NAVIGATION IN THE MORRIS SWIM TASK AS A BASELINE FOR DRUG DISCRIMINATION: A DEMONSTRATION WITH MORPHINE

JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR
2002, 78, 215–223
NUMBER 2 (SEPTEMBER)
NAVIGATION IN THE MORRIS SWIM TASK
AS A BASELINE FOR DRUG DISCRIMINATION:
A DEMONSTRATION WITH MORPHINE
DAVID ZIEGLER, JULIAN R. KEITH,
RAYMOND C. PITTS, AND MARK GALIZIO
UNIVERSIT Y OF NORTH CAROLINA AT WILMINGTON
A morphine versus saline discrimination was demonstrated using the Morris swim task as the behav-
ioral baseline. The apparatus was a large circular pool ?lled with water made opaque by ?oating
polypropylene pellets. Rats were placed in the tank in randomly selected locations (12 trials per
session) and could escape by swimming to a platform submerged 2 cm below the surface. Morphine
(5.6 mg/kg) or saline was injected prior to training sessions. The position of the platform in a given
session depended on the drug condition, thus forming the basis for discriminative responding. Three
of the 4 rats acquired the discrimination, as evidenced by direct swims to the condition-appropriate
platform. Generalization probe sessions were conducted following acquisition. Probe sessions were
preceded by injections of morphine (0, 1.0, 3.0, 5.6, or 10.0 mg/kg) and involved placing the rat in
the pool for 1 min without a platform. Swim patterns revealed a gradient, with probe swimming
more concentrated in the area of the morphine platform position after higher morphine doses. In
addition, dose-dependent increases in the likelihood of swimming ?rst to the morphine-associated
platform location were obtained. These results illustrate the generality of drug discrimination across
different behavioral procedures, and of particular interest with respect to spatial learning, demon-
strate interoceptive stimulus control of navigation.
Key words: drug discrimination, Morris water maze, spatial learning, morphine, rat
Drug-discrimination
procedures
are
sioned after vehicle, but not drug, adminis-
among the most important and widely used
tration. Further evidence of stimulus control
preparations in behavioral pharmacology be-
comes from generalization tests in which
cause of their value in classifying and char-
drug doses higher or lower than that used in
acterizing drugs (see reviews by Riley, 1997;
training are substituted for the training dose,
Stolerman, 1993; Young, 1991a). Contempo-
and are associated with a dose-dependent gra-
rary applications of the procedure generally
dient of drug-lever responding. Drugs of the
involve a conditional discrimination in which
same pharmacological class as the training
a speci?c response (say, pressing the left lever
drug also frequently occasion drug-lever re-
in a two-lever operant chamber) is reinforced
sponding, whereas drugs from different clas-
after administration of a dose of a given drug.
ses typically occasion vehicle-lever responding
During another session, pressing the right le-
(Riley; Stolerman; Young).
ver is reinforced after a vehicle administra-
The results of drug-discrimination studies
tion. With such training, lever pressing comes
have shown considerable generality across
under the control of the drug in that left- or
various methods. For example, similar drug-
drug-lever responding is occasioned after
discrimination patterns have been obtained
drug, but not vehicle, administration, and
with a variety of different species and proce-
right- or vehicle-lever responding is occa-
dures. In addition to the lever-press arrange-
ment with food reinforcement described
This research formed part of a master’s thesis com-
above, drug discrimination has been estab-
pleted by the ?rst author. We thank James Hummel for
lished with other maintaining events (e.g.,
his help with behavioral testing, and the National Insti-
shock avoidance; Shannon & Holtzman,
tute on Drug Abuse for the generous gift of morphine.
1976) and with other apparatus (e.g., t maze;
The procedures were approved by the University Animal
Use and Care Committee. This research was supported
Overton, 1982). The present study sought to
in part by a grant from the National Institute on Drug
determine whether the generality of drug dis-
Abuse (DA12879).
crimination could be extended to an escape
Requests for reprints should be sent to Mark Galizio,
procedure involving spatial navigation: the
Department of Psychology, University of North Carolina
at Wilmington, Wilmington, North Carolina 28403 (e-
Morris swim task (Morris, 1981).
mail: galizio@uncwil.edu).
The Morris swim task involves placing the
215

