Behavioral Neuroscience

1987. Vol. 101, No. 6. 854-857

Copyright 1987 by the American Psychological Association, Inc

0735-7044/~7/$(X) 75

Brief Communications
Conditioned Hyperalgesia Depends on the Pain Sensitivity Measure
Marvin D. Krank
Mount Allison University
Sackville, New Brunswick, Canada
Conflicting reports about the acquisition of conditioned hyperalgesia during the development of
conditioned morphine tolerance have led researchers to suggest that tolerance reflects a reduction
of stimulus processing rather than a compensatory response interaction, l tested conditioned
hyperalgesia on both the hot-plate and tail-flick tests in the same animals. In accordance with
previous reports, the tail-flick responses in drug-free animals failed to reveal a conditioned
compensatory hyperalgesia. Conditioning effects in the tail-flick test were found only when the
animals were challenged with a low dose of morphine. However, the hot-plate responses in drugfree animals replicated earlier demonstrations of conditioned hyperalgesia. The results suggest
that the measurement of conditioned responses in drug-free animals depends on characteristics
of the assessment procedure. These findings are consistent with accounts of morphine tolerance
that depend on compensatory response interactions.

Hinson, Krank, & McCully, 1982). In addition, according to

the associative model, various conditioning procedures with
the environmental CS present at testing should affect the
magnitude of drug tolerance. Evidence to date is that tolerance
development responds as predicted to many conditioning
procedures including latent inhibition (Siegel, 1977). extinction (Siegel, 1977; Siegel, Sherman, & Mitchell, 1980). discrimination training (Hinson & Siegel, 1983: Siegel et al..
1982), partial reinforcement (Krank, Hinson, & Siegel. 1984;
Siegel, 1977), inhibitory training (Fanselow & German, 1982:
Siegel et al., 1981), and blocking and sensory preconditioning
(Dafters, Hetherington, & McCartney, 1983). Taken together.
these data strongly support a role for Pavlovian conditioning
in opiate tolerance.
One prediction of Siegel's compensatory response model,
however, has been challenged by recent observations: Attempts to measure compensatory CRs have not been uniformly successful. Failure to measure compensatory. CRs in
the absence of drug stimulation has been most notable in
studies in which changes in pain sensitivity in response to a
signal for morphine are measured. Although Siegel (1975:
Krank, Hinson, & Siegel, 1981) found evidence for a compensatory hyperalgesic CR elicited by a signal for morphine
administration, Tiffany, Petrie, Baker, and Dahl (1983) consistently did not find a hyperalgesia CR elicited by signals for
morphine administration despite the fact that these signals
control differential levels of tolerance to morphine's analgesic
effects.
Failure to find compensatory CRs under conditions in
which context-specific tolerance has been demonstrated has
led to the development of associative models of tolerance that
do not rely on the response interaction proposed by Siegel.
For example, Baker and Tiffany (1985) proposed an adaptation of Wagner's (1976) priming model of habituation to
morphine tolerance. Baker and Tiffany's model relies on a
reduction of stimulus processing rather than an antagonistic
response interaction. The priming approach to context-spe-

