acquisition phase were 18.3 seconds for lesioned and control groups, respectively; latencies at the end of the reversal phase were 12.5 ? 3.4 seconds and 6.4 ? 1.2 seconds, respectively. Repeated measures MANOVA on the latencies revealed significant differences between groups during both acquisition (F(1,16) = 6.57, P < .02) and reversal (F(l,16) = 5.78, P < .03). As with swim distance, changes in latencies over blocks were significant ( P < .05), but the group by blocked interaction was not significant (P> .lo). Like distances, latencies also increased from the end of the acquisition phase to the start of reversal, and the increase was small in the lesioned group (31%) and large in the control group (169%). Latencies of the two groups in the cue condition when the platform was visible were indistinguishable (Fig. 3B). Swim speed during trials, as assessed by dividing swim distance by escape latency, was unaffected by either training or the PP cut (Fig. 3C). The lesioned group swam slightly slower on average, but there were no significant differences between groups (P> .lo). There were also no major changes or differences in core body temperatures. Mean body temperatures never dropped more than 2°C from normothermia from preswim to postswim measures. Although the body tem-

*

-5

Vis. PI.1.

Reversal

Acauirilion

1500

1200

900

0

0

c

2 .-

600

v)

n 300

s,

-

0

Vis.

50

Reversal

Acquisition

I

40

30

20

10

0 40

Acquisicion

1

Vir. Pllf.

RovertaI

I

@ 20

-

lo

-

0

. 0

.

0

,

CUT SHAM 1

,

1

,

,

1

,

1 2 3 4 5 6 7 8

1

1

1

1

1

1

-

910111213141516

Trial Blocks Fig. 3. Changes in (A) swim distance, (B) escape latency, and (C) swim speed over blocks of four trials during acquisition, reversal, and cue (visible platform: vis. plat.) phases are shown as mean and SEM for rats with PP knife cuts (CUT) and sham operations (SHAM).

peratures of the lesioned rats decreased somewhat more than those of controls, as would be expected from their longer immersions (swims), average temperatures were always within 1°C of the controls. This pattern of results-impaired performance to the submerged platform with normal swim speeds and normal performance to the visible platform-indicates that the cut to the PP did not alter sensory, motor, or motivational processes, but rather specifically disrupted the rats’ ability to use spatial information.

50

-

40

-

30

-

20 10

0

SHAM

CUT

1m m l ......................

(mean = 44% +- 2%; t ( 8 ) = 10.5, P < .001), whereas lesioned rats showed no such preference (mean = 27% 2 3%; t(8) = .62, NS). After reversal training, control rats again showed a clear preference for the now correct quadrant (mean = 42% ? 2%; t(8) = 7.7, P < .001) and avoided the previously correct quadrant (mean = 13% k 2%; t(8) = -7.55, P < .002). In contrast, lesioned rats showed no significant preference for the correct quadrant after reversal training (mean = 32% 3%; t ( 8 ) = 2.25, NS) but did avoid the previously correct quadrant (mean = 20% +- 2%; t(8) = -2.74, P < .05). Further analysis of swimming strategies during the probe

*

50 40

30 20 10 0

7

Post-Reversal 50

1

+

Correct

+

40

30 20

10

Post-Reversal

0 SW NW NE SE

Quadrant

t

SW NW NE SE

6 1

Correct

Quadrant

Previous

Fig. 4. Quadrant preferences during probe trials, expressed

as percentage of total swim distance. The pretraining probe was the habituation trial; the posttraining probe was the last trial on day 5 of acquisition. The postreversal probe was the last trial of day 5 of reversal testing. The northwest quadrant was correct (contained the platform) during acquisition; the southeast quadrant was correct after reversal. *Significantly ( P < .01) above chance (dotted line, 25%). Data from the probe trials revealed that the spatial search strategies of the two groups were clearly different. Figure 4 shows the swim distances (mean 2 SEM) within each quadrant during all three probe trials, expressed as a Percentage of total distance Swum during the trial. During the pretraining probe trial (habituation), neither group preferred any quadrant or half of the pool. During the posttraining probe trial, control rats spent a greater percentage of their swimming in

Y)

m

El Correct

4 -

.-C Y)

cn

e

0

2 -

0 -

SW NW NE SE

Quadrant

SW NW NE SE

Quadrant

Fig. 5, Quadrant preferences during probe trials, as revealed by the number of times the rat crossed Over a ‘‘phantom” platform location in the center of each quadrant. *Significantly (P < .02) greater than second-most-often crossed platform location.

