Psychopharmacology (2000) 148:327335

Springer-Verlag 2000

O R I G I N A L I N V E S T I G AT I O N

Scott E. Bowen Stephen C. Fowler

John A. Stanford Mary Jeanne Kallman

Behavioral tolerance to the force differentiation effects of diazepam

and midazolam in rats

Received: 16 September 1998 / Final version: 23 August 1999

Abstract Rationale: Several benzodiazepines (BZs)

have been shown to increase the peak force of operant
responses at doses that increased, decreased, or had no
effect on response rate, suggesting that operant response
force may be a sensitive index of BZs effects rather than
solely a correlate of rate-dependent effects. In addition,
contingent tolerance to the rate-dependent effects of BZs
has been reported, but the degree of contingent tolerance
that develops when the critical variable of the task is
force of the response has not been explored. Objectives:
These experiments examined the effects of acute and repeated oral administration of diazepam (DZ) and midazolam (MZ) on a force-differentiation task to explore
the importance of task requirements on the development
of contingent tolerance. Methods: Two groups of rats
were trained to press a force-sensing operandum, and responses having peak forces falling within fixed lower
and upper limits [low force (810 g) or high force
(4050 g)] were reinforced with water. Acute effects
of the oral administration of DZ (0.3, 1.0, 3.0, 10.0,
30.0 mg/kg) and MZ (same doses) were determined for
the discriminated-force task before and after a repeatedadministration procedure. Results: When administered
acutely, both drugs increased the peak force of responses
in a dose-related manner and concomitantly reduced the
proportion of reinforced responses, with MZ exhibiting
greater potency. For the next 36 days, one group received drug before experimental sessions and the other
group received drug after the experimental session. A
second doseeffect determination demonstrated that rats
chronically dosed with DZ or MZ pre-session displayed
more contingent tolerance to alterations in peak force
than rats that had received 36 drug injections postsession, where there was no opportunity to practice the
S.E. Bowen S.C. Fowler J.A. Stanford M.J. Kallman
Departments of Psychology and Pharmacology,
University of Mississippi, University, MS 38677, USA
S.E. Bowen ()
Department of Psychology, Wayne State University,
71 West Warren Avenue, Detroit, MI 48202, USA
e-mail: sbowen@sun.science.wayne.edu, Fax: +1-313-577-7636

force-discrimination response while under the drug state.

Conclusions: These results suggest that perceptual motor
difficulty of the task rather than effort may be an important variable in predicting the degree of contingent tolerance that develops. Additionally, these results suggest
that both behavioral and pharmacological mechanisms
are involved in the development of drug tolerance to the
BZs.
Key words Benzodiazepine Operant Force
Tolerance Chronic Rat

Introduction
The benzodiazepines (BZs) have a behavioral profile
which includes sedative, disinhibitory, anticonvulsant,
and muscle relaxation effects. As a result of these pronounced effects, the BZs are a class of drugs with widespread clinical use. In the laboratory, it has been demonstrated that acute and chronic BZ exposure produces disruptions of both learned and unlearned behavior and that
repeated exposure produces tolerance to some of the behavioral and pharmacological actions of these compounds (see File 1985, 1990, for review). While tolerance does develop to some of the effects of BZs, the rate
of tolerance development varies widely for different behavioral effects of the BZs. The largest degree of tolerance is observed for the sedative and ataxic effects, with
little or no tolerance reported for the anti-anxiety effects
of BZs (Margules and Stein 1968; McMillan and Leander 1978; Sepinwall and Cook 1978; File 1981, 1985).
Tolerance to the acute sedative and ataxic effects of
BZs has been shown to develop when the effects are assessed by fixed-ratio (FR) (Sanger and Zivkovic 1987),
variable-interval (VI) (Sepinwall and Cook 1978), or differential reinforcement of low rates (DRL) schedules
(Fowler et al. 1993). In addition, our laboratory recently
reported that analysis of response duration and interresponse times (IRTs) from a multiple FR30DRL 20
operant schedule also reflected tolerance development to