In recent studies of drug tolerance and sensitization, researchers have acknowledged the contribution of drug-associated cues through processes of Pavlovian conditioning. In a
typical drug administration procedure, environmental cues
predictive of drug administration serve as conditioned stimuli
(CSs), and the physiological impact of drug administration
serves as the unconditioned stimulus (UCS). The consequence
of this association between the environmental CS and drug
UCS is often the acquisition of a physiological conditioned
response (CR). Such drug CRs may be measured in drug-free
animals that are exposed to drug-associated cues and given a
placebo injection. Often in these studies the CR is opposite
in direction to the measured response to drug administration
(for reviews, see Obfil, 1966; Siegel, 1983: Siegel, Krank, &
Hinson, 1987; for a discussion of factors that influence the
direction of the CR, see Eikelboom & Stewart, 1982). Because
such drug-opposite CRs occurring in anticipation of drug
administration would be expected to cancel out or compensate for the measured effects of the drug, Siegel (1975, 1977,
1979) suggested that their presence would contribute to drug
tolerance that is defined by a reduced effect of the drug.
Critical predictions derived from the associative account of
opiate tolerance have largely been confirmed in the past few
years (for recent reviews, see Siegel, 1983; Siegel et al., 1987:
Siegel & Macrae, 1984). It is well documented, for example,
that the presence or absence of previously drug-associated
cues often determines the magnitude of tolerance to an opiate's effects (e.g. Siegel, Hinson, & Krank, 1978, 1981 : Siegel,
This research was supported by a National Sciences and Engineering Council (NSERC)grant to Marvin D. Krank.
The author acknowledges the assistance of Susan O'Neill, AnneMarie Wall and Tracy Estabrooks in collection and analysis of the
data.
Correspondence concerning this article should be addressed to
Marvin D. Krank, Department of Psychology, Mount Allison University, Sackville, New Brunswick. Canada E0A 3C0.
854

BRIEF COMMUNICATIONS
cific tolerance postulates that associative activation of the
representation of the drug U C S by the CS reduces subsequent
stimulus processing when the U C S actually occurs and thus
results in a decrement in the U C R . The reduced stimulus
processing explanation of m o r p h i n e tolerance does not depend on the presence of a compensatory C R in a nondrugged
animal and is consistent with failures to find such responses.
More recently, Palletta and Wagner (1986), adapting Wagner's( 1981 ) SOP (Sometimes O p p o n e n t Process) model, proposed a dual-process account of m o r p h i n e tolerance that
includes a role for both response interactions and reduced
stimulus processing. According to the SOP model, like its
predecessor, context-specific habituation occurs because of
reduced stimulus processing of the U C S resulting from prerepresentation of the U C S by the CS. In this model, however,
it is also acknowledged that any C R elicited by the CS will
inevitably interact with the U C R . Depending on the direction
of the CR, this interaction may either diminish or augment
the U C R . Drug CRs that m i m i c the drug U C R would augment the response magnitude. If the magnitude o f the drugmimicking C R is larger than the d e c r e m e n t in response
magnitude resulting from reduced stimulus processing, then
the net result would be sensitization. Otherwise, the net result
would be a reduction in response magnitude. C o m p e n s a t o r y
CRs would always add to the d e c r e m e n t in response magnitude resulting from reduced stimulus processing. In such
cases, two separate processes would contribute to tolerance.
Palletta and Wagner's model is consistent with both the failure
to find compensatory CRs in some preparations in which
conditioned tolerance is reported and the demonstration of
compensatory CRs in others.
Given the possible contribution of two separable processes
to tolerance development, it is i m p o r t a n t to identify the
situations in which compensatory conditioning contributes to
tolerance and those in which no evidence for such conditioning can be found. Although m a n y procedural differences may
be present, the primary difference between successful demonstrations o f conditioned hyperalgesia (Krank et al., 1981 ;
Siegel, 1975) and demonstrations o f no conditioned hyperalgesia (Tiffany et al., 1983) is the response measure used by
the experimenter, the hot-plate test versus the tail-flick test,
respectively. In this experiment, I used a within-subject design
to test whether differences in response measures could account
for the discrepancies between the results of previous assessments of conditioned hyperalgesia.

Method

Subjects a n d Apparatus
The subjects were 36 male Sprague-Dawley (CD) rats obtained
from Canadian Breeding Farms, St. Constant, Quebec. The animals
weighed 300-350 g at the start of the experiment and were allowed
free access to food and water except during the experimental sessions
(described in the Procedure section).
Pain sensitivity was measured in one of two methods: tail-flick
tests and hot-plate tests. For tail-flick tests, restrained animals had
their tails dropped in a constant-temperature hot-water bath (48 ~
Haake Model #D1-13) to a depth of 5 cm. The time until the animal
lifted its tail from the water (tail-flick latency, or TFL) was recorded

