Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Behavioral Neuroscience Copyright 2006 by the American Psychological Association

2006, Vol. 120, No. 3, 641– 650 0735-7044/06/$12.00 DOI: 10.1037/0735-7044.120.3.641

Hippocampal and Caudate Metabolic Activity Associated With Different


Navigational Strategies

Rubén Miranda, Eduardo Blanco, and Sandra Rubio


Azucena Begega Madrid Autonomous University
University of Oviedo

Jorge L. Arias
University of Oviedo

Hippocampal and striatal systems are widely related to spatial tasks. Depending on the strategies used,
different memory systems can be activated. In this study, the authors used the cytochrome c-oxidase
technique as a functional marker of the hippocampal and dorsal striatum activity related to training in
several water maze tasks. Current results show a differential participation of the hippocampal and striatal
systems in navigation. When spatial information is relevant, participation of the hippocampal system is
more important, and when the task is similar to a response learning one, the striatal system is more active.
According to computational models, CA3 seems to be more active when the associative demand is
higher, whereas CA1 and dentate gyrus activity are higher when spatial information processing is
required.

Keywords: spatial learning, Morris water maze, hippocampus, striatum, cytochrome c-oxidase

The hippocampus is a structure involved in memory processes guided system (Compton, 2004; McDonald & White, 1994; Okai-
and learning tasks, such as the spatial ones, that require an elab- chi, 2001; Schroeder et al., 2002). Nevertheless, as Hamilton,
orated processing of the incoming information (Eichenbaum, Rosenfelt, and Whishaw (2004) noted, some neurobiological ma-
Stewart, & Morris, 1990). In contrast, the striatum appears to be nipulations may induce a bias in strategy preference leading to a
specialized in response learning, which is a slower form of learn- misinterpretation of the conclusions relating to the interaction
ing based on stimulus–response associations (White & McDonald, between brain function and behavioral outcomes.
2002). Hippocampal or fimbria-fornix lesioned rats can solve spatial
Previous studies have shown that hippocampal and striatal sys- tasks using nonspatial strategies (e.g., taxon strategies and dead
tems can process information in parallel during a spatial task. Most reckoning; Pouzet, Zhang, Feldon, & Rawlins, 2002) apparently
of the evidence suggesting a functional dissociation between hip- on the basis of an extrahippocampal navigation system (Alyan,
pocampal and striatum systems comes from studies involving Jander, & Best, 2000; Whishaw, Hines, & Wallace, 2001). This
lesions or pharmacological manipulations (Compton, 2004; suggests that involvement of the hippocampus in spatial learning
Eichenbaum et al., 1990; McDonald & White, 1994; Morris, depends on the task and the strategies used. In this way, when a
Garrud, Rawlings, & O’Keefe, 1982; Okaichi, 2001; Packard, task can be solved by nonspatial strategies, the performance may
Cahill, & McGaugh, 1994; Packard & McGaugh, 1992, 1996; depend on other structures such as the striatum, as demonstrated in
Schroeder, Wingard, & Packard, 2002). These systems can inter- response learning paradigms (Packard, 1999; Packard & Mc-
act, and the nature of this interaction seems to be competitive Gaugh, 1996; White & McDonald, 2002). Furthermore, given the
because hippocampal lesions or reversible inactivation improve functional heterogeneity of the hippocampal divisions (Gilbert &
the performance on the basis of the striatum-dependent, cue- Kesner, 2003; Gilbert, Kesner, & DeCoteau, 1998; Gilbert,
Kesner, & Lee, 2001; Kesner, Gilbert, & Wallenstein, 2000; Lis-
man, 1999), a differential involvement of CA1, CA3, and dentate
Rubén Miranda, Eduardo Blanco, Azucena Begega, and Jorge L. Arias, gyrus (DG) probably occurs according to the different demands of
Laboratory of Psychobiology, School of Psychology, University of Oviedo, the tasks or depending on the navigation strategies used. For
Oviedo, Spain; Sandra Rubio, Psychobiology Area, School of Psychology, instance, spatial learning impairments have previously been related
Madrid Autonomous University, Madrid, Spain. to specific deficits in spatial and temporal pattern separation pro-
This research was sponsored by Ministerio de Educación y Ciencia cesses (Eichenbaum et al., 1990) that have been associated with
Grant SEJ2004-07445/PSIC. We thank Angel Nistal for her expert tech- CA1 and DG functioning, respectively (Kesner et al., 2000).
nical support inside the imaging analysis facility of the University of
Although in normal conditions (i.e., natural complex environ-
Oviedo and Luis J. Santı́n for his helpful comments and critical reading of
this article.
ments), rats seem to prefer distal landmarks for guiding their
Correspondence concerning this article should be addressed to Rubén behavior in familiar and well-lighted contexts (Whishaw et al.,
Miranda, who is now at the Department of Psychology, University of the 2001), it is also possible that more than one strategy is being used
Balearic Islands, Ed. Beatriu de Pinós (Room 11), Carretera de Valldemossa simultaneously for more efficient navigation (Cain, Beiko, &
km 7.5, 07122 Palma de Mallorca, Spain. E-mail: ruben.miranda@uib.es Boon, 1997). Likewise, multiple brain memory systems can be