328

the sedative effects of the ultra-short-acting BZ triazolam (Fowler et al. 1993). While the conditions necessary
for the development of tolerance to the behavioral effects
of BZs have been well defined with traditional operant
procedures, in which rate of responding is the reinforced
variable, studies investigating the effects of repeated BZ
administration using micro-analysis techniques are
sparse. Several previous reports have shown that acutely
administered chlordiazepoxide increased the peak force
of operant responses at doses that increased, decreased,
or had no effect on response rate (Fowler 1974; Fowler
et al. 1977, 1983), suggesting that operant response force
may be a sensitive index of BZs effects rather than solely a correlate of rate-dependent effects.
The present experiment examined the development
of tolerance to two BZs using the PRE (animals receive
drug before the session) versus POST (animals receive
drug after the session) drug-exposure strategy in a
force-contingent operant schedule. Other investigators,
using the PREPOST dosing design with rate-contingent learning paradigms (Chen 1968) have shown that
behavioral factors are important modulators in the development of tolerance to the BZs (Herberg and Montgomery 1987; Tang et al. 1988; Lau et al. 1990; Fowler
et al. 1993). These previous experiments have demonstrated that animals that perform a behavior while under the influence of the drug develop a greater degree
of tolerance (i.e., contingent tolerance) than animals
that perform the behavior under the non-drug state. In
the present experiment, tolerance to the effects of BZ
administration was investigated in rats whose operant
responses were reinforced only if the peak force fell
within specified upper and lower force limits (Notterman and Mintz 1962). The purpose of the force-band
limits in the present investigation was to evaluate the
development of acquired tolerance to the effects of BZs
on a rate-free discriminative task that required fine motor control. In addition, two different sets of force-band
requirements were studied in order to evaluate the effects of kinetic requirements as a possible modulator of
learned tolerance to drug effects. Specifically, this task
required one group of rats to work harder by requiring a response with a higher force emission (4050 g)
than a second group whose response forces were restricted to lighter force emissions (810 g) with both
force-bands producing about equal rates of reinforcement. Previous reports have shown that operant response requirements determine, at least in part, the degree of drug disruption seen on operant performances
(Ettenberg et al. 1981; Fowler et al. 1986). In addition
to assessing the effects of BZs on response force, a secondary goal of this experiment was to investigate the
role that duration of drug activity plays in tolerance development, with MZ representing a relatively brief duration of action and DZ representing a longer duration
of action.

Method
Subjects
The subjects were 64 experimentally naive, male Sprague-Dawley
rats (Holtzman Co., Madison, Wis.). Rats weighed approximately
350 g at the beginning of data collection. Subjects were housed individually in stainless steel cages (181824 cm) in a colony
room where a 12-h/12-h light/dark cycle was maintained (light onset approximately 0600 hours). Temperature was controlled to
262C and relative humidity levels were kept between 40% and
70%. All testing was conducted during the light phase of the cycle. Throughout the experiment, all subjects were maintained on
fixed access to water (5 min/day), which occurred after completion of the daily session. This water restriction slowed weight
gain. Animals were allowed free access to Purina Rat Chow in
their home cage.
Apparatus
Experimental sessions were conducted in four computer-interfaced
chambers, which have been previously described (Fowler et al.
1983). Briefly, test chambers contained floors consisting of parallel
stainless-steel rods; the front of the chambers were constructed of
aluminum. A cylindrical recess was mounted in the lower
center of the front panel, which permitted access to a solenoidoperated dipper that delivered water reinforcements at a volume of
0.1 ml for a period of 4 s. Each chamber was equipped with a
24-V house light and a silent, virtually isometric, force-sensing manipulandum (Model 31 load cell, 0250 g range, Sensotec, Columbus, Ohio). Transducers were positioned so that the center of the
disk was 2.5 cm from the outside of the chamber wall and the surface of the disk was 0.5 cm above the lower edge of the access window. Temporal resolution of measurements was 0.01 s and force
resolution of measurement was 1 g equivalent weight. A PC clone
computer equipped with a Labmaster interface simultaneously controlled events of the four experimental chambers and recorded data
collected from each of the chambers. The variables that were recorded using this system included the mean of the peak forces, duration (amount of time the transducer was held down during each
response), IRT, response rate, and number of reinforcers delivered.
Procedure training
Following 3 days of acclimation to the water-deprivation schedule, all rats were trained through successive approximations to
reach through the aperture and press the force-sensing manipulandum with the forepaw to receive water reinforcement. Initially, the manipulandum was positioned close to the opening to facilitate acquisition of the desired response. Once animals had learned
to make contact with the manipulandum, it was gradually moved
back to its final position of 2.5 cm from the inside of the chamber.
After the subjects reliably pressed the disc manipulandum, they
were randomly separated into two groups of 32 rats. A successive
approximation procedure was used to train each rat to press the
disc with a force within a restricted force-band. A strike of the disc
(i.e., a ballistic response) within the restricted force-band defined
the criterion response. For rats in the low-force group, the criterion
force-band was gradually changed to 810 g. For rats in the highforce group, the criterion force-band was gradually changed to
4050 g. For all rats, termination of each press within the criterion
force band was reinforced with water under a FR1 schedule. After
initial shaping of responses, daily (7 days/week) testing sessions of
15-min in duration were conducted.
Initial doseeffect determination
After responding stabilized within a 10% range for each force requirement group, the effects of diazepam (DZ) and midazolam