855

by two independent observers who were blind to the experimental

conditions of the animals. Trials were terminated at 30 s if no response
had occurred. For hot-plate tests, the animals were placed in a round
stainless steel tray (diameter = 21 cm) that had been submerged in
the hot-water bath at a temperature of 52 ~ The tray was encased
by a round Plexiglas cylinder to a height of 31 cm and capped with a
Plexiglas lid. Two independent observers, who were blind to the
experimental conditions of the animals, recorded the time until the
animal either licked its paw or jumped off the hot plate (paw-lick
latency, or PLL). A third observer recorded the number of times that
the animal reared during the hot-plate trial. Hot-plate trials were
terminated at 30 s if no response had occurred. Discrepancies between
the observers (which were generally small) were resolved by means of
averaging the two scores.

Procedure
Tolerance development phase. During tolerance development, all
animals received ten exposures to a "distinctive room" for 60 rain.
The distinctive room was characterized by a higher level of overhead
illumination than the animal's colony room. a different cage (Plexiglas
vs. wire mesh), and a background level of white noise (74 db). Fifteen
minutes into each distinctive-room exposure, all animals were removed from the Plexiglas cage, injected, and returned to the Plexiglas
cage for the remainder of the period. At the end of the distinctiveroom exposure, the animals were returned to their colony room.
Successive distinctive-room exposures were separated by intervals of
2-4 days. On the day after distinctive-room exposure, all animals
received an injection in the colony environment.
The animals were randomly assigned to one of three groups that
differed only during the tolerance development phase of the experiment. The three experimental groups of the animals differed only
with respect to the solutions injected in the distinctive room and in
the colony room. Group paired animals were injected with 5 mg/kg
of a 5-mg/cc solution of morphine sulphate in each exposure to the
distinctive room. These animals received equivalent volume saline
injections in the colony 24 hr later. Group unpaired animals were
injected with saline in the distinctive room, but received 5 mg/kg
injections of morphine sulphate in the colony 24 hr later. Group
control animals were injected with saline both in the distinctive room
and in the colony room.
Tail-flick CR test. In the 1lth exposure to the distinctive room,
animals in each condition were given a saline injection. Thirty
minutes later, pain sensitivity was assessed via the tail-flick measure.
After the test, the animals were returned to the Plexiglas cages for the
remainder of the exposure.
Morphine challenge test. In the 12th exposure to the distinctive
room, all animals received a 1 mg/kg dose of morphine. Thirty
minutes later, tolerance to the analgesic effect of this low dose of
morphine was assessed via the tail-flick method. After the analgesia
test, the animals were returned to the Plexiglas cages, where they
remained for the rest of the distinctive-room exposure.
Hot-plate CR test. In the 13th exposure to the distinctive room,
all animals were again given an injection of saline. Thirty minutes
later, their pain sensitivity was assessed via the hot-plate method.
Data analysis. Group scores on the tolerance test. the tail-flick
CR test, and the hot-plate CR test were analyzed via separate oneway analyses of variance (ANOVAS).Planned orthogonal comparisons
between group paired versus group unpaired and group paired versus
the group control were analyzed for evidence of conditioning effects.
Results
Table 1 shows the mean and standard errors for the three
groups on each of the three pain sensitivity tests. The tail-

856

BRIEF COMMUNICATIONS

Table 1
Mean Response Latencies on Pain Sensitivity Tests
Tail flick
Training
Paired
M
SEM

Unpaired
M
SEM

Saline

Morphine

Paw lick:
Saline

5.65
0.68

6.81 a
1.06

10.56a
2.08

5.81
0.61

9.66
1.31

15.71
2.06

Control
M

5.41
12.49
18.62
0.72
2.78
2.70
Note. Values represent the mean latency to emit the pain-sensitive
response. Lower values indicate greater pain sensitivity.
Differs from the unpaired (p < .05) and control (p < .01) groups.
SEu