641
642 MIRANDA, BLANCO, BEGEGA, RUBIO, AND ARIAS

simultaneously activated to process the different types of informa- cycle with lights on from 08.00 to 20.00) at a constant room temperature
tion relevant for navigation that may determine a particular behav- of 21 ⫹/⫺ 2 °C with ad lib access to food and water. All experimental
ioral output (i.e., navigation strategy). Some controversy exists procedures conducted on the rats were approved by a local veterinary
about which cues are preferred during navigation when multiple committee from the University of Oviedo’s vivarium, and subsequent
handling strictly followed the European Communities Council’s Directive
sources of information are relevant simultaneously. Recently,
86/609 and the American Psychological Association’s ethical code.
Okaichi (2001) proposed that a hierarchy in the use of problem-
Rats were assigned randomly to the different groups that corresponded
solving strategies exists by suggesting that place strategy is the with the different training and control conditions: (a) the hidden platform
favorite one followed by cue strategy and, finally, response strat- (HP) group (Group 1), which performed the HP task (n ⫽ 15 water maze,
egy. Nevertheless, Hamilton et al. (2004) and Baldi, Lorenzini, and n ⫽ 9 cytochrome); (b) the visual fixed (VF) group (Group 2), which
Bucherelli (2003) demonstrated that rats can use a combination of performed the VF task (n ⫽ 10 water maze, n ⫽ 9 cytochrome); (c) the
local and procedural strategies. These results indicate that, in visual moving (VM) group (Group 3), which performed the VM task (n ⫽
conditions in which different cues (i.e., distal and proximal) exist, 15 water maze, n ⫽ 10 cytochrome); (d) the swimming control (SC) group
rats can use both sources of information, which displays a com- (Group 4; n ⫽ 15 water maze, n ⫽ 10 cytochrome); and (e) the caged
bination of strategies. This implies cooperation between cognitive control (CC) group (Group 5; n ⫽ 10 cytochrome).
Rats submitted to a water maze task were assigned pseudorandomly to
strategies that may require a cooperative functional interaction
CO histochemistry groups. Only brain samples in optimal conditions for
between brain memory systems.
densitometric analysis were included in the study.
In this work, behavioral analysis of the strategies displayed in
the different water maze tasks was done in an attempt to clarify the
previous controversy. After delineating the representational or Water Maze
associative demands of the tasks, we analyzed the oxidative me- We trained the rats in several Morris water maze escape tasks (Morris,
tabolism of the hippocampal and striatal systems, trying to better 1981, 1984) using a 1.5-m-diameter black fiberglass pool, which was 75
understand the functional interactions between brain memory sys- cm high and 50 cm above the floor. The pool was filled with tap water to
tems that control spatial abilities. Because the participation of a height of 32 cm, and a black escape platform was placed 2 cm beneath
different brain areas in solving navigation problems can change the water surface. The water temperature was kept at 21 °C during the
when any of them are not functioning (Sutherland & Hamilton, entire test period. The experimental room had numerous visual cues (such
2004), functional studies can provide interesting and useful infor- as colored maps, posters, and plastic dishes fixed on the walls), a shelf,
covered windows, and a table. Lighting was provided by two halogen
mation for describing relationships among multiple memory sys-
spotlights (500 W) placed on the floor and facing the walls. We recorded
tems during normal cognitive processes in the intact brain (Co-
all swim paths of the rats live and later analyzed them using a computerized
lombo, Brightwell, & Countryman, 2003). Considering this, from video-tracking system (Smart, Panlab, Spain).
a noninvasive approach, we decided to test the value of cyto-
chrome c-oxidase (CO) quantitative histochemistry as a potential
functional marker to reveal specific changes in activity of different HP Task
brain areas correlated with the gradual acquisition of complex The entire protocol of the HP task took 5 days. On Day 1, rats received
tasks. CO is a key enzyme in cellular oxidative metabolism be- two habituation trials 60 s in duration with an intertrial interval of 30 s
cause it participates in the mitochondrial electron transport chain remaining in a black plastic bucket. During these trials, rats were released
involved in the direct production of energy as ATP that uses most facing the pool wall from two different points and were allowed to freely
of the available oxygen. Local CO activity is considered a reliable explore the water maze. After habituation (Days 2–5), we trained the rats
marker of brain oxidative metabolic potential because of the close daily using a single six-trial session. In each trial, the rats were released
randomly from one of four start locations and had to search for a hidden
relationship between energy metabolism and neuronal activity
escape platform beneath the water surface. The platform was located in the
(Wong-Riley, 1989). In contrast to other metabolic markers, CO
same position during the training days. Rats failing to find the platform
activity can be used as an index of the sustained energy require- after 60 s were gently guided to it by hand, where they remained for 15 s
ments of nervous tissue associated with brain function (Gonzalez- before being introduced into a black plastic bucket for 30 s. On Day 5 after
Lima & Cada, 1994; Gonzalez-Lima & Jones, 1994). Moreover, the last training trial, rats were submitted to a one-trial probe test in which
this technique has been shown to be sufficiently sensitive in the platform was removed, and the rats were allowed to swim for 30 s.
detecting changes in local cerebral oxidative metabolism induced
after associative learning (Nair, Berndt, Barrett, & Gonzalez-Lima, VF Task
2001; Poremba, Jones, & Gonzalez-Lima, 1998) and has been used
to assess functional changes correlated with spatial reference (Vil- The general procedure for this task was the same as for the HP task,
lareal, Gonzalez-Lima, Berndt, & Barea-Rodriguez, 2002) and which consisted of a habituation day, a training phase, and a probe test. In
working memory (Conejo, Gonzalez-Pardo, Vallejo, & Arias, this task, the escape platform was painted white and projected 2 cm above
the water surface. A cylindrical object (3 cm diameter/5 cm height) was put
2004).
onto the platform as a visual beacon. For each trial the position of the
platform was the same, and only the start point was changed.
Method
VM Task
Rats and Groups
For this task, we used the same procedure previously described but
A total of 65 adult Wistar rats (approximately 90 days old) that weighed changed the position of the visual escape platform. Here the location of the
300 ⫹/⫺ 50 g were obtained from the University of Oviedo’s central platform was changed for each trial to avoid rats using the distal landmarks
vivarium. They were housed under standard conditions (12-hr light– dark of the room.
HIPPOCAMPAL AND CAUDATE METABOLIC ACTIVITY 643

