Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Behavioural Brain Research 193 (2008) 277–288

Contents lists available at ScienceDirect

Behavioural Brain Research


journal homepage: www.elsevier.com/locate/bbr

Research report

Integrating the open field, elevated plus maze and light/dark box to assess
different types of emotional behaviors in one single trial
André Ramos ∗ , Elayne Pereira, Gisele C. Martins, Thaize D. Wehrmeister, Geison S. Izı́dio
Laboratório de Genética do Comportamento, Departamento de Biologia Celular, Embriologia e Genética,
Universidade Federal de Santa Catarina, 88.040-900 Florianópolis, SC, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Current anxiety tests do not provide, individually, a pure and complete picture of an animal’s emotional
Received 13 April 2008 profile. Therefore, many authors test their experimental hypotheses using a series of anxiety-related tests,
Received in revised form 13 May 2008 which are thought to reflect different facets of emotionality. The objective of this study was to investigate
Accepted 5 June 2008
the potential usefulness of integrating three widely used behavioral tests, the open field (OF), elevated
Available online 11 June 2008
plus maze (EPM), and light/dark box (LDB), to assess a wider range of emotional and non-emotional
behaviors within one single trial. A protocol was developed where rats could freely explore an OF that
Keywords:
was physically connected to an EPM and a LDB during 15 min. Classical anxiety- and locomotion-related
Anxiety
Elevated plus maze
behaviors from each test were measured. Lewis and spontaneously hypertensive rats (SHR) inbred strains,
Open field known to present genetic differences in each of the individual tests, differed for all anxiety-related behav-
Light/dark box iors from the combined apparatus. Factor analyses revealed that similar anxiety- and locomotion-related
Behavior genetics factors were produced by the three tests applied either separately or in combination. Under both condi-
Inbred strains tions, each test produced its own anxiety-related factor. Two benzodiazepines, chlordiazepoxide (at 5 and
Rats 10 mg/kg) and midazolam (at 0.75 mg/kg), facilitated the approach towards the EPM open arms, whereas
pentylenetetrazole (10 mg/kg) specifically inhibited exploration of the three aversive areas (OF center,
EPM open arms, LDB light compartment). Together, these results suggest that the new integrated appa-
ratus may contribute to the study of anxiety, by providing a rapid, comprehensive and reliable method of
assessing emotionality-related behaviors and its underlying components.
© 2008 Published by Elsevier B.V.

1. Introduction imize the consequences of such limitations, many authors submit


their experimental subjects to a battery of different tests which,
Since the creation of the open field test, conceived in 1934 to together, are expected to provide a more complete and reliable
provide objective measures of emotionality in rats [18], dozens of picture of an animal’s emotional reactivity [11,27]. Nevertheless,
different tests aiming to assess anxiety-related behaviors in labo- several concerns arise from the use of test batteries, such as the
ratory rodents have appeared in the literature, for reviews see Refs. long-lasting influences of previous test experiences, which were
[1,13,17,23,27,37]. From the field of experimental psychology, the found to be dependent on the animal’s genotype [41], and the need
interest on such tests spread rapidly among neuroscientists, phar- for allowing inter-test intervals (of a few days or weeks) in order to
macologists and, more recently, molecular geneticists. Today, these minimize the effects of test history [24]. In addition to that, ideally, a
tests are part of the screening routine of many laboratories, not only test battery should not be excessively complex or time-consuming
of those working on drug development but also the ones dedicated in order to allow its use in high-throughput behavioral phenotyping
to the behavioral and neurological study of mutant mice and rats studies.
[11,17,40]. Another issue to be considered when using different tests is
Despite their wide diffusion, most users agree that none of how to interpret treatment effects that are seen in one test but
the existing anxiety tests provides a pure, undisputed measure of not in another. Data from correlational studies, in spite of being
emotional reactivity and that each individual test assesses only a somewhat contradictory [28,29], often indicate that there is little
fraction of an animal’s emotional profile [6,13,27]. In order to min- correlation among anxiety-related behaviors measured in different
tests [16,29,38]. Such a lack of correlation is normally interpreted
as a sign of difference in the psychobiological meanings of the var-
∗ Corresponding author. Tel.: +55 48 3721 5153; fax: +55 48 3721 5148. ious tests [16,27,38]. However, because an animal cannot be tested
E-mail address: andre@ccb.ufsc.br (A. Ramos). at the same time in two different tests and as most tests only assess

0166-4328/$ – see front matter © 2008 Published by Elsevier B.V.


doi:10.1016/j.bbr.2008.06.007
278 A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288

temporary emotional states [23], one cannot be sure of how much


of the observed inter-test inconsistency is due to real construct dif-
ferences among tests and how much of it is caused by temporal
changes in the emotional state of an animal that is submitted to
different tests at different times [19].
In view of all the aforementioned matters, several authors have
urged the behavioral neuroscience community to improve the cur- Fig. 1. Diagram representing a top view of the new integrated test, which is a com-
rent methods of behavioral testing. Such an improvement could be bination of the OF (left), EPM (middle) and LDB (right) tests. All three apparatuses
are positioned at the same height (52 cm from the floor) and are interconnected
attained either by conceiving new tests and paradigms [15,21] or by
so that animals can freely move in all areas. Thicker lines represent 1.5-cm-thick
refining the behavioral analysis [22,33] and/or the current design walls; intermediate lines represent the edges of the EPM open arms; and thinner
of existing tests [13,14]. lines represent the lines drawn on the apparatuses’ floors (see Section 2 for more
In the present study, we hypothesize that a more compre- details).
hensive, reliable and rapid assessment of the emotional profile
of laboratory rodents could be obtained by physically combining impermeable formica, had a square floor of 100 cm × 100 cm (divided by black lines
different existing tests that are based on the free exploration of into 25 squares of 20 cm × 20 cm) and white walls, 40-cm high. The EPM was made
aversive environments. To this end, we propose a combination of of wood covered with black formica and had four elevated arms (52 cm from the
floor) 50-cm long and 10-cm wide. The arms were arranged in a cross-like disposi-
the three most widely used tests of anxiety, namely the open field tion, with two opposite arms being enclosed (by 40-cm high walls) and two being
(OF), the elevated plus maze (EPM) and the light/dark box (LDB). open, having at their intersection a central platform (10 cm × 13.5 cm) which gave
With this integrated test, the experimenter should be able to assess access to any of the four arms. The open arms were surrounded by a raised ledge
different aspects of emotionality simultaneously, thus avoiding (1-mm thick and 5-mm high) to avoid rats falling off the arms. Finally, the LDB was
made of wood covered with formica and presented two compartments. One larger
undesirable effects of previous test experiences, inter-test inter-
(27 cm × 27 cm × 27 cm) and white, with the floor divided by black lines in nine
vals and handling between tests. Two pilot studies were devised to squares (9 cm × 9 cm), being strongly illuminated by a 40-W white bulb. The other
determine: whether or not laboratory rats would indeed explore smaller (27 cm × 18 cm × 27 cm high) and black, with the floor divided by white lines
all areas of this integrated test; what would be the best starting in six squares (9 cm × 9 cm), being illuminated by a 40-W red bulb. Both white and
place in the apparatus; and what would be the best duration of a red bulbs were located 30 cm above the apparatus’ floor, thus providing a 750-lx
intensity inside the light compartment and a 25-lx intensity inside the dark com-
test session. Once a general testing protocol was defined, another partment. The two compartments, separated by a wall, were connected by a small
experiment was performed to verify whether the new test, pro- square opening of 7 cm × 7 cm.
visionally called the “triple test”, would be capable of detecting
previously known genetic differences in anxiety-related behaviors. 2.2.2. Integrated format
In all the other experiments (1–3 and 5–8), the tests were used in their new,
Subsequently, factor analyses were performed on behavioral data
integrated format (Fig. 1), where the OF, EPM and LDB apparatuses were physically
obtained from the OF, EPM and LDB, used either separately or in connected by placing them side by side at the same height (52 cm from the floor) and
combination, in order to elucidate the different behavioral dimen- by adding rectangular openings with 10 cm of width at the interface between their
sions assessed under each test condition. Finally, pharmacological respective less aversive areas (i.e. a 10 cm × 40 cm opening between the periphery
experiments were carried out to investigate the effects of three of the OF and one enclosed arm of the EPM and a 10 cm × 7 cm opening between
the other enclosed arm of the EPM and the dark compartment of the LDB). This
anxiety-modulating drugs on the new combined apparatus.
arrangement was intended to encourage transitions between tests. As shown in
Fig. 1, the connections between the OF and EPM and between the EPM and LDB
2. Materials and methods were positioned exactly in the mid portions of the OF and LDB walls, respectively.
A dim light in the test room plus the LDB lights provided 10 lx in the OF, 15 lx in the
2.1. Animals EPM and 15 and 700 lx in the dark and light areas of the LDB, respectively.