flick test with saline revealed no reliable differences between

the groups, F(2, 33) < 1, ns. However, the tail-flick test with
a low dose of morphine showed that the groups differed
reliably on this measure, F(2, 33) = 3.40, p < 05. Comparisons
among the means revealed that group paired tolerance was
greater (shown by faster TFLs) than either group unpaired
tolerance ( p < .05) or group control tolerance ( p < .01).
Analysis of the difference between the saline tail-flick test
scores and morphine tail-flick test scores supported this conclusion. Higher TFLs on the morphine test than on the saline
test were found for both the unpaired group, t(11) = 2.08, p
< .05, and the control group, t ( l l ) = 3.00, p < .01 (onetailed). Group paired, on the other hand, did not have significantly higher TFLs on the morphine test, t(11) = 0.66, ns.
Last, the three groups differed on the saline hot-plate measure,
F(2, 33) = 3.98, p < .05. The paired group exhibited faster
PLLs than either the unpaired group ( p < .05) or the control
group ( p < .01 ). The group means for rearing behavior during
the hot-plate test were 2.16 for group paired, 2.58 for group
unpaired, and 3.08 for group control. The ANOVA indicated
that the groups did not differ reliably on this measure, F(2,
33) < 1, ns. Rearing and paw-lick latencies were not significantly correlated (Pearson p r o d u c t - m o m e n t correlation p =
.06, ns).
Discussion
The pattern of these results is consistent with previous tests
Of conditioned hyperalgesia. As in previous studies with the
tail-flick test (Palletta & Wagner, 1986; Tiffany et al., 1983),
conditioning with morphine did not result in a compensatory
response on a CR test in drug-free animals. Only when the
animals were challenged with a low dose of morphine did the
tail-flick measure reveal the effects of conditioning. However,
a subsequent test of pain sensitivity with saline via the hotplate method in these same animals revealed a hyperalgesia
CR. Conditioned hyperalgesia found in the hot-plate method
of assessment is consistent with previous successful demonstrations of compensatory CRs (Krank et al., 1981; Siegel,
1975). These observations demonstrate within individual animals that the type of response measure used by the experi-

menter is a critical determinant of whether a compensatory

CR will be observed.
Recently, Palletta and Wagner (1986) proposed a dualprocess theory of tolerance in which both compensatory CRs
and reduced stimulus processing contribute to tolerance. They
suggested that compensatory CRs can be measured and contribute to tolerance only in biphasic response systems. Biphasic response measures are defined as those in which the
initial drug effect gives way to a secondary effect in the
opposite direction (cf. opponent process; Solomon & Corbit,
1974). According to the model, CRs in biphasic response
systems mimic this secondary reaction to the UCS. In support
of their model, Palletta and Wagner (1986) demonstrated that
the time course of morphine's effects on the tail-flick measure
does not reveal such a biphasic effect (only analgesia was
observed), whereas activity measures do reveal a biphasic
effect (hypoactivity followed by hyperactivity was observed).
In accordance with their expectations, tail-flick tests did not
provide evidence for compensatory CRs in drug-free animals,
but activity tests did demonstrate compensatory CRs in drugfree animals.
Several investigators (Mucha, Volkovskis, & Kalant, 1981;
Palletta & Wagner, 1986; Tiffany et al., 1983) have proposed
that compensatory responses are found on the hot-plate measure of analgesia because it is confounded with activity effects.
According to this interpretation, the compensatory hyperactivity CR, which occurs on activity measures, produces apparent hyperalgesia CR on the hot-plate test. Although plausible, this explanation is not consistent with the rearing behavior found during the hot-plate test in this experiment.
There is no suggestion in the group means for rearing that a
general increase in activity accounts for the differences in
specific defensive reactions to nociception. Nor is there any
correlation between rearing behavior and paw-lick latencies.
Although rearing behavior may not be a good indicator of
conditioned activity, this observation at least illustrates the
possibility that paw-lick behavior may not be directly dependent on a general increase in behavior. Until data demonstrate
that the hot-plate measure of pain sensitivity is positively
correlated with other activity measures, the contribution of
conditioned activity to the hyperalgesic CR remains an open
question. Research on the precise relationship between pawlick responses and activity is necessary in order to decide
whether conditioned hyperalgesia is merely a reflection of
conditioned activity or a compensatory CR.
An alternative explanation of my findings is that compensatory CRs may be difficult to measure in certain response
systems (Hinson, Poulos, & Cappell, 1982; Siegel & Macrae,
1984). This interpretation is similar to that of Palletta and
Wagner (1986) except the emphasis here is on the ability to
measure the CR rather than on its contribution to drug
tolerance. Although hyperalgesic CRs cannot be measured
across a wide range of temperatures in the tail-flick test (cf.
Palletta & Wagner, 1986; Tiffany et al., 1983), these observations may demonstrate only that the tail-flick response
system does not react to conditioned hyperalgesia; that is, the
tail-flick may be a unidirectional response measure, which
responds primarily to manipulations that produce decreases
in pain sensitivity. The presence of a compensatory CR may