SC Task Germany). These thin sections were put on glass slides previously cleaned
with alcohol (100%) and then stored at ⫺80 °C.
In this task, rats were placed in the water maze in the absence of an
escape platform, which allowed them to swim freely. The procedure was
conducted during the same number of trials and training days as in the HP CO Histochemistry
task. The duration of each trial coincided with the mean escape latencies
performed by the rats trained in the HP task. Sections from the brains were used to perform CO histochemistry. We
used a modified version of the method originally described by Wong-Riley
(1979), which was based on the method developed by Sukekawa (1991). In
Behavioral Analysis brief, slides were lightly fixed for 5 min with a 1.5% glutaraldehyde
solution, rinsed 3 times in phosphate buffer, and incubated at 37 °C for 2
During the training phase in each task, we recorded the distance covered
hr in the dark and with continuous stirring in a solution that contained 50
to reach the platform and the escape latencies. We performed statistical
mg 3,3⬘-diaminobenzidine, 15-mg cytochrome c (Sigma Chemical, St.
analyses using the mean values of these parameters for each daily training
Louis, Missouri), 4 g sucrose per 100 ml phosphate buffer (pH 7.4; 0.1 M).
session. For analysis of the probe test performance, two different virtual
The slides were rinsed 3 times with phosphate buffer, dehydrated, and
divisions of the pool were used: quadrant and ring divisions. First, the pool
coverslipped with Entellan (Merck, Whitehouse Station, New Jersey).
was divided virtually into four quadrants (A, B, C, and D/escape), and
second, it was divided into concentric circular regions or rings with the
same area (border, Ring 2, Ring 3/platforms, and nucleus; see Figure 1). Quantitative Densitometric Analysis
On the basis of these divisions, the mean percentage of the time swam in
each sector was analyzed. We measured CO histochemical staining intensity by densitometric
analysis using a charge-coupled device digital camera (Leica DC-300; E
Histochemistry of the CO Licht Company, Denver, Colorado) coupled to a light microscope (Olym-
pus, Tokyo, Japan) and image analysis software (Leica Q-Win; E Licht
Tissue Preparation Company, Denver, Colorado). A minimum of six measurements, from 8
bits gray scale images, were taken in the dorsal striatum (i.e., caudate-
Rats from control and trained groups were deeply anesthetized by putamen) and CA1, CA3, and DG of the antero-dorsal hippocampus from
intraperitoneal injection of sodium pentobarbital (70 mg/kg) 1 hr after each rat and expressed as arbitrary units of optical density. Delimitation of
finishing the behavioral procedure. Rats were transcardially perfused with the hippocampus and striatum was performed according to Paxinos and
a phosphate buffer solution (pH 7.4; 0.1 M). Their brains were quickly Watson’s (1986) atlas. Antero-posterior coordinates for the studied regions
removed and coronally sectioned to obtain thick slices extending from were approximately as follows: from ⫺1.30 mm to ⫺2.80 mm for caudate-
optic chiasm to the caudal hypothalamic area according to Paxinos and putamen complex and from ⫺2.56 mm to 3.80 mm for hippocampus from
Watson’s (1986) atlas. Then, the slices were immersed in a 20% sucrose bregma. Similar to the procedure described by Nair and Gonzalez-Lima
buffer solution at 4 °C until sinking (approximately 12 hr). Tissue was (1999), to establish comparisons and count for possible staining variations
embedded in a cryoprotective gel (optimal cutting temperature compound; between brain sections from different staining baths, we normalized mea-
Miles, Elkhart, Indiana) frozen in Freon-22 (chlorodifluoromethane, surements by means of a ratio between the values of the mean hippocampal
CIF2CH) and stored at ⫺70 °C. Next, 20-␮m-thick sections from the brains activity and subregion activity. To compare the activity measurements of
were obtained at ⫺20 °C with a cryostat microtome (Microm, Heidelberg, the striatum with the hippocampus, we used subtracted values.

Figure 1. Virtual divisions of the Morris water maze used for probe test analyses. Left: Quadrant division (A,
B, C, and D/escape). The escape platform was placed in Quadrant D in tasks with the platform in a fixed position
(i.e., hidden platform and visual fixed). Right: Ring division (border, Ring 2, Ring 3/platforms, and nucleus).
Escape platform was always placed in Ring 3.
644 MIRANDA, BLANCO, BEGEGA, RUBIO, AND ARIAS