A total of 375 animals from four rat lines (Wistar, Lewis, spontaneously hyperten-
2.3. Testing protocols
sive rats and a local heterogeneous stock) were used in a series of eight experiments.
In two pilot studies (experiments 1 and 2), Wistar rats were obtained from the 2.3.1. Experiment 1
colonies of the local animal breeding center. Lewis (LEW) and spontaneously hyper- This was a pilot study designed to verify whether the animals would really
tensive rats (SHR), used in experiment 3, are inbred strains and have been kept explore all three apparatuses as well as to address two basic aspects of the new
in our laboratory under a system of brother-sister mating for more than 20 gen- triple test: the starting point and the duration of the testing sessions. To this end,
erations. In experiment 4, a heterogeneous stock, also born in our laboratory, was 24 Wistar females were randomly dived in three experimental groups that were
used. It derived from the intercross between three different rat lines, as previously placed in one of three starting points: the center of the OF; the central platform of
described [32]. Briefly, male Wistar rats were mated with female Hooded rats and the EPM facing an open arm; and the white compartment of the LDB. The following
their female offspring were mated with male Lewis rats. Males and females of this behaviors were registered through direct visual observation from 0 to 15 and from 15
last generation were intercrossed, thus producing a genetically heterogeneous stock to 30 min. In the OF: number of peripheral squares (adjacent to the walls) crossed
of animals. For experiments 5 and 6, Wistar rats were purchased from UNIVALI (Ita- (outer locomotion) and number of central squares (away from the walls) crossed
jai, Brazil) and were used after a period of 3 weeks of rest inside our laboratory. (inner locomotion). In the EPM: the number of entries with all four paws inside
Their Wistar offspring, born in our laboratory, were used in experiments 7 and 8. All each type of arm and the time spent with all four paws inside the open arms. In the
animals were weaned and separated by sex at 3–4 weeks of age and thereafter were LDB: time spent with all four paws in each compartment, and number of transitions
kept in collective plastic cages (5–6 rats per cage) with food and water available ad between compartments (one entry into the light, plus one return to the dark). The
libitum, under a 12-h light/dark schedule (lights on at 7:00 a.m.), at 22 ± 2 ◦ C. All apparatus was cleaned with water using a wet sponge and a paper towel before the
animals, naı̈ve to any experimental procedure, were tested between 9 and 11 weeks introduction of each animal, a procedure that was kept for all experiments.
of age with all tests being carried out between 1:00 and 6:00 p.m. All procedures
were carried out according to the guidelines of the local Commitee for Animal Care
2.3.2. Experiment 2
in Research (CEUA/UFSC).
Once defined, based on experiment 1, that a 15-min trial initiated in the OF was
enough to produce exploration of all three apparatuses, a second pilot study was
2.2. Test apparatus designed to determine the best starting point inside the OF and to investigate the
effects of even shorter trials. To this end, 20 Wistar males were randomly divided in
The new test apparatus (Fig. 1) was an integration of three well-known tests two experimental groups that were placed either in the center or the periphery (at
of anxiety/emotionality, namely the open field (OF), elevated plus maze (EPM) and the corner of two walls, away from the EPM opening) of the OF. The same behaviors
light/dark box (LDB). listed above were measured from 0 to 5, 5 to 10 and 10 to 15 min.

2.2.1. Single format 2.3.3. Experiment 3


In experiment 4, each of these tests was used in its regular, individual format, Once defined, based on the two pilot studies, that the standard test sessions
as previously described [31]. Briefly, the OF was made of wood covered with white would start in the center of the OF and last 15 min, an experiment was performed to
A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288 279

Fig. 2. Behavioral variables representing the general exploration level in the three main areas of the triple test: total locomotion in the OF (A), total number of arm entries in
the EPM (B) and total number of transitions in the LDB (C). Bars and vertical lines (top) represent the means and S.E.M. of the behavioral scores accumulated during 30 min,
for Wistar female rats grouped (n = 8) according to the point where they started the test (OF, EPM or LDB). Symbols and error bars connected by lines (bottom) represent the
means and S.E.M. of the behavioral scores accumulated during each half of the test session. No significant effects of the starting point were found. (***) Represents significant
effects of test duration (0–15 vs. 15–30 min, two-way mixed ANOVA, p < 0.001).