BRIEF COMMUNICATIONS
be revealed only if such a system is challenged (cf. Hinson et
al., 1982) or if another response is measured. My results are
consistent with the view that conditioned tolerance to morphine's effects on the tail-flick test are based on a hyperalgesic
C R that can be measured on the hot-plate test but not on the
tail-flick test.
This study is not unique in demonstrating differential sensitivity of these two response systems to analgesic and hyperalgesic events. A variety of pharmacological manipulations
also produce divergent effects on different measures of pain
sensitivity (e.g., Fasmer, Berge, Tveiten, & Hole, 1986; Ogren
& Berge, 1984: Post, Minor, Davies, & Archer, 1986). These
findings indicate that different tests measure different types
of pain sensitivity. Specifically, the tail-flick response is mediated by a spinal withdrawal reflex, whereas the paw-lick
response involves the supra-spinal integration of several m o t o r
response systems. The pharmacological and physiological differentiation of these two response systems is consistent with
the view that failure to observe conditioned hyperalgesia in
tail-flick tests reflects an inherent m e a s u r e m e n t property of
this response rather than the absence of a CR.

References
Baker, T. B., & Tiffany, S. T. (1985). Morphine tolerance as habituation. P,~3'chologicalReview, 92, 78-108.
Dafters, R., Hetherington, M., & McCartney, H. (1983). Blocking
and sensory preconditioning effects in morphine analgesic tolerance: Support for a Pavlovian conditioning model of tolerance.
Quarterly Journal ~f Experimental Psychology, 35B, 1- 11.
Eikelboom, R., & Stewart, J. (1982). Conditioning of drug-induced
physiological responses. Psychological Review, 89, 507-528.
Fanselow, M. S., & German, C. (1982). Explicitly unpaired delivery
of morphine and the test situation: Extinction and retardation of
the suppressing effects of morphine on locomotor activity. Behavioral and Neural Biology, 35, 231-242.
Fasmer, O. B., Berge, O. G., Tveiten, L., & Hole, K. (1986). Changes
in nociception after 6-hydroxydopamine lesions of descending
catecholaminergic pathways in mice. Pharmacology, Biochemisto'
and Behavior, 24, 1441-1444.
Hinson, R. E., Poulos, C. X., & Cappell, H. (1982). Effects of
pentobarbital and cocaine in rats expecting pentobarbital. Pharmacology Biochemisto' and Behavior, 16, 661-666.
Hinson, R. E., & Siegel, S. (1983). Anticipatory hyperexcitability and
tolerance to the narcotizing effect of morphine in the rat. Behavioral Neuroscience, 97, 759-767.
Krank, M. D., Hinson, R. E., & Siegel, S. (1981). Conditional
hyperalgesia is elicited by environmental signals of morphine.
Behavioral and Neural Biology, 32, 148-157.
Krank, M. D., Hinson, R. E., & Siegel, S. (1984). Effect of partial
reinforcement on tolerance to morphine-induced analgesia and
weight loss in the rat. Behavioral Neuroscience, 98, 72-78.
Mucha, R. F., Volkovskis, C., & Kalant, H. (1981). Conditioned
increases in locomotor activity produced with morphine as an
unconditioned stimulus, and the relation of conditioning to acute
morphine effect and tolerance. Journal ~f Comparative and Physiological t'sychology, 95, 351-362.
Ob~l, F. (1966). The fundamentals of the central nervous system
control of vegetative homeostasis. Acta Physiologica Academiae

857

Scientiarium Hungaricae, 30, 15-29.