Statistical Analysis differences between groups, F(2, 37) ⫽ 12.385, p ⬍ .001, with no
significant differences in the interaction between both factors, F(6,
We analyzed escape latencies or path lengths in each trained group using
70) ⫽ 1.076, ns. Bonferroni’s post hoc test showed statistical
one-way repeated measures analyses of variance (ANOVAs), with training
significant differences between Day 1 and Days 2, 3, and 4, and
day as the repeated-measures factor and group as the intersubjects factor.
Bonferroni’s post hoc tests were used to assess differences between days, between Day 2 and Days 1, 3, and 4 ( p ⬍ .05). No significant
and Dunnett’s tests were used to assess differences between groups be- differences existed between Days 3 and 4. Dunnett’s post hoc test
cause the variances obtained were unequal. For the probe test performance, revealed that HP rats’ latencies were significantly greater than
two different analyses were made: (a) We performed intragroup analysis those of VF and VM groups ( p ⬍ .05) and that no differences
using a one-way repeated measures ANOVA, with mean percentage time existed between VF and VM groups (see Figure 2).
spent by sector as the repeated-measures factor, and (b) we performed
intergroup analysis using a one-way ANOVA, with group as the intersub-
Distance Covered
jects factor and mean percentage time spent in escape quadrant as the
dependent variable. Bonferroni’s post hoc tests were performed to assess As shown in Figure 2, the mean path lengths measured signif-
differences in time spent in each sector for each group and to assess icantly decreased during training days in all groups, F(3, 35) ⫽
differences between groups in the time spent in Quadrant D (escape).
27.809, p ⬍ .001. Group effect was significant, F(2, 37) ⫽ 17.035,
We analyzed CO activity differences between main hippocampal divi-
p ⬍ .001, but the interaction was not, F(6, 70) ⫽ 1.079, ns.
sions (CA1, CA3, and DG) in each group separately using repeated
measures ANOVAs. Posteriori pairwise comparisons were made to reveal Bonferroni’s post hoc test revealed significant differences between
activity differences between regions. Group differences in CO-transformed Day 1 and Days 2, 3, and 4, and between Day 2 and Days 1, 3, and
activity from each main hippocampal division were assessed by one-way 4 ( p ⬍ .05), which corresponded with the greater distances with
ANOVAs. Bonferroni’s post hoc tests were used to assess differences in Days 1 and 2. The optimal performance was reached at the end of
CO activity between means when ANOVAs indicated significant overall Day 2 of training because no differences existed between Days 3
differences. and 4. Dunnett’s post hoc test showed that the distance covered by
Finally, one-way ANOVAs were used to assess differences in CO VF rats was significantly less than that covered by HP and VM rats
hippocampal-striatum subtracted activity, and Bonferroni’s post hoc tests ( p ⬍ .05).
were used to assess differences in CO activity between groups.
Overall activity of control rats (SC) was similar to HP rats
because a one-way repeated measures ANOVA showed nonsig-
Results nificant differences in the mean distance covered between both
groups, F(1, 26) ⫽ 1.368, ns.
Behavioral Tests
Escape Latencies Probe Test
Mean escape latencies significantly decreased during training Quadrant division. Intragroup analyses for the mean percent-
days, F(3, 35) ⫽ 31.666, p ⬍ .001, which resulted in significant age time spent in each quadrant of the pool showed statistically

40 1000
Mean distance covered (+/-SEM) (cm.)
Mean escape latency (+/- SEM) (sec.)

HP HP
35 VF VF
VM 800 VM
30

25 600

20

400
15

10
200
5

0 0
1 2 3 4 1 2 3 4

Days Days

Figure 2. Left: Mean escape latencies performed during training. Visual fixed (VF) and visual moving (VM)
groups displayed significantly lower latencies compared with the hidden platform (HP) group. All groups
reached asymptotic behavior at the end of the 2nd training day. There were significant differences between Day
1 compared with Days 2, 3, and 4, and between Day 2 compared with Days 1, 3, and 4, which corresponded to
the greater latencies with Days 1 and 2. Right: Mean distance covered during training. VF rats covered
significantly lower distances compared with HP and VM rats. All groups reached asymptotic behavior at the end
of the 2nd training day. Significant differences were found between Day 1 compared with Days 2, 3, and 4, and
between Day 2 compared with Days 1, 3, and 4. No differences were found between the HP and swimming
control groups. Error bars represent standard error of the means (SEMs).
HIPPOCAMPAL AND CAUDATE METABOLIC ACTIVITY 645

significant differences in the HP group, F(3, 12) ⫽ 49.674, p ⬍ tween groups, F(3, 51) ⫽ 39.034, p ⬍ .001. Post hoc analysis
.001; VF group, F(3, 7) ⫽ 6.993, p ⫽ .016; and SC group, F(3, showed that rats trained in the HP task spent significantly more
12) ⫽ 8.705, p ⬍ .001. Rats trained in the VM task showed no time in the quadrant of the maze where the escape platform was
preference for swimming in any of the quadrants because no constantly located in the HP and VF tasks compared with the other
significant differences were found, F(3, 12) ⫽ 1.118, ns. groups ( p ⬍ .05; see Figure 3).
Bonferroni’s post hoc analysis revealed that HP rats swam Ring division. Intragroup analyses for the mean percentage
significantly longer in Quadrant D (escape) compared with the rest time spent in each circular sector of the pool showed statistically
of the quadrants ( p ⬍ .05). Rats trained in the VF task spent more significant differences in the HP group, F(3, 12) ⫽ 14.722, p ⬍
time swimming in Quadrants A and D/escape compared with .001; VF group, F(3, 7) ⫽ 6.767, p ⫽ .018; VM group, F(3, 12) ⫽
Quadrants B and C ( p ⬍ .05). SC rats spent significantly less time 17.399, p ⬍ .001; and SC group, F(3, 12) ⫽ 548.682, p ⬍ .001.
in swimming in Quadrant B compared with Quadrants A, C, and Post hoc analyses revealed that HP rats spent significantly more
D/escape ( p ⬍ .05; see Figure 3). time in Rings 2 and 3 compared with the nucleus ( p ⬍ .05). Rats
Intergroup analysis for the mean percentage time spent in Quad- trained in the VF task swam significantly longer in Ring 2 com-
rant D (escape) revealed statistically significant differences be- pared with the border ( p ⬍ .05). VM rats spent more time in Rings

70 QUAD A
QUAD B
Mean percentage time in quadrants

QUAD C
60
QUAD D (Escape)

50

40

30

20 * * *
* *
*
10

0
HP VF VM SC

Figure 3. Results of the probe test analyses that used the virtual quadrant division. Top: Representative
trajectories during the probe test of the hidden platform (HP) group (left), the visual moving (VM) group
(middle), and the swimming control (SC) group (right). Bottom: HP rats spent significantly more time
(percentage) in Quadrant D/escape compared with Quadrants A, C, and B. Visual fixed (VF) rats spent significantly
more time (percentage) in Quadrant D/escape compared with Quadrants B and C. No differences existed in the
time spent in the escape quadrant compared with Quadrant A/start. VM rats showed no differences in the time
spent in each quadrant. SC rats spent significantly less time in Quadrant B during the last training day. Group
comparison of the mean percentage time spent in Quadrant D/escape showed that the HP rats spent significantly
more time swimming in the escape quadrant (error bars represent standard error of the means; *p ⬍ .05).
646 MIRANDA, BLANCO, BEGEGA, RUBIO, AND ARIAS