investigate whether rats of different genotypes, known to display consistent genetic 2.3.5. Experiments 6–8
differences in anxiety-related behaviors from each of the individual tests [28,29,31], These three experiments were carried out to assess the effects of three phar-
would show similar differences also in the triple test. To this end, 19 LEW rats (12 macological treatments on the behaviors measured in the new triple test. To this
females and 7 males) and 21 SHR rats (11 females and 10 males) were tested in the end, Wistar male rats were submitted to this test after being injected with either a
integrated apparatus, with those behaviors registered in the previous experiments vehicle or the assigned pharmacological compound. Experiments 6 and 7 included
being measured for the total period of 15 min. Males and females were tested in two benzodiazepine drugs known to have anxiolytic-like effects in the EPM, chlor-
alternate days. diazepoxide and midazolam, respectively [7], whereas experiment 8 assessed the
anxiogenic-like effects of pentylenetetrazole. The same protocol used in exper-
2.3.4. Experiments 4 and 5 iment 5 was used in this series of experiments, excepting that here behavior
In order to get an insight into the different behavioral dimensions assessed in was recorded through the use of two video cameras mounted above the appara-
the triple test, in comparison with the underlying behavioral dimensions of the OF, tus (one for the OF and EPM and one for the LDB) and monitored on a TV set
EPM and LDB applied as separate tests, two experiments were carried out using two by an experimenter located in an adjacent room. Adaptation of this protocol to
independent groups of 60 rats (30/sex). In both cases, behavioral data were submit- automated video-tracking systems needs to be considered in future studies. Test
ted to a principal component analysis, as described below. In experiment 4, each order was totally counterbalanced for treatment condition and the experimenter
animal was submitted to the three separate tests successively, with 1-week interval who scored behavior was not informed about the treatment applied to each ani-
between tests. The apparatuses were set up individually, as already described, each mal.
test session lasted 5 min and all aforementioned variables – plus the number of fecal
boli (defecation) dropped in the OF and the number of squares crossed (locomotion)
in each compartment of the LDB – were registered through direct visual observation. 2.4. Drugs
The same general procedure was used in experiment 5, except for the fact that in the
Chlordiazepoxide (CDZ) (Centaur, India) was suspended in 0.5% methylcellu-
latter each animal was submitted once to a single 15-min session in the triple test,
lose (methocel), as previously described [39], and administered intraperitoneally
starting in the center of the OF. Males and females were tested in alternate days.
(i.p.) (2 ml/kg) 30 min prior to the test, at doses of 0.5, 5 and 10 mg/kg. The control
group received a similar volume of 0.5% methocel. Injection volume and tim-
Table 1 ing were chosen based on two previous rat studies testing the effects of CDZ
Behavioral scores (X ± S.E.M.) of female Wistar rats (n = 24) measured at the three in the EPM and LDB [8,25]. The dose range was designed to provide at least
different areas of the triple test from 0 to 15 and 15 to 30 min of test one dose that was known to be effective in the OF and EPM (5 mg/kg) [8,25,39],
one dose that had shown anxiolytic effects in the LDB (10 mg/kg) [8] plus one
Test area Behavior 0–15 min 15–30 min F(1,21) lower dose (0.5 mg/kg) [10]. Midazolam (MDZ) (Cristalia, Brazil) was dissolved
in saline (NaCl 0.9%) and administered i.p. (1 ml/kg) 30 min prior to the test,
OF Inner locomotion 9.0 ± 2.9 3.6 ± 1.4** 8.88
at doses of 0.1875, 0.375 and 0.750 mg/kg. The control group received a simi-
OF Outer locomotion 91.3 ± 13.5 41.4 ± 8.0*** 30.56
lar volume of saline. Volume, doses and timing of MDZ treatment were based
EPM Closed arm entries 22.0 ± 2.2 13.6 ± 2.0** 13.92
in a previous rat study [2] showing anxiolytic effects of this drug in the EPM,
EPM Open arm entries (%) 6.8 ± 1.1 5.0 ± 1.8 1.47
with the present dose range being slightly narrower than the one used in that
EPM Time in open arms (s) 12.9 ± 3.6 9.6 ± 3.9 1.79
previous report (i.e. 0.187–0.75 instead of 0.125–1.0 mg/kg) due to the smaller num-
LDB Time in light (s) 13.8 ± 3.0 6.0 ± 2.4*** 23.35
ber of treatment groups available in the present study. Pentylenetetrazole (PTZ)
LDB Time in dark (s) 251.4 ± 45.8 289.3 ± 63.4 0.57
(Sigma Aldrich, Germany) was dissolved in saline (NaCl 0.9%) and administered
OF = open field; EPM = elevated plus maze; LDB = light/dark box. (**) and (***) repre- i.p. (2 ml/kg) 5 min before testing, as previously described in three rat studies
sent a significant decrease in behavior from first (0–15 min) to second (15–30 min) [12,25,28], at doses of 10 and 20 mg/kg, with the control group receiving a simi-
half of the test session (p < 0.01 and p < 0.001, respectively; time effect in a two-way lar volume of saline. These doses were shown to be effective in rats tested in the
mixed ANOVA). EPM [12,25].
280 A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288

Table 2
Behavioral scores (X ± S.E.M.) of male Wistar rats (n = 20) measured at the three different areas of the triple test from 0 to 5, 5 to 10 and 10 to 15 min of test

Test area Behavior 0–5 min 5–10 min 10–15 min F(2,36)

OF Inner locomotion 7.6 ± 1.1 8.5 ± 1.1 5.8 ± 1.1 2.11


OF Outer locomotion 57.1 ± 4.0 45.7 ± 2.9** 29.3 ± 2.5*** 33.66
EPM Closed arm entries 10.0 ± 0.7 11.4 ± 0.9 9.2 ± 1.0 2.74
EPM Open arm entries (%) 15.6 ± 2.6 6.9 ± 1.6* 6.4 ± 2.3** 5.15
EPM Time in open arms (s) 14.3 ± 1.8 7.6 ± 1.9* 7.0 ± 2.3* 4.50
LDB Time in light (s) 0.1 ± 0.1 0.5 ± 0.4 0.5 ± 0.3 2.20
LDB Time in dark (s) 13.5 ± 3.1 29.0 ± 4.0** 24.1 ± 4.7* 4.90

OF = open field; EPM = elevated plus maze; LDB = light/dark box. (*), (**) and (***) represent a significant change in behavior in comparison with the first 5 min of the test
session (p < 0.05, p < 0.01 and p < 0.001, respectively; time effect in a two-way mixed ANOVA followed by Duncan test).

2.5. Statistics of the starting point (OF, EPM or LDB) and of interaction between
starting point and testing time (0–15 and 15–30) on any of the
In experiments 1 and 2, two-way mixed ANOVAs for repeated measures were
behavioral variables considered. However, animals starting the test
used to analyze the effects of the starting point (OF, EPM, LDB in experiment 1; inner
vs. outer area of the OF in experiment 2) and the effects of time (0–15 vs. 15–30 min session in the OF tended to explore more all the three apparatuses
in experiment 1; 0–5, 5–10, 10–15 min in experiment 2). Whenever necessary, the than animals starting in the EPM or LDB (see Fig. 2). Actually, 100% of
mean values were compared using the Duncan test. Data from experiment 3 were the animals starting in the OF visited all three apparatuses within
submitted to a two-way ANOVA for strain (LEW and SHR) and sex effects. Duncan the first 15 min of test. Conversely, some animals starting in the
test was used for post hoc comparisons whenever a significant interaction between
strain and sex was detected. Data from experiments 4 and 5 were analyzed using
EPM never visited the OF and took more than 15 min to visit the
Pearson correlation and principal component analysis (PCA), followed by an orthog- LDB and some animals starting in the LDB never visited the OF and
onal Varimax rotation, to investigate the relationship among data obtained from the never re-entered the light compartment of the LDB. Together, these
three tests used separately or in combination. In order to keep a balanced repre- preliminary data suggest that starting the test session inside the OF
sentation of each test, the PCA included four variables from each test. Only factors
area would be the best choice to induce maximum exploration of
with eigenvalues greater than 1 were kept. Finally, a one-way ANOVA for treatment
effect, followed by Duncan test whenever necessary, was used in experiments 6–8. the whole integrated apparatus.
All analyses were performed using the Statistica® 8.0 software package. The same ANOVA revealed a significant decrease from the first
to the second half of the testing session for total OF locomo-
tion (F(1,21) = 32.98, p < 0.001), total EPM arm entries (F(1,21) = 15.35,
3. Results p < 0.001) and total LDB transitions (F(1,21) = 20.75, p < 0.001) (see
Fig. 2). This negative time effect was also found for inner OF loco-
3.1. Experiment 1 motion (F(1,21) = 8.88, p < 0.01), outer OF locomotion (F(1,21) = 30.56,
p < 0.001), EPM closed-arm entries (F(1,21) = 13.92, p < 0.01) and time
The results of the first pilot study are presented in Fig. 2 and in the light compartment of the LDB (F(1,21) = 23.35, p < 0.001) (see
Table 1. The two-way mixed ANOVA revealed no significant effects Table 1). Together, these results suggest that the test session could