---

216
DAVID ZIEGLER et al.
rat in a large pool with a small escape plat-
to exert rapid stimulus control (e.g., Gallistel,
form slightly submerged beneath water made
1990; O’Keefe & Nadel, 1978). From this
opaque so that the platform is not visible.
standpoint, learning relations between inter-
Morris (1981) showed that rats rapidly learn
oceptive cues and spatial locations might be
to navigate directly to the platform from any
seen as contraprepared, and thus, dif?cult or
release spot and, thus, the procedure has be-
impossible to learn. The present study ex-
come a widely used tool in the study of spatial
amined the possibility of establishing intero-
learning (Cain & Saucier, 1996; Gallistel,
ceptive control over navigation by training
1990). Accounts of spatial learning in the
rats to swim to a hidden platform that was
swim task differ, but most emphasize the con-
located in one position on sessions following
trol of navigation by stimuli located outside
injections of 5.6 mg/kg morphine and in an-
the pool (e.g., Morris). However, analysis of
other location following saline injections.
stimulus control in the swim task has been
limited, and it remains dif?cult to specify with
METHOD
precision the nature of the stimuli that con-
trol navigation (Cain, Beiko, & Boon, 1997;
Subjects
Whishaw & Mittlelman, 1986).
Four albino Holtzman Sprague-Dawley rats,
Nonetheless, Keith and Galizio (1997)
120 days old at the onset of testing, served as
brought navigation itself under stimulus con-
subjects. Their coats were marked with harm-
trol. They demonstrated conditional control
less black ink to enhance video tracking.
of navigation by visual stimuli in the swim task
by training rats in two physically separate
Apparatus
pools: one painted black and the other paint-
The apparatus was a circular white pool
ed white. A multiple-component repeated-ac-
(1.5 m diameter, 30.5 cm deep). The pool
quisition procedure was used in which the
was ?lled to a level of 23 cm with water, and
hidden platform was always in a ?xed loca-
a layer of polypropylene pellets ?oated on the
tion in one pool (performance component)
surface of the water to prevent rats from see-
and varied from session to session in the oth-
ing the submerged escape platform (see Cain
er (acquisition component). After 10 to 20
et al., 1997). A small white platform (10 cm
training sessions, rats showed evidence of
diameter, 20 cm high) was submerged 2 cm
conditional control. In the performance-com-
below the surface of the water in a location
ponent pool, rats swam rapidly and directly
determined by experimental condition (see
to the platform. In the acquisition-compo-
below). The water temperature was main-
nent pool, latencies were long and swim
tained at approximately 30 C. The swimming
paths erratic on the ?rst trial of a session, but
pool was illuminated using indirect halogen
after just one or two trials, rapid direct swims
lighting (500 W) re?ected off the white paint-
to the new platform position were observed.
ed ceiling. Black-and-white vinyl shower cur-
These results showed conditional control of
tains, marked with large distinctive patterns,
navigation by visual stimuli, and it seemed
completely enclosed the water tank in a cir-
plausible to try to adapt the procedure to
cular fashion (see Figure 1).
drug discrimination. Indeed, the rapid learn-
A closed circuit camera centered over the
ing seen in the swim task suggests the predic-
tank was aimed directly downward, capturing
tion that drug discrimination might be read-
the entire surface of the water. The video sig-
ily acquired with this procedure. Consider
nal was transduced by an interface made to
that Riley (1997) has shown that drug-dis-
track a contrasting object (a rat) in the visual
crimination training can be accelerated using
?eld of the camera. Software (San Diego In-
the conditioned taste aversion procedure,
struments) permitted the evaluation of sev-
which, like spatial navigation, is characterized
eral dependent measures including latency to
by rapid learning. On the other hand, the
reach the platform, path length, initial head-
evolutionary history that shaped spatial learn-
ing, and overall swim speed.
ing may suggest a different prediction. Some
classic treatments view spatial learning as in-
Procedure
volving a preparedness to develop a ‘‘map’’
Pretraining. During two preliminary ses-
that permits landmarks in the environment
sions, rats were trained to swim to a ?xed lo-