2 and 3 compared with the nucleus ( p ⬍ .05). Control rats (SC) CO subtracted activity between hippocampus and striatum as the
spent significantly more time in the border compared with the rest dependent variable. Overall, statistically significant differences
of the sectors ( p ⬍ .05; see Figure 4). were detected, F(4, 42) ⫽ 19.904, p ⬍ .001, and a Bonferroni’s
post hoc test revealed that the activity in the CC group was
CO Activity significantly lower than the rest of the groups and that the activity
in the VM group was higher than the rest of the groups ( p ⬍ .05).
Hippocampal Activity No significant differences were detected between the SC, VF, and
Intragroup analyses for CO activity differences among main HP groups (see Figure 6).
hippocampal divisions (CA1, CA3, and DG) showed statistically
significant differences in each group ( p ⬍ .001). Posterior pair- Discussion
wise comparisons revealed a higher CO activity in CA3 followed
The main goal of the current work was to assess the functional
by DG and CA1, with the lowest CO activity in the HP group ( p ⬍
activity, by means of the CO quantitative histochemistry, of the
.05). In the remaining groups, CO activity of CA1 was signifi-
hippocampus and dorsal striatum related to different water maze
cantly lower than CA3 and DG ( p ⬍ .05). No significant differ-
tasks involving multiple navigation strategies. The behavioral per-
ences existed between CA3 and DG in these groups.
formance of the rats was measured as distance covered and escape
To assess group differences in CO activity of the main hip-
latency. All rats mastered the tasks after 2 days of training (i.e., 12
pocampal divisions, we performed one-way ANOVAs that com-
trials), although VF rats displayed shorter latencies and trajectories
pared the mean CO activity that resulted from transforming CO
compared with HP rats, and VM rats displayed long trajectories
values into activity ratios. These ratios were established as follows:
similar to HP rats with short latencies similar to VF rats (see
hippocampal division CO activity (CA1, CA3, or DG)/mean hip-
Figure 2). These results reflect that when proximal and distal cues
pocampal CO activity.
are available (i.e., VF task), rats can use both sources of informa-
ANOVAs showed statistically significant differences among
tion to locate a place. It also seems that the presence of a proximal
groups in CA1, F(4, 43) ⫽ 5.345, p ⫽ .001; CA3, F(4, 43) ⫽
cue can facilitate acquisition of the task. This effect was previously
4.694, p ⫽ .003; and DG, F(4, 43) ⫽ 3.708, p ⫽ .011. Post hoc
reported by testing hippocampal rats in a water maze (Morris et al.,
analyses revealed a lower activity in CA1 of SC rats compared
1982) and in a radial maze (Ramos, 2002), suggesting a differen-
with VF and CC groups ( p ⬍ .05). In CA3, the VM group
tial involvement of the hippocampal system in the standard water
presented a higher activity compared with the VF group ( p ⬍ .05).
maze task (i.e., HP) and in the visual version (i.e., VF). However,
Finally, the CO activity of DG in the SC and HP groups was higher
the behavior displayed by VM rats suggests that these rats might
compared with the CC group ( p ⬍ .05; see Figure 5).
have been using another strategy as well as the cue-guided one. For
instance, according to the results of Baldi et al. (2003), they could
Hippocampal to Caudate Activity
have used a procedural strategy based on the distance relationship
To compare the CO activity of the hippocampal and striatal between the platform location and the wall of the pool. Moreover,
among groups, we performed a one-way ANOVA with the mean it is also possible that rats tried to use the distal landmarks as

80 Border (B)
Ring 2 (R2)
70 Ring 3 (R3)
Mean percentage time in rings

Nucleus (N)
60

50

40

30

20
HP: R2, R3 > N
VF: R2 > B
10
VM: R2, R3 > N
SC: B > R2 > R3 > N
0
HP VF VM SC

Figure 4. Results of the probe test analyses that used the virtual ring division. Hidden platform (HP) rats spent
significantly more time (percentage) swimming in Rings 2 and 3 compared with the nucleus. Visual fixed (VF)
rats spent significantly more time (percentage) in Ring 2 compared with the border. Visual moving (VM) rats
swam significantly more time in Rings 2 and 3 compared with the nucleus. The swimming control (SC) rats
showed significant differences in the mean percentage of time spent in each sector during the last training session
(error bars represent standard error of the means).
HIPPOCAMPAL AND CAUDATE METABOLIC ACTIVITY 647

1.10
CC
SC
1.05 HP
VF
VM
CO activity ratio (a.u.)

1.00

0.95

0.90

0.85

0.80
CA1 CA3 DG

Figure 5. Cytochrome c-oxidase (CO) activity in hippocampal subfields. CA1: There were significant differ-
ences between the swimming control (SC) group and caged control (CC) and visual fixed (VF) groups. CA3: The
activity of the VF group was significantly lower compared with the visual moving (VM) group. Dentate gyrus
(DG): There were significant differences between the CC group compared with the SC and hidden platform (HP)
groups (error bars represent standard error of the means).