Fig. 3. Behavioral variables representing the general exploration level in the three main areas of the triple test: total locomotion in the OF (A), total number of arm entries in
the EPM (B) and total number of transitions in the LDB (C). Bars and vertical lines (top) represent the means and S.E.M. of the behavioral scores accumulated during 15 min,
for Wistar male rats grouped (n = 10) according to the point where they started the test (inner or outer area of the OF). Symbols and error bars connected by lines (bottom)
represent the means and S.E.M. of the behavioral scores accumulated during each 5-min period of the test session. No significant effects of the starting point were found. (*)
and (***) Represent significant effects of test duration (0–5 vs. 5–10 or 10–15 min, two-way mixed ANOVA followed by Duncan test, p < 0.05 and p < 0.001, respectively).
A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288 281

Fig. 4. Behavioral variables representing anxiety- (A-F) and locomotion-related (G-H) behaviors in the three main areas of the triple test: inner locomotion in the OF (A), %
entries (B) and time spent (C) in the open arms of the EPM, time spent in the light compartment (D), number of transitions (E) and time spent in the dark compartment (F)
of the LDB, outer locomotion in the OF (G) and number of closed-arm entries in the EPM (H). Bars and vertical lines represent the means and S.E.M. of the behavioral scores
accumulated during 15 min for animals grouped by strain (LEW and SHR) and sex (F and M) (n = 7–12). (*), (**) and (***) Represent significant differences between strains.
(#), (##) and (###) Represent significant differences between sexes (two-way ANOVA, p < 0.05, p < 0.01 and p < 0.001, respectively, with Duncan post hoc test being used in
G only due to a significant strain × sex interaction).

be limited to the first 15 min if one wants to avoid a drastic drop in ing point (inner vs. outer area of the OF) and of interaction between
the animals’ general exploration level. starting point and testing time (0–5, 5–10 and 10–15) on any of the
behavioral variables considered. Overall, there was a slight trend
3.2. Experiment 2 towards higher exploration levels of all apparatuses in rats starting
the test session in the inner area of the OF (Fig. 3). As one of the two
The results of this experiment are shown in Fig. 3 and Table 2. The OF starting points needed to be chosen, the central square seemed
two-way mixed ANOVA revealed no significant effects of the start- to be the safest choice.
282 A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288

Table 3
Factor analysis (PCA) of 12 behavioral variables measured in 60 rats (30/sex) from a genetically heterogeneous stock tested in the open field (OF), elevated plus maze (EPM)
and light/dark box (LDB) as separate tests

Test/behavior Factor 1 Factor 2 Factor 3 Factor 4 Factor 5

OF
Inner locomotion 0.96
Outer locomotion −0.76
Inner locomotion (%) 0.92
Defecation 0.78

EPM
Time in open arms (s) 0.49 0.72
Open arm entries (%) 0.93
Closed arm entries 0.89
Total arm entries 0.88

LDB
Time in light (s) 0.91
Transitions 0.92
Locomotion in light 0.91
Locomotion in dark 0.71

% of total variance 33.6 15.6 14.4 12.9 8.6

Factor loadings higher than 0.4, produced by a varimax rotation, are shown for each factor.

The same ANOVA showed a significant decrease along the 15 min had lower scores when compared with both SHR males (p < 0.01)
of test for total OF locomotion (F(2,36) = 27.96, p < 0.001) and total and LEW females (p < 0.001). For all the other variables, the two-
EPM arm entries (F(2,36) = 3.29, p < 0.05) (results of the correspond- way ANOVA revealed a significant effect of strain, regardless of
ing post hoc tests are shown in Fig. 3). This negative time effect was sex. SHR rats scored higher than LEW for inner locomotion in the
also found for outer OF locomotion (F(2,36) = 33.66, p < 0.001) and % OF (F(1,37) = 46.42, p < 0.001), % entries (F(1,37) = 5.98, p < 0.05) and
entries (F(2,36) = 5.15, p < 0.05) and time spent (F(2,36) = 4.50, p < 0.05) time spent (F(1,37) = 4.91, p < 0.05) in the open arms of the EPM,
in the open arms of the EPM (results of the corresponding post hoc and time spent in the light (F(1,37) = 10.07, p < 0.01) and number of
tests are shown in Table 2). On the other hand, several other vari- transitions (F(1,37) = 20.05, p < 0.001) in the LDB. LEW rats scored
ables did not show any clear tendency towards a decrease along the higher than SHRs for the time spent in the dark compartment of
15 min of test (Fig. 3 and Table 2). Finally, there was an increase in the LDB (F(1,37) = 24.42, p < 0.001) and number of closed arm entries
the time spent in the dark compartment of the LDB (F(2,36) = 4.90, in the EPM (F(1,37) = 14.24, p < 0.001). Overall sex effects were found
p < 0.05) throughout the test session. Together, these results sug- for inner locomotion (F(1,37) = 8.52, p < 0.01), % entries in the open
gest that limiting the test session to the first 5 or 10 min may not be arms (F(1,37) = 5.86, p < 0.05), time spent in the dark (F(1,37) = 5.11,
advantageous in relation to the proposed 15-min trial if one wants p < 0.05) and number of transitions (F(1,37) = 4.79, p < 0.05). In all
to assess a range of behaviors as wide as possible. cases, females approached more the aversive environments than
males (Fig. 4).
3.3. Experiment 3
3.4. Experiments 4 and 5
The results of experiment 3 are shown in Fig. 4. There was a sig-
nificant strain × sex interaction (F(1,37) = 11.74, p < 0.01) for the outer As shown in Tables 3 and 4, the two PCA performed on a set of
locomotion in the OF, with post hoc tests showing that LEW males 12 behavioral variables (four from each test), representing different

Table 4
Factor analysis (PCA) of 12 behavioral variables measured in 60 rats (30/sex) from a Wistar stock tested in the open field (OF), elevated plus maze (EPM) and light/dark box
(LDB) integrated as parts of the triple test

Test/behavior Factor 1 Factor 2 Factor 3 Factor 4 Factor 5

OF
Inner locomotion 0.93
Outer locomotion 0.95
Inner locomotion (%) 0.96
Defecation

EPM
Time in open arms (s) 0.95
Open arm entries (%) 0.97
Closed arm entries 0.97
Total arm entries 0.87

LDB
Time in light (s) 0.95
Transitions 0.91
Locomotion in light 0.94
Locomotion in dark 0.54 0.49 −0.40