78

,

HIPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992

posttraining and postreversal probes (mean f SEM). During habituation, there were very few platform crossings in any quadrant by either group. By the end of the acquisition phase, both groups were crossing the platform positions in all quadrants more often ( P < .05),but only the control group crossed the correct position significantly more often than any other platform position (t(8) = 4.60, P < .002 against the secondmost often crossed position). The lesioned group did not cross the correct platform significantly more often than the other positions (t(8) = 1.97, NS against the second-most often crossed position), and crossed the correct platform less often than sham-operated rats, though the contrast was not quite significant (t(8) = 2.05, P < ,057). A similar pattern of results was obtained during the postreversal probe, with the addition that control rats crossed the correct platform position significantly more often than lesioned rats did ( t ( 8 ) = 3.21, P < .005). These results show that both groups learned to search the circular region of the pool where the platform might be, but that the lesioned rats were unable to determine which quadrant of the pool would be most likely to contain the platform. This, in turn, suggests that the relatively efficient swim paths of the lesioned rats during submerged platform testing were due to the adoption of a cue- or responsebased strategy, such as searching the pool at a given distance from the wall. Indeed, during probe trials, lesioned rats tended to swim in large loops that rarely crossed the center of the pool, whereas control rats tended to swim to the correct quadrant, search it, then swim over all other regions of the pool including the center.

DISCUSSION In this experiment, rats with bilateral knife cut transections of the PP showed an impaired ability to acquire the location of a submerged platform. The impairment was evident in the longer swim paths and longer escape latencies during both original acquisition and reversal to a new platform location. The impairment was evident also in the reduced specificity of the search strategies during the probe trials. These deficits did not appear to be due to impairments of sensory, motor, or motivational systems, because the lesioned rats showed normal swim speeds and were unimpaired in finding a visible platform. Rather, the deficit appeared to be due to an impairment in spatial learning that prevented the rats from navigating to the platform location using extramaze cues, forcing them to rely on less efficient strategies based on local cues (e.g., the wall of the pool), particular patterns of swimming (e.g., in large loops), or both (e.g., swim in loops at approximately the right distance from the wall). It could be argued that the slower acquisition and poorer asymptotic performance were not due to the use of a less efficient strategy, but rather were due to a lesion-induced slowing of a memory forming process and an insufficiently long test duration. However, the pattern of results argues against this and suggests a true and near-complete deficit in spatial localization. Both groups showed the most rapid decreases in swim distances during the first two to three blocks of acquisition, and then showed only minor improvements thereafter. Yet, on the postacquisition probe trial, lesioned rats showed no preference for the previously correct quadrant. Lesioned rats did cross platform positions more often