329
(MZ) were assessed. Half of the high-force trained group (n=16)
and half of the low-force trained group (n=16) were assigned to
DZ treatment (n=32) and the remaining animals (n=32) were assigned to MZ treatment. An acute dose effect determination was
obtained for MZ and DZ such that three intervening days were interspersed between days when dose effects were determined. Doses of 0.3, 1.0, 3.0, 10.0, and 30.0 mg/kg and the appropriate propylene glycol vehicle were administered once by gavage in a completely counterbalanced dosing order. Oral drug gavage was conducted 30-min pre-session and was based on data obtained from
previous studies (Hironaka et al. 1984; Friedman et al. 1986; Lau
et al. 1996).
Chronic exposure
Upon completion of the initial doseeffect determinations, each of
the low and high groups (n=16) for DZ and MZ were further separated into pre-session squads (PRE, n=8 per squad) and postsession squads (POST, n=8) matched for baseline response rates.
For the next 36 days, PRE subjects were administered 10 mg/kg of
DZ or MZ 30 min before the experimental session. The POST animals received the same doses, but 30 min after the test session.
The doses of DZ and MZ selected for chronic administration were
based on the data obtained during the initial doseeffect determination. To maintain identical gavage experience, both the PRE and
POST groups received a vehicle gavage when their counterparts
received the drug solution (i.e., all animals were subjected to gavage both before and after test sessions). The dose of MZ and DZ
administered chronically was selected because it produced approximately a 50% reduction in the ratio of reinforcements to responses on the force-differentiation task when the drugs were given
acutely before the session.
Doseeffect re-determination
Following the period of chronic dosing, doseeffect functions for
MZ and DZ were re-determined using the same procedure as in
the initial determination. However, subjects continued testing and
received their chronic gavage of MZ or DZ on the days between
doseeffect test days to maintain the tolerance condition. Each
subject received only one dose of the drug on a test day by administering the drug 30 min before the session in a counterbalanced
Latin square design. Higher doses than those used in the initial
doseeffect assessment were not examined for re-determining the
doseeffect curves for MZ and DZ.
Drugs
Diazepam (valium; Hoffman-La Roche) and midazolam (dormicum; Hoffman-La Roche) were dissolved in a propylene glycol
vehicle to produce a total dosage volume of 1 ml/kg for oral
administration. The vehicle consisted of propylene glycol (40%),
ethyl alcohol (10%), benzyl alcohol (1.5%), sodium benzoate
(4.9 g), and benzoic acid (120 mg). Water was added to produce a
total volume of 100 ml.

Table 1 Initial mean (+SEM)

control values for low- and
high-force groups

Dependent variable

Response force(g)
Response duration(s)
Responses
Reinforcements

Data analysis
The effects of MZ and DZ on performance during each experimental session were represented by four dependent measures:
mean peak force of responses (MPF), mean duration of responding, number of responses emitted, and a ratio of reinforcers to
responses. Drug effects were expressed as a percentage of the vehicle control. Separate repeated-measures analyses of variance
(ANOVA) were performed on data collected for each drug during
individual phases of the experiment. Prior to the application of
ANOVA procedures on peak force data, the force proportions
were logarithmically transformed, which normalized the data and
produced near homoscedasticity. Due to hardware problems, the
responses of animals tested in chamber number 4 were not properly reinforced on chronic days 2225. In order to maintain equal N
values, the data recorded for all animals were not included in the
analysis for those days. Degrees of freedom for ANOVA procedures were corrected for all incidences of lost data values due to
equipment problems.

Results
Similar to previous work with force tasks (Fowler et al.
1977, 1986), operant responding in the present experiments stabilized for each force-requirement group within
approximately 25 sessions. Characteristics of response
patterns prior to drug testing are presented in Table 1.
Whereas the distribution of forces was disbursed about
each force-band requirement, response forces for the
high-force group were distributed over a wider force
range than was observed for the low-force group (data
not shown). This is similar to the results of a previous
report which utilized a separate force-band criterion
(Kallman and Fowler 1994). However, this variation in
the force distributions for the low and high-force groups
did not result in a significantly different number of reinforcements delivered to the two groups.
Initial doseeffect determination
The initial determination of the effects of DZ and MZ on
peak force of responses is shown in Fig. 1. DZ and MZ
produced dose-dependent increases in operant force for
both high- and low-force criterion groups (F4,116=4.29,
P<0.01; F4,108=9.59, P<0.01). Post-hoc tests indicated
that doses of 330 mg/kg MZ and 1030 mg/kg DZ produced over-productions of force, with MZ being roughly
threefold more potent than DZ. While no main effect
was observed for low- versus high-force groups, a significant force requirement dose interaction was observed
for MZ (F4,108=2.47, P<0.05). Post-hoc analysis of this