% of total variance 36.3 16.2 14.9 12.5 9.8

Factor loadings higher than 0.4, produced by a varimax rotation, are shown for each factor.
A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288 283

Fig. 5. Correlations among six selected variables from the OF, EPM and LDB tests applied either separately (single) or in combination (triple). Two variables potentially
related with anxiety/emotionality were selected from each test: inner locomotion and defecation in the OF; time spent and % entries in the open arms of the EPM; number
of transitions and locomotion in the light compartment of the LDB. Bars represent the correlation coefficients (Pearson r) for each pair of variables measured under the two
different testing conditions. (*) Represent a significant (p < 0.05) correlation.

facets of the behavioral reactivity of rats submitted to the OF, EPM the two main indices of anxiety of the EPM. Also in both cases, fac-
and LDB applied individually (Table 3) or in combination (Table 4), tor 3 correlated with locomotion-related variables from the EPM
produced very similar factor structures. In both situations, five and LDB, being slightly contaminated by the time spent in the open
factors emerged, representing 85.1 and 89.7% of the total variabil- arms of the EPM under the single-test condition. Factor 5, which
ity, under the single- and triple-test conditions, respectively. The represented less than 10% of the total variability, related to outer
factor loadings for each behavioral measure, which represent the locomotion in the OF and either defecation (single) or locomotion in
correlation between the measure and a specific factor, are shown. the dark (triple condition). This last variable, while relating mostly
To facilitate the interpretation, only the more significant loadings with the locomotion-related factor (factor 3) in the individual tests,
(higher than 0.4) are presented. In both conditions, factor 1 received loaded reasonably evenly over factors 1, 3 and 5 in the triple test.
the most significant contributions from anxiety-related measures Fig. 5 illustrates a comparison between the new triple test and
of the LDB; factor 2 was essentially related to the approach towards the original single tests in terms of the correlation coefficients of a
the central area of the OF; and factor 4 received high loadings from matrix made of six selected variables (two from each test) which
284 A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288

Fig. 6. Anxiety-related behaviors of Wistar male rats treated with either vehicle or different doses of chlordiazepoxide (CDZ at 0.5, 5 or 10 mg/kg, top), midazolam (MDZ
at 0.1875, 0.375 or 0.750 mg/kg, middle) or pentylenetetrazole (PTZ at 10 or 20 mg/kg, bottom). Behavioral variables were measured in a 15-min session in the triple test:
inner locomotion in the OF, % entries and % time spent in the open arms of the EPM, locomotion in the light compartment and number of transitions in the LDB. Bars and
vertical lines represent the means and S.E.M. of animals grouped by pharmacological treatment (n = 13–16). (*), (**) and (***) Represent significant differences between a
given treatment and its respective control group (0 mg/kg) (ANOVA followed by Duncan post hoc test, p < 0.05, p < 0.01 and p < 0.001, respectively).

are potentially related with anxiety/emotionality. Major increases Results from the MDZ experiment (experiment 7) showed
in inter-test correlations, due to the tests being applied combined basically the same profile as the previous experiment, with no
rather than individually, were only observed between defecation significant treatment effects on the OF variables and a significant
in the OF and each of the other variables (negative correlations) drug effect on the % entries (F(3,62) = 4.34, p < 0.01) and % time spent
and between the % entries in the open arms of the EPM and (F(3,62) = 3.73, p < 0.05) in the open arms (Fig. 6) and on the num-
each of the two variables from the LDB (positive correlations). In ber of entries in the closed arms (Fig. 7) (F(3,62) = 4.46, p < 0.01)
all these cases, significance levels were only attained when the of the EPM. Post hoc tests showed that only the highest dose of
tests were applied in their integrated format. Correlations between 0.75 mg/kg affected those behaviors, increasing (p < 0.01) prefer-
inner OF locomotion and anxiety-related measures from the EPM ence for the open arms and decreasing (p < 0.05) closed-arm entries.
were very low in both testing conditions, whereas some improve- No significant treatment effects were found for the LDB variables.
ment (although not reaching significance) could be observed Finally, the analysis of the PTZ data (experiment 8) showed
between the former variable and anxiety-related measures from significant effects of this compound on all behaviors, except
the LDB in the triple test. Finally, the time spent in the open for locomotion in the dark compartment of the LDB. The one-
arms of the EPM was significantly and positively related with both way ANOVA effects were the following: OF inner locomotion
anxiety-related measures from the LDB, regardless of the testing (F(2,35) = 5.11, p < 0.05);% entries (F(2,35) = 9.89, p < 0.001) and % time
condition. spent (F(2,35) = 7.82, p < 0.01) in the open arms of the EPM; locomo-
tion in the light compartment (F(2,35) = 4.02, p < 0.05) and transitions
3.5. Experiments 6–8 (F(2,35) = 3.76, p < 0.05) in the LDB (Fig. 6); OF outer locomotion
(F(2,35) = 20.83, p < 0.001); and EPM closed-arm entries (F(2,35) = 6.91,
Results of experiments 6, 7 and 8 are combined in Fig. 6 p < 0.01) (Fig. 7). Post hoc tests revealed that 10 mg/kg of PTZ caused
(anxiety-related measures) and 7 (locomotion-related measures). significant reductions only on the inner locomotion in the OF
The one-way ANOVAs for the behavioral effects of CDZ (experi- (p < 0.05), % entries (p < 0.01) and % time spent (p < 0.01) in the open
ment 6) revealed no significant effects for the OF variables but a arms of the EPM and locomotion in the light compartment of the
significant drug effect on the % entries (F(3,60) = 5.98, p < 0.01) and LDB (p < 0.05) (see Fig. 6). On the other hand, the highest dose of
% time spent (F(3,60) = 3.03, p < 0.05) in the open arms (Fig. 6) and 20 mg/kg had diminishing effects on all variables (p < 0.05), except
on the number of entries in the closed arms (Fig. 7) (F(3,60) = 5.13, for the locomotion in the dark part of the LDB (Figs. 6 and 7).
p < 0.01) of the EPM. A significant drug effect was also found for the
locomotion in the dark part of the LDB (F(3,60) = 3.16, p < 0.05). Post 4. Discussion
hoc tests comparing treatments with the control group indicated
that the intermediate dose of 5 mg/kg increased (p < 0.05) the first The results of the present study reinforce the assumption that
two variables and had no effects on the third variable. The highest the physical integration of three well-known tests of anxiety – the
dose of 10 mg/kg, however, had significant (p < 0.05) effects on all OF, EPM and LDB – into one combined apparatus can be useful for
three EPM variables, increasing preference for the open arms and the experimental study of anxiety. Rats of both sexes and from dif-
decreasing closed-arm entries. This same dose also decreased the ferent genetic backgrounds were shown to freely explore all areas
locomotion in the dark compartment of the LDB (p < 0.05) (Fig. 7). of this combined test within a timeframe that is compatible with
A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288 285