---

SWIM TASK AND DRUG DISCRIMINATION
217
although no injections were given during the
two pretraining sessions, the location of the
platform was the same as that later used for
the saline condition.
Discrimination training. General procedures
were as described above, but after pretraining
rats received intraperitoneal injections of 5.6
mg/kg morphine or saline 15 min prior to
each training session. On sessions that fol-
lowed saline injections, the platform was al-
ways placed in the east quadrant, 33 cm to-
wards the center of the tank. After morphine
injections, the platform was located in the
west quadrant, again 33 cm towards the cen-
ter of the tank (see Figure 1). For analysis
purposes, a correct (condition-appropriate)
response was de?ned by two criteria. The ?rst
criterion was that the rat swim to the correct
platform without breaking the plane of a cir-
Fig. 1.
The pool (1.5 m diameter; not drawn to scale)
cle (20 cm diameter) centered on the con-
is depicted. Within the pool, small gray circles indicate
dition-inappropriate platform (see Figure 1).
drop-off points, and larger dark circles the location of
The second criterion for a condition-appro-
the platforms (M
morphine; S
saline). Surrounding
priate response was that the latency to reach
the platforms are two rings generated by the data analysis
software (not visible to the rats). The inner ring (20 cm
the escape platform was less than or equal to
diameter) was used to determine criterion responding in
10 s. Training criteria required that 90% or
training, and the outer ring (45 cm diameter) was used
better condition-appropriate responding be
to determine dwell time during probe trials. Outside this
achieved across at least eight of ten sessions
graphic representation of the apparatus, the visual pat-
terns displayed on a curtain that encircled the apparatus
and also that in eight of those ten sessions
are represented by squares that indicate the pattern dis-
the ?rst trial was correct. When these criteria
played on the entire quadrant of the curtain (north:
were met, probe testing sessions replaced the
large white cross on black; east: vertical stripes; south:
regular training session on two of the ?ve
solid black; west: solid white).
weekly sessions.
Generalization probes. As with training ses-
cation in the pool. This phase of training in-
sions, probe sessions were preceded by injec-
volved placing the rat in the water until it
tions, but to assess generalization, the range
reached the platform or until 1 min elapsed.
of doses included 1.0, 3.0, and 10.0 mg/kg
At the onset of each trial, the rat was placed
morphine in addition to 5.6 mg/kg and sa-
in the tank, facing the outside, at any of six
line. The rat was placed in the pool 15 min
release locations (see Figure 1). After reach-
after injection on probe sessions, but no plat-
ing the platform, rats were permitted to re-
form was present. The swim path was ob-
main on it for 15 s, and then were removed
served for 60 s in the absence of the platform,
from the pool area for a 2-min intertrial in-
and the session was then terminated. For pur-
terval. If the rat failed to reach the platform
poses of generalization data analysis, the rat’s
within 1 min, it was placed on the platform
swim path during the 60 s was divided into
by the experimenter. Twelve trials were con-
time spent within a circle (45 cm diameter)
ducted in each session, with each release lo-
around the saline platform location (saline
cation used twice per session. Sessions were
dwell time), a circle (45 cm diameter) around
conducted 5 days per week. Release locations
the morphine platform (morphine dwell
were randomly determined, with the con-
time), and time spent outside both circles.
straints that no more than three northerly or
Each probe performance was also analyzed
three southerly release points were used con-
with a variation of the criteria used to deter-
secutively, and the same release point was not
mine correct responding in baseline (i.e., if
used twice in a row. The location of the plat-
the swim path intercepted the position of the
form was ?xed during any given session, and
morphine platform ?rst, it was characterized