than they had before training, but did not discriminate the correct position from the incorrect ones. Further, lesioned rats concentrated their search in the radial area of the pool where the platform might be found (without regard for quadrant), unlike control rats who searched first in the correct quadrant, and then over the entire surface of the pool. These results suggest that the performance improvements of the lesioned rats was due to acquisition of an improved strategy for searching the pool for the platform, rather than acquisition of the spatial location of the platform. The results from the reversal phase were consistent with this interpretation. At the start of the reversal phase, control rats showed substantial increases in both distance and latency, whereas lesioned rats showed very little change, indicating that the control rats had to relearn the platform location but the lesioned rats had only to make a minor change in their search procedure. On the first blocks of both the acquisition and reversal phases, when neither group could "know" the location of the platform, both groups showed virtually identical swim distances and latencies. During the reversal phase, normal rats reached asymptotic performance after only one block of trials, whereas lesioned rats showed gradual improvements over the entire seven blocks. During the postreversal probe trial, lesioned rats failed to show a consistent preference for the correct quadrant and crossed the correct platform location only slightly more often than noncorrect positions. When tested with the visible platform, the lesioned rats were unimpaired, swimming directly to the platform from the first trial onward, just like the controls. Taken together, the deficits in performance with the submerged platform, the nonspecific searching during probe trials, and the unimpaired performance to the visible platform all indicate that the lesioned rats were able to learn to escape onto a platform using local cues (e.g., the wall of the pool), or a procedure (e.g., swim around in large loops), but were unable to learn the spatial location of the submerged platform relative to distal extramaze cues. There are several reasons why it is reasonable to attribute the deficits in the lesioned rats to damage to the PP. First, the structures whose damage appeared to be most disruptive to spatial learning were the posterior portions of the corpus callosum (specifically the forceps majora) and the angular bundle, where the PP has been described to pass (Blackstad, 1956; Hjorth-Simonsen and Jeune, 1972). Extensive cell loss was clearly observable in the superficial layers of both enthorhinal cortices, indicating retrograde degeneration of the cells of origin of the PP (Steward and Scoville, 1976; Schwartz and Coleman, 1981). Only one rat had knife cuts completely posterior to the angular bundle; this rat had only minimal cell loss in the entorhinal cortex and performed as well as controls on all behavioral measures. Second, the deficits were probably not due to damage to the presubiculum and parasubiculum, as these structures were severely damaged in only three of the nine lesioned rats. Third, while it is possible that the deficit could have been due to transection of hippocampal efferents to the entorhinal cortex (Hjorth-Simonsen, 1971), entorhino-entorhinal commissural connections (Blackstad, 1956), or even crossed temporo-ammonic pathways (Goldwitz et al., 1975; Zimmer and Hjorth-Simonsen, 1975), these pathways are much smaller than the massively projecting PP.

PERFORANT PATH CUTS AND SPATIAL LEARNING / Skelton and McNarnara It is more reasonable to attribute the deficit to the large pathway than the smaller ones. Fourth, while it is possible that the observed deficits were due to alterations in intrinsic processing in either the hippocampus or entorhinal cortex, resulting from orthograde or retrograde degeneration and concomitant anatomical reorganization, such an interpretation assumes that the hippocampus or entorhinal cortex is capable of independently mediating spatial learning, and only fails to do so because of anatomical reorganization. Since lesions of either hippocampus (Morris et al., 1982; Sutherland et al., 1982) or entorhinal cortex (Schenk and Morris, 1985) alone produces the same spatial learning deficit, it is more reasonable to assume that these two structures function interdependently and to conclude that the observed deficits were due to severing the connection between them. The behavioral effects of PP lesions observed in the present study are entirely congruent with the large body of evidence documenting the critical role played by the hippocampal formation in spatial learning in the water maze (see Morris et al., 1982; Sutherland et al., 1982). More specifically, they agree very well with the deficits observed after lesions of the retrohippocampal formation by Schenk and Morris (1983, who were able to dissociate the place learning shown by control rats from procedural learning shown by rats with lesions in the entorhinal cortex, or entorhinal cortex and subiculum. They too found that the place learning deficit produced by retrohippocampal lesions was almost identical to the place learning deficit produced by lesions of the hippocampal formation (Morris et al., 1982; Sutherland et al., 1982). The reasonable conclusion to draw from these findings is that the PP is a critical link between the entorhinal cortex and the hippocampus for spatial learning in the water maze. In a wider sense, the present results confirm the importance of the integrity of the hippocampal region (hippocampal formation, subicular regions, and entorhinal cortices) to spatial learning in general. Although the effect of PP lesions on spatial learning has not been studied previously, there is considerable evidence that the entorhinal cortex is important for spatial learning (Myhrer, 1975; Thompson, 1976; Olton et al., 1978; Jarrard et al., 1984; Ramirez and Stein, 1984; Reeves and Smith, 1987; Ramirez et al., 1988; Goodlett et al., 1989; Rasmussen et al., 1989). However, it is not clear whether the entorhinal cortex is required for postlesion retention and performance of preoperatively acquired spatial locations, as lesions of the entorhinal cortex disrupt retention in the Morris water maze (Schenk and Morris, 1985) but not radial arm maze (Bouffard and Jarrard, 1988) or brightness-oddity discriminations (Myhrer and Naevdal, 1989). The present study confirms results from hippocampal and entorhinal cortex lesions, filling a rather important gap in our knowledge of the functional interactions of the components of the hippocampal region. The importance derives not so much from what it reveals about the functions of the hippocampus, entorhinal cortex, and PP, but rather from its substantiation of a common assumption that the PP plays an important role in spatial learning, an assumption made in several previous investigations of the role of PP-DG LTP in learning. Previous studies have shown that blockade of LTP in the PPDG by~high-frequency stimulation to the point of “saturation” of synaptic plasticity disrupts subsequent spatial learn-