Fig. 7. Locomotion-related behaviors of Wistar male rats treated with either vehicle or different doses of chlordiazepoxide (CDZ at 0.5, 5 or 10 mg/kg, top), midazolam (MDZ
at 0.1875, 0.375 or 0.750 mg/kg, middle) or pentylenetetrazole (PTZ at 10 or 20 mg/kg, bottom). Behavioral variables were measured in a 15-min session in the triple test:
outer locomotion in the OF, number of entries in the closed arms of the EPM, locomotion in the dark compartment of the LDB. Bars and vertical lines represent the means and
S.E.M. of animals grouped by pharmacological treatment (n = 13–16). (*), (**) and (***) Represent significant differences between a given treatment and its respective control
group (0 mg/kg) (ANOVA followed by Duncan post hoc test, p < 0.05, p < 0.01 and p < 0.001, respectively).

high-throughput behavioral testing. Once the basic protocol was behaviors dropped with time, probably as a result of familiarization,
defined, 97.5% of all the animals submitted to the triple test visited others remained stable or even tended to increase throughout the
all three apparatuses during the regular 15-min trial. Two inbred rat 15 min of test. In summary, the present results, together with pre-
strains, LEW and SHR, known to display high and low anxious-like liminary pharmacological data (not shown), suggested that 15-min
profiles, respectively, in all three individual tests, also differed in a sessions would be adequate for allowing a satisfactory expression
consistent manner for all anxiety-related measures from the triple of all behaviors. Thus, from the two pilot studies on, a protocol
test. Multivariate analyses suggested that the three tests used in of 15 min starting in the center of the OF was chosen for the new
combination assess the same basic behavioral dimensions as those integrated test.
assessed by the individual tests. Finally, pharmacological data pro- The results obtained with the inbred strains LEW and SHR
vided further support to the use of the triple test in the preclinical (experiment 3) corroborated the hypothesis that this new paradigm
investigation of anxiety. is capable of detecting genetic differences in anxiety-related behav-
The first experiment showed that rats explore the three major iors. LEW and SHR rats have been used as a genetic model of anxiety
areas of the triple test regardless of where they were placed in the for more than 10 years based on the fact that they respond dif-
beginning of the test session. However, a non-significant but con- ferently to several anxiety- and stress-related tasks, including the
sistent trend towards higher general exploration levels was found OF, EPM and LDB [3,4,28,29,31]. LEW rats, when compared with
in animals initially placed inside the OF. Such a tendency indicated SHRs, have repeatedly shown higher avoidance of the aversive
that the OF could be considered as a good starting point for this test. compartments of each individual test [28,29,31]. Pharmacological
As far as the test duration is concerned, most standard protocols experiments indicated that such differences were anxiety-related
of the OF, EPM and LDB use sessions of 5 min [9,25,26]. Thus, one [28,39], whereas genetic analyses showed them to result from the
could suppose that a 15-min session should be adequate to evaluate direct effects of genes [29,30]. The present results are in total agree-
animals in an integrated set including the three tests. It was shown ment with the aforementioned previous studies, as LEW rats of both
here, indeed, that a 15-min period was enough for all animals start- sexes, when tested in the triple test, were shown to ambulate less
ing in the OF to visit all three apparatuses. Increasing the time from in the inner OF area, to do less entries and spend less time in the
15 to 30 min, however, produced a drop in most behavioral scores, EPM open arms, to spend less time in the light and more time in
which is suggestive of habituation [1]. The further refinement of the the dark compartments of the LDB and to perform less transitions
new protocol, which was the object of a second study, showed that in the LDB.
starting the test session in either the center or the periphery of the As far as locomotion measures are concerned, the present results
OF did not cause any changes in behavior, although a slight trend are not in total agreement with previous studies, which have shown
towards higher general exploration was seen in animals starting in LEW and SHR rats not to differ in general locomotion in the OF
the inner OF area. Reducing even further the test sessions to either [28,29,31,39]. Herein, SHR males displayed more peripheral loco-
10 or 5 min did not bring a real benefit to the test. Whereas some motion in the OF than LEW males. In the EPM, on the other hand,
286 A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288