---

218
DAVID ZIEGLER et al.
Fig. 2.
Acquisition of the morphine discrimination for each rat. Percentage of correct trials per session are in-
dicated by either circles (saline training sessions) or triangles (morphine training sessions). Filled symbols indicate
that performance on the ?rst trial of that session was correct, and open symbols indicate incorrect initial trials. The
dotted line represents the criterion cut-off for 10 of 12 correct responses.
as a morphine response; a swim path that in-
tion probes was determined randomly, with
tercepted the saline platform position ?rst
the constraint that the end of each cycle of
was classi?ed as a saline response).
determinations (one exposure to each dose
Even after the probe sessions had begun,
including a saline injection) was completed
training sessions were conducted on three of
before beginning the next cycle.
the ?ve weekly sessions, and probe sessions
were conducted only when the criteria for
maintenance of discrimination were met. For
RESULTS
a probe test to be conducted, responses on 4
Acquisition of discrimination. Extensive train-
of the last 5 training days had to be 90% con-
ing was required to establish the morphine–
dition appropriate. Furthermore, on the ses-
saline discrimination. In fact, only 3 of the 4
sion immediately preceding the planned
rats met acquisition criteria. Figure 2 shows
probe session, the ?rst response had to be
the acquisition data for each rat with per-
correct, with at least 90% condition-appro-
centage of correct swims per session plotted
priate responding. Each dose was tested three
on the vertical axis (recall that a correct swim
times. The order of doses during generaliza-
was de?ned as one in which the rat reached

---

SWIM TASK AND DRUG DISCRIMINATION
219
the platform in less than 10 s without enter-
of morphine during the ?rst cycle of testing.
ing the 20-cm circle that de?ned the condi-
The gradient obtained from Rat Y1 was a
tion-inappropriate zone). For 3 of the rats
somewhat different shape. The gradient was
(Rats Y1, Y2, and Z1) accuracy increased
?at up through 3.0 mg/kg, with higher levels
across sessions. Rat Y1 met the acquisition cri-
of morphine responding occurring at the
teria in 64 sessions, Y2 required 42 sessions,
training dose and above. In general, gradi-
and Z1 required 49 sessions. Although Rat Z2
ents obtained during the second cycle of de-
showed improvement across sessions, its per-
terminations were similar in form to those
formance never reached criterion levels, and
obtained in the ?rst. One notable exception
it was dropped from the study after 70 ses-
was for Rat Y1, whose gradient was ?at until
sions.
the highest morphine dose was tested. The
Generalization tests. The left panels of Figure
third cycle of doses for Rats Y1 and Y2 (right
3 show percentage of morphine responding
panels) resulted in gradients that were less
(morphine area dwell time divided by mor-
clearly under morphine stimulus control. Al-
phine plus saline area dwell times multiplied
though the peak of Rat Y1’s gradient was at
by 100%) averaged across the three deter-
5.6 mg/kg, morphine responding was rough-
minations and plotted as a function of mor-
ly equivalent after saline and 10 mg/kg. Rat
phine dose. Thus, scores higher than 50% on
Y2 showed twin peaks at 3.0 and 10.0 mg/kg,
this measure re?ect more time spent swim-
but responding at 5.6 and 1.0 mg/kg and sa-
ming in the morphine area relative to time
line was equivalent. However, stimulus con-
spent in the saline area (with 100% indicating
trol remained sharp for Rat Z1, which showed
that the animal failed to enter the saline area
continued high levels of morphine respond-
during the probe session). The generaliza-
ing for the three highest doses tested. In sum-
tion gradients showed a dose-dependent in-
mary, there was some evidence of a ?attening
crease in morphine responding for all 3 rats.
of generalization gradients by the third cycle
Rat Y1’s average morphine area responding
of determinations for 2 rats (Y1 and Y2).
was approximately 50% after saline injections
The dwell-time measures presented in Fig-
as well as after 1.0 mg/kg of morphine, with
ure 3 re?ect the swimming patterns of sub-
gradual increases in morphine dwell time at
jects for the full 60-s probe. Such extended
the higher doses. Steeper gradients were ob-
exposure to extinction conditions might have
tained from Rats Y2 and Z1. Rat Y2 (middle
interfered with stimulus control, particularly
left panel) spent relatively little time in the
late in the probe trial. An alternative measure
morphine area after saline and 1.0 mg/kg in-
that is less in?uenced by extinction condi-
jections, but relatively more time there at all
tions is the initial platform area choice on
higher doses. Similar results were apparent
each probe trial. Using the criteria that were
for Rat Z1 (bottom left panel). The absolute
operative in determining correct responding
dwell-time data presented in Table 1 show
in training (Figure 1), each probe trial was
that the percentages shown in Figure 3 re-
analyzed in terms of which area was ?rst en-
?ected dose-related increases in time spent in
tered. Figure 4 shows the percentage of initial
the morphine area and decreases in time
swims to the morphine area. All 3 rats swam
spent in the saline area. For example, rats
to the saline area ?rst during each saline test
spent from 4.2 to 8.5 s in the morphine area
session and to the morphine area ?rst on all
after saline and 11.5 to 15.1 s after 5.6 mg/
three 10 mg/kg test sessions. Responding to
kg morphine. In summary, all 3 rats provided
intermediate doses revealed a fairly steep gra-
evidence of generalization gradients along
dient, indicating that stimulus control in the
the dimension of morphine dose.
initial portion of the probes remained fairly
The right panels of Figure 3 show the mor-
strong throughout the three cycles of probe
phine area dwell times for each individual
determinations.
probe session. In the ?rst cycle, as morphine
Table 2 shows swimming speeds as function
dose was increased, the relative time spent in
of morphine dose, and the absence of reli-
the morphine area increased in all 3 rats. Rats
able effects indicates that the gradients ob-
Y2 and Z1 each responded approximately
served in Figures 3 and 4 were obtained at
30% to 35% less in the morphine area at sa-
doses that did not produce impairments in
line and 1.0 mg/kg than at the higher doses
overall swimming performance.