another source of information to navigate. Because the platform approximately 75% of time during the last training session (see
position changed in each trial and also its relation to distal land- Figure 4). These data reflect that VM rats, in contrast to SC rats,
marks, using this information, the rats’ initial trajectories would acquired some knowledge about the constant relationship between
have been mistaken and, after not finding the platform in its the platform locations and the walls. Finally, VF rats during the
previous location, they would have to correct the trajectory to probe test showed a preference for swimming in two quadrants of
reach the new platform’s location on the basis of the proximal cue the pool (start and escape). According to a recent report by
(i.e., visible platform). This strategy selection, recently reported by Hamilton et al. (2004), these results suggest that these rats prob-
Hamilton et al. (2004), causes rats to take longer and less accurate ably used a combination of place and cue strategies during the
trajectories that cover greater distances, which possibly explains training and probe test. If these rats were only using a place
the differences compared with the VF group. However, from the strategy during training, probably due to overshadowing of the
sole analysis of the distance covered, we cannot confirm or disre- intramaze cues (Chamizo, Sterio, & Mackintosh, 1985), then the
gard that rats are using this kind of place strategy involving spatial results of the probe test should have been similar to the HP group,
working memory. swimming preferentially only in the escape quadrant. However,
Probe test analyses confirmed that HP rats preferentially swam they swam a similar time in both escape and start quadrants, and
in the escape quadrant, indicating that they presumably were using group comparisons showed that the HP rats swam significantly
distal landmarks and a spatial strategy to navigate. However, rats longer in the escape quadrant. In comparison, if these rats were
trained in the VM task, in the absence of the visible platform, navigating using only a cue strategy, probably due to a reduction
showed no preference for any quadrant and displayed a circular in the attention and associability of the spatial cues derived from
swimming behavior similar to SC rats (see Figure 3). This indi- the habituation phase, then they would have shown no preference
cates that in the VM task, the distal landmarks of the room were for any quadrant, as occurred with the VM group (see Figure 3).
not important to identify a target place defined by quadrants; Together, the results of the behavioral analyses show that rats use
however, this does not necessarily imply that no spatial relation- different navigation strategies depending on the water maze train-
ships were established between room landmarks and different ing. When rats are trained to find a hidden platform and distal cues
water maze locations. With the aid of the virtual ring division of are available, the rats preferentially use a local strategy. If rats
the pool, we observed that all groups trained with the escape have to reach a visible platform placed at a fixed location, they use
platform (i.e., HP, VF, and VM) in the probe trial spent nearly 60% a combination of place and cue-guided strategies. Finally, if the
of the time swimming inside Rings 2 and 3 (platforms ring), visible platform is placed at different locations and at a specific
whereas the SC group remained swimming in the border area for distance from the wall of the pool, then the rats use a cue-guided
648 MIRANDA, BLANCO, BEGEGA, RUBIO, AND ARIAS

50
CC
SC *
HP
40

CO substracted activity (a.u.)


VF
VM

30

20 *

10

Figure 6. Comparison of the cytochrome c-oxidase (CO) activity of the striatum relative to the hippocampus.
The caged control (CC) group showed significantly lower activity compared with all groups, and visual moving
(VM) rats showed significantly higher activity also compared with all groups. No significant differences were
found between the swimming control (SC), hidden platform (HP), and visual fixed (VF) groups (error bars
represent standard error of the means; *p ⬍ .05).

strategy combined with a procedural strategy. In general, it seems CA3 after VM training can be related to a spatial working memory
that rats, instead of using only one strategy as the best one, can processing or more likely with a prolonged associative demand
combine several strategies that attend to all the available informa- during training. Because in this task the platform location, as well
tion to achieve a more efficient navigation. Taking this into ac- as the release points, varied in each trial, possibly new associations
count, in this work we decided to study brain functional correlates between the visible platform, spatial landmarks, and motor re-
between the different tasks and their demands with the activity of sponses would be expected to be formed in each trial. However, it
the hippocampal and striatal systems. For this purpose, we used appears that the mere activation of one hippocampal subfield is not
CO quantitative histochemistry as a potential functional marker, sufficient to explain a task-dependent processing. For example, SC
which is useful to reveal changes in brain oxidative metabolism rats submitted to a simple forced swimming task presented an
induced by sustained energy requirements derived from the grad- unforeseen high activity in DG, similar to HP rats, but both groups
ual acquisition of complex (water maze) tasks (Gonzalez-Lima & clearly differed in the CA3 and CA1 levels of activity (see Figure
Cada, 1994; Gonzalez-Lima & Jones, 1994; Wong-Riley, 1979, 5). Conversely, HP and VF rats presented a similar pattern of
1989). activity with low CO activity in CA3 and high CO activity in CA1
The posttraining functional study revealed a putative differential and DG. Nevertheless, from our data, it is unclear which kind of
participation of hippocampal main divisions and dorsal striatum in demand can alter hippocampal CO levels in SC rats, as it is
the water maze tasks. CO activity of CA1 and DG was signifi- possible that the increase in CO activity observed in DG reflects a
cantly higher in VF and HP groups, respectively, whereas CA3 spatial processing of the room containing the water maze.
activity was higher in the VM group (see Figure 5). As discussed Cumulative evidence has suggested that the hippocampus is
previously, VF and HP groups can be considered as place learners, involved in multiple memory types (Squire, 1992) and has empha-
and the VM group demonstrated a navigation based on a cue sized its important role in spatial learning (Eichenbaum et al.,
strategy, although they also exhibited acquisition of procedural 1990; Morris et al., 1982; O’Keefe & Nadel, 1978). Nevertheless,
skills (i.e., circular swimming at a specific distance from the wall it is becoming clear that spatial learning may not depend exclu-
of the pool) that may have involved some kind of intramaze sively on the hippocampus because rats with hippocampal lesions
allothesis (Baldi et al., 2003). Hence, it cannot be disregarded that can solve spatial tasks using nonspatial strategies (Alyan et al.,
these rats have not learned the task completely independently of 2000; Pearce, Roberts, & Good, 1998; Pouzet et al., 2002;
spatial processing; indeed, spatial working memory can also be Whishaw et al., 2001). These findings suggest that multiple brain
involved. systems interact during normal navigation and that a relationship
Spatial and temporal pattern separation processes have been exists between the systems’ activity and behavioral strategies. The
reported to be essential to solve spatial tasks, and a failure in these striatum is continuously associated with cue navigation and re-
functions can explain spatial learning deficits (Eichenbaum et al., sponse learning, and many studies have reported functional disso-
1990; Kesner et al., 2000). Our results confirm that CA1 and DG ciations and interactions between hippocampal and striatal activity
have an important role in the acquisition of water maze tasks on navigation (Colombo et al., 2003; Compton, 2004; McDonald
requiring spatial processing. However, the increased activity of & White, 1994; Mizumori, Yeshenko, Gill, & Davis, 2004; Okai-
HIPPOCAMPAL AND CAUDATE METABOLIC ACTIVITY 649