79

ing in the escape-hole board task (McNaughton et al., 1986) and the Morris water maze (Castro et al., 1989), but neither task had been shown to depend upon the PP. Perhaps future analyses of electrophysiological events within the synapses of the PP onto hippocampal cells may reveal that LTP is occurring in the synapses as a consequence of learning in the Morris water maze, and may even establish a causal link between LTP-like electrophysiological changes and behavioral learning. Certainly, a precondition for establishing such a link is evidence showing that the structure being studied actually participates in the learning taking place. The present study establishes just such a precondition for the PP and the Morris water maze.

ACKNOWLEDGMENTS The authors thank Trent Fisher and Astrid Duren for their expert technical assistance. They also thank Robert J. Sutherland for his advice on the design of this experiment and comments on the manuscript. This work was supported by an operating grant from the Natural Sciences and Engineering Research Council of Canada and a research grant from the University of Victoria.

References Berger, T. W. (1984) Long-term potentiation of hippocampal synaptic transmission affects rates of behavioral learning. Science 224:627630. Blackstad, T. W. (1956) Commissural connections of the hippocampal region in the rat with special reference to their mode of termination. J. Comp. Neurol. 105:417-537. Bouffard, J. P., and L. E. Jarrard (1988) Acquisition of a complex place task in rats with selective ibotenate lesions of hippocampal formation: Combined lesions of subiculum and entorhinal cortex versus hippocampus. Behav. Neurosci. 1022328-834. Castro, C. A , , L. H. Silbert, B. L. McNaughton, and C. A. Barnes (1989) Recovery of spatial learning deficits after decay of electrically induced synaptic enhancement in the hippocampus. Nature. 342:545-548. Goldwitz, D. A., W. F. White, 0. Steward, G. Lynch, and C. Cotman (1975) Anatomical evidence for a projection from the entorhinal cortex to the contralateral dentate gyrus of the rat. Exp. Neurol. 47:433-441. Goodlett, C. R., J. M. Nichols, R. W. Halloran, and J. R. West (1989) Long-term deficits in water maze spatial conditional alteration performance following retrohippocampal lesions in rats. Behav. Brain Res. 32:63-67. Green, E. J., and W. T. Greenough (1986) Altered synaptic transmission in dentate gyrus of rats reared in complex environments: Evidence from hippocampal slices maintained in vitro. J. Neurophysiol. 55939-750. Hjorth-Simonsen, A. (1971) Hippocampal efferents to the ipsilateral entorhinal area: An experimental study in the rat. J. Comp. Neurol. l42:4 17-438. Hjorth-Simonsen, A., and B. Jeune (1972) Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. Comp. Neurol. 144:215-231. Jarrard, L. E., H. Okaichi, and 0. Steward (1984) On the role of hippocampal connections in the performance of place and cue tasks: Comparisons with damage to hippocampus. Behav. Neurosci. 98:946-954. McNaughton, B. L., C. A. Barnes, G. Rao, 1.Baldwin, and M. Rasmussen (1986) Lone-term enhancement of himocamual svnautic

-

1.

. , .