---

220
DAVID ZIEGLER et al.
Fig. 3.
Left panels show mean percentage of morphine dwell time (number of seconds spent in morphine target
area divided by number of seconds spent in saline plus morphine target areas multiplied by 100%) during probe
sessions as a function of morphine dose for the 3 rats tested. Right panels show percentage of morphine dwell time
during individual probe sessions (Test 1
black circles; Test 2
white triangles; Test 3
black squares) for each
rat.

---

SWIM TASK AND DRUG DISCRIMINATION
221
Table 1
Mean absolute dwell time (s) and standard deviation (in parentheses) in saline and morphine
areas as a function of morphine dose during generalization probe tests.
Rat Y1
Rat Y2
Rat Z1
Probe
dose
Saline
Morphine
Saline
Morphine
Saline
Morphine
Saline
9.3 (1.8)
8.5 (0.3)
10.5 (1.4)
5.1 (0.6)
9.8 (1.0)
4.2 (0.8)
1.0
8.2 (0.7)
7.1 (1.5)
9.7 (1.7)
4.6 (0.5)
9.8 (0.5)
5.4 (2.4)
3.0
8.0 (1.5)
9.9 (0.9)
7.5 (0.3)
12.6 (0.2)
6.4 (0.8)
12.0 (1.3)
5.6
4.9 (0.9)
11.5 (3.7)
8.1 (1.9)
11.3 (2.7)
4.7 (1.2)
15.1 (1.4)
10.0
5.3 (1.0)
10.5 (1.7)
7.4 (1.8)
14.3 (2.1)
4.6 (0.9)
13.5 (2.4)
DISCUSSION
man, 1976, 1979; Young et al., 1992). The
One major question posed in this study was
gradients based on percentage of morphine-
whether a morphine versus saline discrimi-
area dwell time across sessions (Figure 3)
nation could be learned using the Morris
ranged from approximately 30% at saline and
swim task. With 3 of the 4 rats tested, the re-
low doses of morphine to 75% for 5.6 and 10
sults were af?rmative. Although there was
mg/kg. The shapes of these gradients varied
some evidence of acquisition with the 4th rat,
somewhat within rats, but for 2 rats (Y2 and
it failed to meet the discrimination criteria
Z1), gradients were sigmoidal with low and
after 70 sessions. For the 3 rats that met the
roughly equivalent morphine dwell time after
discrimination criterion, acquisition was not
saline and 1.0 mg/kg and high and roughly
particularly rapid, with a range of 42 to 64
equivalent morphine dwell time at all other
sessions to criterion required. In fact, these
doses. The gradient for Y1 was ?atter, but
rates of acquisition were similar to those ob-
stimulus control was still revealed by the high-
served with rats in other studies using two-
er values at 5.6 and 10 mg/kg.
choice operant procedures with morphine.
The differences in gradient shape revealed
For example, Young, Masaki, and Geula
by the two indexes of generalization used in
(1992) trained 17 rats in a saline versus 5.6
the present study (dwell time and percentage
mg/kg morphine discrimination using a
of initial swims) are reminiscent of the con-
?xed-ratio schedule of lever pressing with
troversy involving appropriate measurement
food reinforcement, and found that the
procedures in drug discrimination, and the
mean number of sessions to acquisition cri-
resulting conceptualization of the nature of
terion was 54. In a t-maze study, Overton
stimulus control by drugs (see Colpaert,
(1982) found that a saline versus 4.0 mg/kg
1991; Stolerman, 1991; Young, 1991b). The
morphine discrimination required more than
initial-swim gradients (Figure 4) tended to re-
31 sessions, and several rats were dropped
semble those obtained from previous studies
from the study after failing to meet acquisi-
employing nominal response-selection mea-
tion criteria in fewer than 60 sessions. Thus,
sures (e.g., Colpaert, 1977); that is, these gra-
acquisition of stimulus control of navigation
dients tended to be relatively steep. On the
in the swim task by morphine appears to be
other hand, dwell-time gradients (Figure 3)
comparable to that seen in more standard
tended to resemble those using graded mea-
drug-discrimination procedures.
sures (e.g., D’Mello & Stolerman, 1978). It is
A second concern of the present study was
important to note that these graded functions
whether swimming patterns trained in the
were not simply the result of averaging across
presence of 5.6 mg/kg morphine would gen-
subjects or averaging across dose determina-
eralize to other morphine doses. Generaliza-
tions within subjects (see Figure 3). Thus, it
tion gradients were obtained using two dif-
is possible that the present procedure may of-
ferent dependent measures, and both
fer a general method by which graded drug-
revealed dose-dependent increasing func-
discrimination functions may be obtained
tions that were similar to gradients obtained
within subjects (see also McMillan & Hard-
with morphine using more traditional two-
wick, 2000; Snodgrass & McMillan, 1996).
choice procedures (e.g., Shannon & Holtz-
Although the average dwell-time gradients