LEW rats of both sexes did more closed-arm entries than SHRs. the correlation between emotionality measures from different tests
This is in agreement with some previous studies [20,39] but not varies from low (r < 0.25) to moderate (r < 0.5) and that testing ani-
with others [28,29,31], which showed no strain differences for this mals in an integrated apparatus may improve such a correlation in
measure. It is important to recognize that the new test is not only some cases.
a sum of the individual tests, because in each separate appara- Finally, a preliminary pharmacological investigation was carried
tus an animal would not have the choice to exit the test area, out using three different drugs with known effects on anxiety-
whereas here it can choose to be anywhere at any point in time. related behaviors. The benzodiazepine chlordiazepoxide (CDZ), at
This fact, added to other unavoidable design and protocol differ- 5 mg/kg, produced anxiolytic-like effects on two anxiety measures
ences, might account for any inconsistency between the triple test from the EPM (% entries and % time spent in the open arms), thus
and the regular single tests. However, the present strain results confirming previous studies [8,25,39], without showing any influ-
strongly suggest that many similarities exist between these two ences on other variables, either locomotion- or anxiety-related. At
testing conditions. 10 mg/kg, CDZ produced the same anxiolytic-like effects but also
In order to further investigate the degree of similarity between seemed to provoke sedation, as shown by the reduced number of
the triple and single test conditions, principal component analy- closed-arm entries in the EPM and of squares crossed in the dark
ses (PCA) were performed for each situation. The results showed a compartment of the LDB. The effects of midazolam (MDZ) were very
striking similarity between the five-factor solutions found in both similar to those of CDZ, except that only the highest dose of MDZ
experimental conditions. The LDB, OF and EPM, applied either sepa- (0.75 mg/kg) had significant anxiolytic-like effects and these were
rately or in combination, produced three independent, test-specific not devoid of sedative influences in the EPM. Such anxiolytic and
anxiety factors (factors 1, 2 and 4, respectively), if one considers sedative effects agree with a previous study using similar doses
that the variables loading on these factors have been validated as of MDZ in the regular EPM [2]. However, based on that previous
indices of anxiety [9,10,25,26]. Factor 3, on the other hand, could be study, an anxiolytic-like effect was expected to be seen also with the
seen as a locomotion-related factor, for it related with measures of present intermediate dose (0.375 mg/kg), but this was not observed
locomotion in two protected areas, the closed arms of the EPM [12] herein. In summary, the present results are consistent with the
and the dark part of the LDB. Outer OF locomotion loaded heavily known anxiolytic effects of CDZ and MDZ and the possible sedative
on Factor 5, together with either defecation or locomotion in the effects of CDZ in the common EPM [7,12,25,39]. Nevertheless, these
dark, suggesting that this factor is at least partially related with anti-anxiety agents did not have the expected effects on behav-
locomotion. These results are in agreement with previous factor iors measured in the OF and LDB areas. Two alternative hypotheses
studies that showed that different anxiety tests seem to reflect dif- could be put forward to explain this lack of effect: (1) in rats, the
ferent facets of emotionality [16,29,38]. The present PCA analyses OF and LDB, even individually, are not consistently sensitive to ben-
also suggest that the triple test preserves important features of its zodiazepines; (2) these tests, individually, respond consistently to
individual components and therefore it can be used to assess differ- benzodiazepines, but when physically combined they change their
ent types of emotional states as well as locomotor activity in novel nature and loose such sensitivity.
but not highly aversive areas. The main difference between the fac- In the single OF test, anxiolytic effects of acute CDZ and MDZ
tors obtained under the two testing conditions was related to the treatments were found in about only half of the rat studies reviewed
locomotion in the dark part of the LDB, which loaded only on factor by Prut and Belzung [26], which brings some support to the first
3 in the individual tests, but on factors 1, 3 and 5 in the triple test. aforementioned hypothesis. Such an inconsistency, however, might
Factor 1 reflected mostly the approach towards the LDB light com- be partially explained by the huge inter-laboratory variation found
partment. Thus, the relatively high loading (0.54) of locomotion in in OF protocols [36]. Yet, data from our own laboratory [39] have
the dark compartment on this factor means that animals exploring shown that CDZ at 5 mg/kg can have anxiolytic-like effects in rats
more the light also explored more the dark LDB areas. This rela- tested in the single OF test, depending on the genotype and gender
tionship makes sense in the triple test only, where there is a choice of the animals. As far as the LDB is concerned, CDZ and other benzo-
between the LDB and the other test areas and a higher preference diazepines have been shown to increase exploration of mice in the
for the LDB, as a whole, might cause an increase in all LDB variables. aversive light compartment [9,10], an effect also produced by sero-
The correlation analyses further clarified the relationships tonin reuptake inhibitors [35]. In rats, however, CDZ at 5 mg/kg,
among variables from the three different tests applied either a treatment that reduced anxiety in the EPM, did not change the
individually or in combination. Fig. 5 shows the correlation coef- approach towards the white area of the regular LDB test [8], which
ficients among six measures (two from each test) selected for is in total agreement with the present results of the triple test.
being anxiety- or emotionality-related [9,10,18,25,26]. Because in Likewise, at 10 mg/kg, CDZ was shown to reduce the latency to
the triple apparatus an animal can express its emotional state in all enter the light area of the LDB but it did not change other clas-
tests within the timeframe of a single trial, one could expect the sical indices of anxiety from this test [8], which also agrees with
correlations to increase in comparison with the single-test condi- the present results. Citalopram, which was shown to be anxiolytic
tion. Such an increase, however, was only observed in some pairs in mice [35], had anxiogenic-like effects in the rat LDB [34] and,
of variables, such as OF defecation in relation to all the others vari- in another comprehensive study, rats receiving a wide range of
ables and % of entries in the open arms of the EPM in relation to anxiolytic drugs (including benzodiazepine and serotonin-related
the two variables from the LDB. Defecation was the first measure compounds) showed no signs of anxiolysis in a LDB [5]. Together,
of emotionality proposed by Hall [18], but it is rarely considered these results support the aforementioned hypothesis 1 and suggest
as an index of anxiety in pharmacological studies [26,27]. Yet, the that the LDB may not be as efficient for anxiolytic drug screening
present results suggested a mild but significant negative correla- in rats as it is in mice.
tion between OF defecation and classical anxiety measures from The aforementioned issues need to be addressed in future
the EPM and LDB. These, in turn, seem to be more strongly related to studies, by testing rats of the same strain and under the same phar-
each other, which is compatible with their extensive use as anxiety macological treatments used herein, in each of the individual tests,
models [13,17,27], than with the most popular measure of anxiety specially the OF and LDB. If hypothesis 1 is confirmed, this may
from the OF. Actually, the central OF locomotion was not signif- bring some concerns about the meaning of the regular individual
icantly correlated with any of the other variables, regardless of tests for rats. On the other hand, if hypothesis 2 is corroborated,
the experimental condition. Overall, the present data suggest that future studies will have to investigate the potential usefulness of
A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288 287