chi, 2001; Packard, 1999; Packard & McGaugh, 1992, 1996; Colombo, P. J., Brightwell, J. J., & Countryman, R. A. (2003). Cognitive
Schroeder et al., 2002; White & McDonald, 2002). strategy-specific increases in phosphorylated cAMP response element-
In our work, we have introduced a novel functional approach to binding protein and c-Fos in the hippocampus and dorsal striatum.
determine in which training or control condition either the striatum Journal of Neuroscience, 23, 3547–3554.
or the hippocampus has a greater participation. All groups that Compton, D. M. (2004). Behavior strategy learning in rat: Effects of
lesions of the dorsal striatum or dorsal hippocampus. Behavioural Pro-
were submitted to a water maze task presented an increase in CO
cesses, 67, 335–342.
activity in the striatum (relative to the hippocampus) compared Conejo, N. M., Gonzalez-Pardo, H., Vallejo, G., & Arias, J. L. (2004).
with the CC group. This increase was especially higher in the VM Involvement of the mammillary bodies in spatial working memory
group, suggesting that cue/response learning involves a greater revealed by cytochrome oxidase activity. Brain Research, 1011, 107–
energy demand for the striatum than for the hippocampus (see 114.
Figure 6). Although it is well known that the striatum plays a key Eichenbaum, H., Stewart, C., & Morris, R. G. M. (1990). Hippocampal
role in motor activity, it is unlikely that our results can be ex- representation in place learning. Journal of Neuroscience, 10, 3531–
plained by a specific greater locomotor activity or motor experi- 3542.
ence in the VM rats. During water maze training, the longer escape Gilbert, P. E., & Kesner, R. P. (2003). Localization of function within the
latencies and distances covered corresponded to the SC group that dorsal hippocampus: The role of the CA3 subregion in paired-associate
presented low CO activity levels, similar to HP and VF rats. learning. Behavioral Neuroscience, 117, 1385–1394.
Gilbert, P. E., Kesner, R. P., & DeCoteau, W. E. (1998). Memory for
Furthermore, it has been reported that a long period of physical
spatial location: Role of the hippocampus in mediating spatial pattern
exercise (6 months) is required to observe a motor activity-related
separation. Journal of Neuroscience, 18, 804 – 810.
increase in the CO levels of the striatum (McCloskey, Adamo, & Gilbert, P. E., Kesner, R. P., & Lee, I. (2001). Dissociating hippocampal
Anderson, 2001). subregions: A double dissociation between dentate gyrus and CA1.
In summary, our results indicate that rats may use as much Hippocampus, 11, 626 – 636.
information as they can to find an escape platform by combining Gonzalez-Lima, F., & Cada, A. (1994, November). Cytochrome oxidase
more than one navigation strategy. Hippocampal and striatal sys- activity in the auditory system of the mouse: A qualitative and quanti-
tems can also be activated in parallel, suggesting a functional tative histochemical study. Neuroscience, 63(2), 559 –578.
participation of both systems to solve spatial problems with dif- Gonzalez-Lima, F., & Jones, D. (1994). Quantitative mapping of cyto-
ferent cognitive demands. CO histochemistry appears to be suffi- chrome oxidase activity in the central auditory system of the gerbil: A
ciently sensitive to detect changes in prolonged neuronal activity study with calibrated activity standards and metal-intensified histochem-
as a consequence of relatively more complex tasks compared with istry. Brain Research, 660, 34 – 49.
Hamilton, D. A., Rosenfelt, C. S., & Whishaw, I. Q. (2004). Sequential
conditioning paradigms in which this technique was used previ-
control of navigation by locale and taxon cues in the Morris water maze.
ously (Nair et al., 2001; Poremba et al., 1998). Our results are in
Behavioural Brain Research, 154, 385–397.
agreement with a previous functional study by Colombo et al. Kesner, R. P., Gilbert, P. E., & Wallenstein, G. V. (2000). Testing neural
(2003), which reported more phosphorylation of cAMP response network models of memory with behavioral experiments. Current Opin-
element binding protein and c-Fos expression in the hippocampus ion in Neurobiology, 10, 260 –265.
and striatum of place and response learners, respectively. Here, Lisman, J. E. (1999). Relating hippocampal circuitry to function: Recall of
with the aid of the CO technique, we observed increased energy memory sequences by reciprocal dentate–CA3 interactions. Neuron, 22,
metabolism in (a) hippocampal CA1 and DG, after the processing 233–242.
of local landmarks underlying a place strategy as well as in (b) McCloskey, D. P., Adamo, D. S., & Anderson, B. J. (2001). Exercise
hippocampal CA3 and dorsal striatum, following associative pro- increases metabolic capacity in the motor cortex and striatum, but not in
cessing of proximal (taxon) cues underlying a cue-guided or re- the hippocampus. Brain Research, 891, 168 –175.
sponse strategy. Further studies with functional markers such as McDonald, J., & White, N. M. (1994). Parallel information processing in
the water maze: Evidence for independent memory systems involving
CO, with a focus on other cerebral regions, could be useful to
dorsal striatum and hippocampus. Behavioral and Neural Biology, 61,
clarify the interaction between multiple brain memory systems
260 –270.
during spatial learning in intact brains. Mizumori, S. J. Y., Yeshenko, O., Gill, K. M., & Davis, D. M. (2004).
Parallel processing across neural systems: Implications for a multiple
References memory system hypothesis. Neurobiology of Learning and Memory, 82,
278 –298.
Alyan, S. H., Jander, R., & Best, P. J. (2000). Hippocampectomized rats Morris, R. G. M. (1981). Spatial localization does not require the presence
can use a constellation of landmarks to recognize a place. Brain Re- of local cues. Learning and Motivation, 12, 239 –260.
search, 876, 225–237. Morris, R. G. M. (1984). Development of a water-maze procedure for
Baldi, E., Lorenzini, C. A., & Bucherelli, C. (2003). Task solving by studying spatial learning in the rat. Journal of Neuroscience Methods,
procedural strategies in the Morris water maze. Physiology & Behavior, 11, 47– 60.
78, 785–793. Morris, R. G. M., Garrud, P., Rawlings, J. N. P., & O’Keefe, J. (1982, June
Cain, P. D., Beiko, J., & Boon, F. (1997). Navigation in the water maze: 24). Place navigation impaired in rats with hippocampal lesions. Nature,
The role of proximal and distal visual cues, path integration, and mag- 297, 681– 683.
netic field information. Psychobiology, 25, 286 –293. Nair, H. P., Berndt, J. D., Barrett, D., & Gonzalez-Lima, F. (2001).
Chamizo, V. D., Sterio, D., & Mackintosh, N. J. (1985). Blocking and Metabolic mapping of brain regions associated with behavioral extinc-
overshadowing between intra-maze and extra-maze cues: A test of the tion in pre-weaning rats. Brain Research, 903, 141–153.
independence of locale and guidance learning. Quarterly Journal of Nair, H. P., & Gonzalez-Lima, F. (1999). Extinction of behavior in infant
Experimental Psychology: Journal of Comparative and Physiological rats: Development of functional coupling between septal, hippocampal,
Psychology, 39(B), 107–125. and ventral tegmental regions. Journal of Neuroscience, 19, 8646 – 8655.
650 MIRANDA, BLANCO, BEGEGA, RUBIO, AND ARIAS