---

222
DAVID ZIEGLER et al.
Table 2
Mean swimming speeds (cm/s) and standard deviations
(in parentheses) as a function of morphine dose during
generalization probe tests.
Probe
dose
Rat Y1
Rat Y2
Rat Z1
Saline
32.4 (1.3)
29.9 (0.3)
29.7 (4.4)
1.0
30.9 (1.3)
28.8 (1.3)
30.2 (7.3)
3.0
36.0 (1.8)
28.2 (1.5)
24.3 (1.1)
5.6
33.3 (0.7)
28.1 (3.3)
28.1 (8.5)
10.0
34.9 (1.3)
27.2 (0.4)
24.5 (2.8)
Y2 appeared to develop a response pattern
on probe trials that was characterized by
swimming to the condition-appropriate plat-
form area, and in the absence of the plat-
form, then swimming about the periphery of
the tank (thigmotaxis). Thus, the extinction
conditions prevailing during probe trials may
have been discriminated from training trials.
Interestingly, data from recent studies suggest
that stimulus control obtained in traditional
discrete-response procedures is relatively un-
affected by repeated exposure to extinction
(Ator, 1990; Zarcone & Ator, 2000). For ex-
ample, when food reinforcement no longer
followed lever presses, Zarcone and Ator
found that response rates declined, but that
degree of stimulus control based on a loraz-
epam training stimulus remained high. Note,
however, that in the present procedure the
rat must continue to swim throughout the ex-
tinction probe trial, even if stimulus control
declines. Thus, although a graded function
can be obtained in individuals with the dwell-
time measure, the initial-swim gradients may
provide a closer counterpart to traditional
discrete-response procedures.
The relatively slow acquisition observed in
the present study, along with the logistical
problems posed by the swim task, may limit
its utility as a tool to study drug discrimina-
tion. Variations on the present procedure
(e.g., shorter probe sessions to reduce expo-
Fig. 4.
Percentage of probe trials on which the initial
swim response was to the morphine target area as a func-
sure to extinction conditions), however,
tion of morphine dose in each rat tested.
might improve on the present results. In any
case, the present results do provide a dem-
onstration of the control of spatial navigation
generally were representative of individual
by interoceptive stimuli. The relatively pro-
sessions during the ?rst two cycles, some ?at-
longed acquisition might be related to the
tening of these gradients was apparent by the
complexity of the discrimination. Consider
third test cycle in 2 of the 3 rats. Observation
that although navigation in the Morris swim
revealed that, by the third cycle, Rats Y1 and
task is generally understood to be controlled