80

HZPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992

transmission and the acquisition of spatial information. J. Neurosci. 6:563-571. Morris, R. G. M. (1981) Spatial localization does not require the presence of local cues. Learning and Memory 12:239-260. Moms, R. G. M., P. Garrud, J . N. P. Rawlins, and J. O’Keefe (1982) Place navigation impaired in rats with hippocampal lesions. Nature 297:681-683. Myhrer, T. (1975) Locomotor and avoidance behavior in rats with partial or total hippocampal perforant path sections. Physiol. Behav. 15:217-224. Myhrer, T. (1988a) Exploratory behavior and reaction to novelty in rats with hippocampal perforant path systems disrupted. Behav. Neurosci. 102:356-362. Myhrer, T. (1988b) The role of medial and lateral hippocampal perforant path lesions and object distinctiveness in rats’ reaction to novelty. Physiol. Behav. 42:371-377. Myhrer, T., and G. A. Naevdal (1989) The temporal-hippocampal region and retention: The role of temporo-entorhinal connections in rats. Scand. J. Psychol. 30:72-80. O’Keefe, J., and L. Nadel (1978) The Hippocampus as a Cognitive M a p . Oxford University Press, Oxford. Olton, D. S., J. A. Walker, and F. H. Gage (1978) Hippocampal connections and spatial discrimination. Brain Res. 139:295-308. Paxinos, G., and C. Watson (1988) The Rat Brain in Stereotaxic Coordinates. Academic Press, Orlando, FL. Ramirez, J. J., and D. G. Stein (1984) Sparing and recovery of spatial alternation after entorhinal cortex lesions in rats. Behav. Brain Res. 1333-61. Ramirez, J. J., R. Labbe, and D. G. Stein (1988) Recovery from perseverative behavior after entorhinal cortex lesions in rats. Brain Res. 459:153-156. Rasmussen, M., C. A. Barnes, and B. L. McNaughton (1989) A systematic test of cognitive mapping, working memory, and temporal discontiguity theories of hippocampal function. Psychobiol. 17~335-348. Reeves, T. M., and D. C. Smith (1987) Reinnervation of the dentate gyrus and recovery of alternation behavior following entorhinal cortex lesions. Behav. Neurosci. 101:179-186. Ruthrich, H., H. Matthies, and T. Ott (1982) Long-term changes in synaptic excitability of hippocampus cell populations as a result of training. In Neuronal Plasticity and Memory Formation, C. A.

Marsan and H. Matthies, eds., pp. 589-594, Raven Press, New York, NY. Schenk, F., and R. G. M. Morris (1985) Dissociation between components of spatial memory in rats after recovery from the effects of retrohippocampal lesions. Exp. Brain Res. 58:ll-28. Schwartz, S . P., and P. D. Coleman (1981) Neurons of origin of the perforant path. Exp. Neurol. 74:305-312. Sharp, P. E., B. L. McNaughton, and C. A. Barnes (1985) Enhancement of hippocampal field potentials in rats exposed to a novel, complex environment. Brain Res. 339:361-365. Sharp, P. E., B . L. McNaughton, and C. A. Barnes (1989) Exploration-dependent modulation of evoked responses in fascia dentata: Fundamental observations and time course. Psychobiol. 17:257269. Skelton, R. W., J. J. Miller, and A. G. Phillips (1985) Long-term potentiation facilitates behavioral responding to single-pulse stimulation of the perforant path. Behav. Neurosci. 99:603-620. Skelton, R. W., A. S. Scarth, D. M. Wilkie, J. J . Miller, and A. G. Phillips (1987) Long-term increases in dentate granule cell responsivity accompany operant conditioning. J. Neurosci. 7:3081-3087. Steward, O., and S. A. Scoville (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol. 169:347-370. Sutherland, R. J., and J. W. Rudy (1989) Configural association theory: The role of the hippocampal formation in learning, memory, and amnesia. Psychobiol. 17:129-144. Sutherland, R. J . , B. Kolb. and 1. Q. Wishaw (1982) Spatial mapping: Definitive disruption by hippocampal or medial frontal cortical damage in the rat. Neurosci. Lett. 31:271-276. Teyler, T. J., and P. DiScenna (1987) Long-term potentiation. Annu. Rev. Neurosci. 10:131-161. Thompson, R. (1976) Entorhinal-subicular lesions: Amnestic effects on an assortment of learned responses in the white rat. Bull. Psych. SOC.8:433-434. Weisz, D. G., G . A. Clark, and R. F. Thompson (1984) Increased responsivity of dentate granule cells during nictitating membrane response conditioning in rabbit. Behav. Brain Res. 12:145-154. Zimmer, J . , and A. Hjorth-Simonsen (1975) Crossed pathways from the entorhinal area to the fascia dentata: 11. Provokable in the rat. J . Comp. Neurol. 161:71-102.