Okaichi, H. (2001). Effects of dorsal-striatum lesions and fimbria-fornix Ramos, J. M. J. (2002). Es necesario el hipocampo para el aprendizaje
lesions on the problem-solving strategies of rats in a shallow water maze. espacial? [Is the hippocampus necessary for spatial learning?]. Revista
Cognitive, Affective, & Behavioral Neuroscience, 1, 229 –238. de Neurologı́a, 34, 1142–1151.
O’Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map. Schroeder, J. P., Wingard, J. C., & Packard, M. G. (2002). Post-training
Oxford, England: Oxford University Press. reversible inactivation of hippocampus reveals interference between
Packard, M. G. (1999). Glutamate infused posttraining into the hippocam- memory systems. Hippocampus, 12, 280 –284.
pus or caudate-putamen differentially strengthens place and response Squire, L. R. (1992). Memory and the hippocampus: A synthesis from
learning. Proceedings of the National Academy of Sciences, USA, 96, findings with rats, monkeys, and humans. Psychological Review, 99,
12881–12886. 195–231.
Packard, M. G., Cahill, L., & McGaugh, J. L. (1994). Amygdala modula- Sukekawa, K. (1991). A fresh mount method for cytochrome oxidase
tion of hippcampal-dependent and caudate nucleus-dependent memory histochemistry. Biotechnic & Histochemistry, 1, 99 –101.
processes. Proceedings of the National Academy of Sciences, USA, 91, Sutherland, R. J., & Hamilton, D. A. (2004). Rodent spatial navigation: At
the crossroads of cognition and movement. Neuroscience & Biobehav-
8477– 8481.
ioral Reviews, 28, 687– 697.
Packard, M. G., & McGaugh, J. L. (1992). Double dissociation of fornix
Villareal, J. S., Gonzalez-Lima, F., Berndt, J., & Barea-Rodriguez, E. J.
and caudate nucleus lesions on acquisition of two water maze tasks:
(2002). Water maze training in aged rats: Effects on brain metabolic
Further evidence for multiple memory systems. Behavioral Neuro-
capacity and behavior. Brain Research, 939, 43–51.
science, 106, 439 – 446.
Whishaw, I. Q., Hines, D. J., & Wallace, D. G. (2001). Dead reckoning
Packard, M. G., & McGaugh, J. L. (1996). Inactivation of hippocampus or
(path integration) requires the hippocampal formation: Evidence from
caudate nucleus with lidocaine differentially affects expression of place spontaneous exploration and spatial learning tasks in light (allothetic)
and response learning. Neurobiology of Learning and Memory, 65, and dark (idiothetic) tests. Behavioural Brain Research, 127, 49 – 69.
66 –72. White, N. M., & McDonald, R. J. (2002). Multiple parallel memory
Paxinos, G., & Watson, C. (1986). The rat brain in stereotaxic coordinates systems in the brain of the rat. Neurobiology of Learning and Memory,
(2nd ed.). San Diego, CA: Academic Press. 77, 125–184.
Pearce, J. M., Roberts, A. D. L., & Good, M. (1998, November 5). Wong-Riley, M. T. T. (1979). Changes in the visual system of monocularly
Hippocampal lesions disrupt navigation based on cognitive maps but not sutured or enucleated cats demonstrable with cytochrome oxidase his-
heading vectors. Nature, 396, 75–77. tochemistry. Brain Research, 171, 11–28.
Poremba, A., Jones, D., & Gonzalez-Lima, F. (1998). Classical condition- Wong-Riley, M. T. T. (1989). Cytochrome oxidase: An endogenous met-
ing modifies cytochrome oxidase activity in the auditory system. Euro- abolic marker for neuronal activity. Trends in Neurosciences, 12, 94 –101.
pean Journal of Neuroscience, 10, 3035–3043.
Pouzet, B., Zhang, W. N., Feldon, J., & Rawlins, J. N. P. (2002). Hip- Received May 31, 2005
pocampal lesioned rats are able to learn a spatial position using nonspa- Revision received January 23, 2006
tial strategies. Behavioural Brain Research, 133, 279 –291. Accepted January 24, 2006 䡲

You might also like