---

SWIM TASK AND DRUG DISCRIMINATION
223
by con?gurations of extrapool stimuli (e.g.,
guishing acquisition and performance effects. Psycho-
Morris, 1981; Sutherland & Rudy, 1989), it is
biology, 25, 217–228.
McMillan, D. E., & Hardwick, W. C. (2000). Drug dis-
dif?cult to specify the precise contingencies
crimination in rats under concurrent variable-interval
that control the performance. For example,
variable-interval schedules. Journal of the Experimental
the spatial relation between the platform and
Analysis of Behavior, 73, 103–120.
extrapool cues depends upon the position of
Morris, R. G. M. (1981). Spatial localization does not
the rat in the pool at any point in time. With
require the presence of local cues. Learning & Moti-
vation, 12, 317–338.
the procedure used in the present study, an
O’Keefe, J., & Nadel, L. (1978). The hippocampus as a
additional level of complexity is added in
cognitive map. Oxford: Claredon Press.
which the signi?cance of the extrapool con-
Overton, D. A. (1982). Comparison of the degree of dis-
?guration to navigation is under the contex-
criminability of various drugs using the T-maze drug
discrimination paradigm. Psychopharmacology, 76, 385–
tual control of an interoceptive stimulus.
395.
Stimulus control of this nature may be re-
Riley, A. L. (1997). Drug discrimination learning: As-
quired to explain some examples of naviga-
sessment of opioid receptor pharmacology. In M. E.
tion in the natural environment. For exam-
Bouton & M. S. Fanselow (Eds.), Learning, motivation,
ple, the seasonal control of direction of bird
and cognition: The functional behaviorism of Robert C.
Bolles (pp. 225–254). Washington, DC: American Psy-
migration is, in part, under the in?uence of
chological Asscociation.
different hormonal states (Berthold, 1993).
Shannon, H. E., & Holtzman, S. G. (1976). Evaluation
These and the present results indicate that
of the discriminative stimulus effects of morphine in
theories of spatial learning that have empha-
the rat. Journal of Pharmacology and Experimental Ther-
apeutics, 198, 54–65.
sized cognitive maps of distal stimuli will have
Shannon, H. E., & Holtzman, S. G. (1979). Morphine
to account for contextual control by intero-
training dose: A determinant of stimulus generaliza-
ceptive stimuli as well (Gallistel, 1990;
tion to narcotic antagonists in the rat. Psychopharma-
O’Keefe & Nadel, 1978). The techniques de-
cology, 61, 239–244.
veloped in the present study might be further
Snodgrass, S. H., & McMillan, D. E. (1996). Drug dis-
crimination under a concurrent schedule. Journal of
developed to provide a laboratory analysis of
the Experimental Analysis of Behavior, 65, 495–512.
such processes.
Stolerman, I. P. (1991). Measures of stimulus generaliza-
tion in drug discrimination experiments. Behavioural