the triple test by verifying, for example, if its OF and LDB areas [6] Blizard DA, Takahashi A, Galsworthy MI, Martin B, Koide T. Test standardiza-
are sensitive to other classes of anxiolytic drugs, such as buspirone tion in behavioural neuroscience: a response to Stanford. J Psychopharmacol
2007;21:136–9.
and selective serotonin reuptake inhibitors, which tend to produce [7] Carobrez AP, Bertoglio LJ. Ethological and temporal analyses of anxiety-like
inconsistent results when tested in each of the individual compo- behavior: the elevated plus-maze model 20 years. Neurosci Biobehav Rev
nents of the triple test [5,7,26]. Moreover, the effects of different 2005;29:1193–205.
[8] Chaouloff F, Durand M, Mormède P. Anxiety-and activity-related effects of
treatment schedules (i.e. acute vs. chronic) will have to be consid- diazepam and chlordiazepoxide in the rat light/dark and dark/light tests. Behav
ered. Brain Res 1997;85:27–35.
The effects of pentylenetetrazole (PTZ), on the other hand, were [9] Costall B, Jones BJ, Kelly ME, Naylor RJ, Tomkins DM. Exploration of mice in a
black and white test box: validation as a model of anxiety. Pharmacol Biochem
clear-cut and test-independent. At 10 mg/kg, PTZ had anxiogenic- Behav 1989;32:777–85.
like effects in all aversive areas of the triple test, without showing [10] Crawley JN. Neuropharmacologic specificity of a simple animal model
any influences on measures of protected locomotion. It signifi- for the behavioral actions of benzodiazepines. Pharmacol Biochem Behav
1981;15:695–9.
cantly reduced inner locomotion in the OF, % entries and % time
[11] Crawley JN, Paylor R. A proposed test battery and constellations of specific
spent in the open arms of the EPM and locomotion in the light behavioral paradigms to investigate the behavioral phenotypes of transgenic
compartment of the LDB, but it did not affect outer locomotion, and knockout mice. Horm Behav 1997;31:197–211.
closed-arm entries and locomotion in the dark compartment of the [12] Cruz APM, Frei F, Graeff FG. Ethopharmacological analysis of rat behavior on
the elevated plus-maze. Pharmacol Biochem Behav 1994;49:171–6.
three respective tests. At 20 mg/kg, however, PTZ reduced all vari- [13] Cryan JF, Holmes A. The ascent of mouse: advances in modeling human depres-
ables (except for locomotion in the dark), which is in agreement sion and anxiety. Nat Rev Drug Discov 2005;4:775–90.
with its known anxiogenic and hypolocomotor effects [12,25,28]. [14] Ennaceur A, Michalikova S, Chazot Pl. Models of anxiety: responses of
rats to novelty in an open space and an enclosed space. Behav Brain Res
The present pharmacological results demonstrate that the triple 2006;171:26–49.
test provides anxiety-related indices, which respond to anxliolytic [15] Ennaceur A, Michalikova S, Van Rensburg R, Chazot Pl. Models of anxiety:
and/or anxiogenic treatments (e.g. CDZ 5 mg/kg and PTZ 10 mg/kg), responses of mice to novelty and open spaces in a 3D maze. Behav Brain Res
2006;174:9–38.
as well as locomotion-related indices, which are not influenced by [16] File SE. The biological basis of anxiety. In: Meltzer HY, Nerozzi D, editors. Cur-
these treatments. However, these data should be considered with rent practices and future developments in the pharmacotherapy of mental
caution because they are quite preliminary and narrow in terms of disorders. Amsterdam: Elsevier; 1991. p. 159–65.
[17] Finn DA, Rutledge-Gorman MT, Crabbe JC. Genetic animal models of anxiety.
tested compounds. Neurogenetics 2003;4:109–35.
In summary, the present study demonstrates that the inte- [18] Hall CS. Emotional behavior in the rat. Defecation and urination as measures
gration of three widely used tests of anxiety, the OF, EPM and of individual differences in the emotionality. J Comp Psychol 1934;18:385–
403.
LDB, into one single experimental set can be useful for measur-
[19] Izidio GS, Lopes DM, Spricigo Jr L, Ramos A. Common variations in the pretest
ing, in a relatively simple way, a wide range of unconditioned environment influence genotypic comparisons in models of anxiety. Genes
exploratory behaviors in just one trial, minimizing common con- Brain Behav 2005;4:412–9.
cerns about previous test experiences, inter-test intervals and [20] Izidio GS, Spricigo Jr L, Ramos A. Genetic differences in the elevated plus-maze
persist after first exposure of inbred rats to the test apparatus. Behav Processes
handling between tests. The behaviors measured were shown to 2005;68:129–34.
reflect distinct emotional and motor dimensions, in line with the [21] Kalueff AV, Wheaton M, Murphy Dl. What’s wrong with my mouse model?
recently proposed notion of “hybrid models”, which are expected Advances and strategies in animal modeling of anxiety and depression. Behav
Brain Res 2007;179:1–18.
to assess different behavioral domains simultaneously [21]. The [22] Lipkind D, Sakov A, Kafkafi N, Elmer GI, Benjamini Y, Golani I. New replicable
present results also point to this new integrated paradigm as a anxiety-related measures of wall vs. center behavior of mice in the open field.
promising tool for genetic and pharmacological screening, as it J Appl Physiol 2004;97:347–59.
[23] Lister RG. Ethologically-based animal models of anxiety disorders. Pharmacol
seems to provide a more informative snapshot of an animal’s emo- Ther 1990;46:321–40.
tional status in comparison with each of the original tests. Further [24] Paylor R, Spencer CM, Yuva-Paylor LA, Pieke-Dahl S. The use of behavioral test
studies involving behavioral, pharmacological and genetic exper- batteries II: effect of test interval. Physiol Behav 2006;87:95–102.
[25] Pellow S, Chopin P, File SE, Briley M. Validation of open: closed arm entries in
iments will be necessary before the usefulness of this model can
an elevated plus-maze as a measure of anxiety in the rat. J Neurosci Methods
be confirmed and generalized to other species, in particular to 1985;14:149–67.
mice. [26] Prut L, Belzung C. The open field as a paradigm to measure the effects of drugs
on anxiety-like behaviors: a review. Eur J Pharmacol 2003;463:3–33.
[27] Ramos A, Mormède P. Stress and emotionality: a multidimensional and genetic
approach. Neurosci Biobehav Rev 1998;22:33–57.
Acknowledgements [28] Ramos A, Berton O, Mormède P, Chaouloff F. A multiple-test study of anxiety-
related behaviours in six inbred rat strains. Behav Brain Res 1997;85:57–
This work was supported by a grant (Ed 612005) from 69.
[29] Ramos A, Mellerin Y, Mormède P, Chaouloff F. A genetic and multifactorial anal-
CNPq/Brazil. A. Ramos was recipient of a fellowship from CNPq and ysis of anxiety-related behaviours in Lewis and SHR intercrosses. Behav Brain
G.S. Izidio and E. Pereira had scholarships from CNPq. The authors Res 1998;96:195–205.
wish to thank Ligia F.G. de Oliveira for her technical assistance. [30] Ramos A, Moisan MP, Chaouloff F, Mormède C, Mormède F. Identification of
female-specific QTL’s affecting an emotionality-related behavior in rats. Mol
Psychiatry 1999;4:453–62.
[31] Ramos A, Kangerski AL, Basso PF, Da Silva Santos JE, Assreuy J, Vendruscolo LF,
References et al. Evaluation of Lewis and SHR rat strains as a genetic model for the study
of anxiety and pain. Behav Brain Res 2002;129:113–23.
[1] Archer J. Tests for emotionality in rats and mice: a review. Anim Behav [32] Ramos A, Correia EC, Izidio GS, Bruske GR. Genetic selection of two new
1973;21:205–35. rat lines displaying different levels of anxiety-related behaviors. Behav Genet
[2] Bertoglio LJ, Carobrez AP. Prior maze experience required to alter midazolam 2003;33:657–68.
effects in rats submitted to the elevated plus-maze. Pharmacol Biochem Behav [33] Rodgers RJ, Cao B-J, Dalvi A, Holmes A. Animal models of anxiety: an ethological
2002;72:449–55. perspective. Braz J Med Biol Res 1997;30:289–304.
[3] Berton O, Ramos A, Chaouloff F, Mormède P. Behavioral reactivity to social and [34] Sánchez C, Méier E. Behavioral profiles of SSRIs in animal models of
nonsocial stimulations: a multivariate analysis of six inbred rat strains. Behav depression, anxiety and aggression. Are they all alike? Psychopharmacology
Genet 1997;27:155–66. 1997;129:197–205.
[4] Berton O, Aguerre S, Sarrieau A, Mormede P, Chaouloff F. Differential effects of [35] Sánchez C, Bergqvist PBF, Brennum LT, Gupta S, Hogg S, Larsen A, et al. Esci-
social stress on central serotonergic activity and emotional reactivity in Lewis talopram, the S-(+)-enantiomer of citalopram, is a selective serotonin reuptake
and spontaneously hypertensive rats. Neuroscience 1998;82:147–59. inhibitor with potent effects in animal models predictive of antidepressant and
[5] Bilkei-Gorzó A, Gyertyán I, Lévay G. mCPP-induced anxiety in the light-dark box anxiolytic activities. Psychopharmacology 2003;167:353–62.
in rats—a new method for screening anxiolytic activity. Psychopharmacology [36] Stanford SC. The open field test: reinventing the wheel. J Psychopharmacol
1998;136:291–8. 2007;21:134–5.
288 A. Ramos et al. / Behavioural Brain Research 193 (2008) 277–288

[37] Treit D. Animal models for the study of anti-anxiety agents: a review. Neurosci [40] Vitaterna MH, Pinto LH, Takahashi JS. Large-scale mutagenesis and pheno-
Biobehav Rev 1985;9:203–22. typic screens for the nervous system and behavior in mice. Trends Neurosci
[38] Trullas RE, Skolnick P. Differences in fear motivated behaviors among inbred 2006;29:233–40.
mouse strains. Psychopharmacology 1993;44:463–9. [41] Võikar V, Vasar E, Rauvala H. Behavioral alterations induced by repeated testing
[39] Vendruscolo LF, Takahashi RN, Bruske GR, Ramos A. Evaluation of the anxiolytic- in C57BL/6J and 129S2/SV mice: implications for phenotyping screens. Genes
like effect of NKP608, a NK1-receptor antagonist, in two rat strains that differ Brain Behav 2004;3:27–38.
in anxiety-related behaviors. Psychopharmacology 2003;170:287–93.

You might also like