Ogren, S. O., & Berge, O. G. (1984). Test-dependent variations in the
antinociceptive effect of p-chloroamphetamine-induced release of
5-hydroxytryptamine. Neuropharmacology, 23. 915-924.
Palletta, M. S., & Wagner, A. R. (1986). Development of contextspecific tolerance to morphine: Support for a dual-process interpretation. Behavioral Neuroscience, 100, 611-623.
Post, C., Minor, B. G., Davies, M., & Archer, T. (1986). Analgesia
induced by 5-hydroxytryptamine receptor agonists is blocked or
reversed by noradrenaline-depletion in rats. Brain Re~earch. 363,
18-27.
Siegel, S. (1975). Evidence from rats that morphine tolerance is a
learned response. Journal of Comparative and Physiological Psychology, 89, 498-506.
Siegel, S. (1977). Morphine tolerance acquisition as an associative
process. Journal of Experimental P.s3'choh~g)': Animal Behavior
Processes, 3, 1- 13.
Siegel, S. (1979). The role conditioning in drug tolerance and addiction. In J. D. Keehn (Ed.), Psychopathology in animals Re.search
and treatment implications (pp 143-168). New York: Academic
Press.
Siegel, S. (1983). Classical conditioning, drug tolerance and drug
dependence. In Y. Israel, R. B. Glasser, H. Kalant, R. E. Popham,
W. Schmidt, & R. G. Smart (Eds.), Research advance.~ in alcohol
and drug problems (Vol. 7, pp. 207-246). New York: Plenum.
Siegel, S., Hinson, R. E., & Krank, M. D. (1978). The role of predrug
signals in morphine analgesic tolerance: Support for a Pavlovian
conditioning model of tolerance. Journal t?[E,~7~erimentalP~)'chology. Animal Behavior Processes, 4, 188-196.
Siegel, S., Hinson, R. E., & Krank, M. D. (1981). Morphine-induced
attenuation of morphine tolerance. Science, 212, 1533-1534.
Siegel, S., Hinson, R. E., Krank, M. D., & McCully, J. (1982). Heroin
"overdose" death: The contribution of drug-associated environmental cues. Science, 216, 436-437.
Siegel, S., Krank, M. D., & Hinson, R. E. (1987). Anticipation of
pharmacological and nonpharmacological events: Classical conditioning and addictive behavior. Journal ~/ Drug 1.~ue.~, 17, 83110.
Siegel, S., & Macrae, J. (1984). Environmental specificity of tolerance.
Trends in Neuroscience, 7, 140-143.
Siegel, S., Sherman, J. E., & Mitchell, D. (1980). Extinction of
morphine analgesia. Learning and Motivation, 11, 289-293.
Solomon, R. L. & Corbit, J. D. (1974). An opponent-process theo~
of motivation. Psychological Review, 81, 119-145.
Tiffany, S. T., Petrie, E. C., Baker, T. B., & Dahl, J. L. (1983).
Conditioned morphine tolerance in the rat: Absence of a compensatory response and cross-tolerance with stress. Behavioral ,Vettroscience, 97, 335-353.
Wagner, A. R. (1976). Priming in STM: An information processing
mechanism for self-generated depression in performance. In T. J.
Tighe & R. N. Leaton (Eds.), Habituation: Per.spective~/kom child
development, animal behavior, and neurophysioh~gy (pp. 95-128).
Hillsdale, N J: Erlbaum.
Wagner, A. R. ( 1981 ). SOP: A model of automatic memoD processing
in animal behavior. In N. E. Spear & R. R. Miller (Eds.), Inlbrmation processing in animals: Memoo' mechani~m~ (pp. 5-47).
Hillsdale, N J: Erlbaum.
Received January 7, 1987
Revision June 9, 1987
Accepted June 18, 1987 9