Pharmacology, 2, 265–282.
REFERENCES
Stolerman, I. P. (1993). Drug discrimination. In F. van
Haaren (Ed.), Techniques in the behavioral and neural
Ator, N. (1990). Drug discrimination and drug stimulus
sciences: Vol. 10. Methods in behavioral pharmacology (pp.
generalization with anxiolytics. Drug Development Re-
217–243). New York: Elsevier.
search, 20, 189–204.
Sutherland, R. J., & Rudy, J. W. (1989). Con?gural as-
Berthold, P. (1993).Bird migration: A general survey. Ox-
sociation theory: The role of the hippocampal-for-
ford: Oxford University Press.
mation in learning, memory, and amnesia. Psychobiol-
Cain, D. P., Beiko, J., & Boon, F. (1997). Navigation in
ogy, 17, 129–144.
the water maze: The role of proximal and distal visual
Whishaw, I. Q., & Mittleman, G. (1986). Visits to starts,
cues, path integration, and magnetic ?eld informa-
routes, and places by rats (Rattus norvegicus) in swim-
tion. Psychobiology, 25, 286–293.
ming pool navigation tasks. Journal of Comparative Psy-
Cain, D. P., & Saucier, D. (1996). The neuroscience of
chology, 100, 422–431.
spatial navigation: Focus on behavior yields advances.
Young, A. M. (1991a). Discriminative stimulus pro?les
Reviews in the Neurosciences, 7, 215–231.
of psychoactive drugs. In N. K. Mello (Ed.), Advances
Colpaert, F. C. (1977). Drug produced cues and states:
in substance abuse (Vol. 4, pp. 139–203). London:
Some theoretical and methodological inferences. In
Kingsley.
H. Lal (Ed.), Discriminative stimulus properties of drugs
Young, A. M. (1991b). The time is ripe for an experi-
(pp. 5–21). New York: Plenum.
mental analysis of measurement issues. Behavioural
Colpaert, F. C. (1991). The discriminative response: An
Pharmacology, 2, 287–291.
elementary particle of behavior. Behavioural Pharma-
Young, A. M., Masaki, M. A., & Geula, C. (1992). Dis-
cology, 2, 283–286.
criminative stimulus effects of morphine: Effects of
D’Mello, G. D., & Stolerman, I. P. (1978). Methodolog-
training dose on agonist and antagonist effects of mu
ical issues in drug discrimination research. In F. C.
opioids. Journal of Pharmacology and Experimental Ther-
Colpaert & J. A. Rosecrans (Eds.), Stimulus properties
apeutics, 261, 246–257.
of drugs: Ten years of progress (pp. 243–252). Amster-
Zarcone, T. J., & Ator, N. A. (2000). Drug discrimina-
dam: Elsevier.
tion: Stimulus control during repeated testing in ex-
Gallistel, C. R. (1990). The organization of learning. Cam-
tinction. Journal of the Experimental Analysis of Behavior,
bridge, MA: Bradford Books/MIT Press.
74, 283–294.
Keith, J. R., & Galizio, M. (1997). Acquisition in the Mor-
ris swim task is impaired by a benzodiazepine but not
Received August 3, 2001
an NMDA antagonist: A new procedure for distin-
Final acceptance May 1, 2002

---