Diazepam

Midazolam

Low force

High force

Low force

High force

10.22 (0.57)
0.06 (0.006)
302.53 (20.00)
83.40 (5.03)

38.27 (1.05)
0.101 (0.008)
249.81 (17.19)
88.06 (5.53)

10.04 (0.425)
0.05 (0.008)
286.64 (21.59)
83.07 (4.44)

39.04 (1.52)
0.14 (0.015)
233.33 (24.25)
74.20 (2.56)

330

Fig. 1 The acute effects of orally administered diazepam (DZ)

(upper panels) and midazolam (MZ) (lower panels) on operant
force before (initial) and after repeated administration either preceding (pre-session) or following (post-session) behavioral sessions. The left side of the figure depicts the low-force band task
for both drugs and the right side represents the high-force band
task. The mean values plotted are a proportion of control forces
(SEM)

Fig. 2 The acute effects of orally administered diazepam (DZ)

(upper panels) and midazolam (MZ) (lower panels) on number of
responses before (initial) and after repeated administration either
preceding (pre-session) or following (post-session) behavioral sessions. The left side of the figure depicts the low-force band task
for both drugs with the right side representing the high-band task.
Mean proportion of control responses (SEM) is shown

interaction indicated that MZ disrupted responding under

the low-force requirement to a greater degree than under
the high-force requirement. The increase in response
force was accompanied by significantly longer response
durations for the MZ-treated animals, but not for the DZtreated animals (F4,108=5.65, P<0.01) (data not shown).
Response rates were increased in a dose-dependent fashion (Fig. 2) with concurrent decreases in the number of
reinforcements for both DZ and MZ (Fig. 3 and Fig. 4).

across the repeated-exposure period, indicating a decline

in the force-increasing effects of DZ and MZ. DZ dosing
produced a significant low/high PRE/POST days interaction (F15,390=1.92, P0.05), while MZ exposure produced a significant PRE/POST days interaction
(F15,390=5.59, P0.01) and a low/high days interaction
(F15,375=2.45, P0.01). Post-hoc analysis of these interactions revealed that rats treated with DZ before the daily
session and working under the low-force requirement had
slight increases in response forces across days of repeated
administration. Rats orally treated with 10 mg/kg MZ
initially displayed higher forces than rats treated with
10 mg/kg of DZ, regardless of whether they were given
drug before or after the session, but this difference dissipated as the length of chronic exposure increased.
The initial effects of DZ and MZ, to elevate response
rates and decrease the ratio of reinforcers delivered to
responses, decreased for both BZs with repeated exposure but was more pronounced in animals treated with
MZ. As seen in Fig. 5, MZ-treated animals that were ad-

Chronic drug exposure

While chronic administration of DZ (10 mg/kg/day) produced no PREPOST differences for response force, repeated MZ administration increased forces for rats treated
PRE relative to those treated POST (F1,25=8.64, P0.01)
(Fig. 5). With the exception of rats treated with DZ before the daily session and working under the low-force
requirement, there was a tendency for force to decrease

331

Fig. 3 The acute effects of orally administered diazepam (DZ)

(upper panels) and midazolam (MZ) (lower panels) on number of
reinforcers before (initial) and after repeated administration either
preceding (pre-session) or following (post-session) behavioral sessions. The left side of the figure depicts the low-force band task
for both drugs with the right side representing the high-band task.
Mean proportion of control reinforcers (SEM) is shown

Fig. 4 The acute effects of orally administered diazepam (DZ)

(upper panels) and midazolam (MZ) (lower panels) on the ratio of
responses/reinforcements before (initial) and after repeated administration either preceding (pre-session) or following (post-session) behavioral sessions. The left side of the figure depicts the
low-force band task for both drugs with the right side representing
the high-band task. Mean proportion of the control ratio of reinforcements/responses (SEM) is shown

ministered drug before the session (PRE) displayed an

across days decrease in responses (F15,375=4.67, P0.01)
and the ratio of reinforcers delivered to responses
(F15,375=3.43, P0.01), while rats treated after the daily
session displayed little change in responses across the
days of repeated exposure. For DZ, PRE-session treatment disrupted response rates to a greater extent than
did POST-session treatment (F1,26=13.42, P0.01), and
the ratio of reinforcers to responses was lower when rats
were treated with DZ PRE-session (F1,26=4.99, P0.05).
Overall, the effect of PRE-session DZ administration
was to increase frequency of responding and to decrease
the ratio of reinforcers delivered when force of responding was not different for the PRE/POST conditions. Animals trained under the low force requirement generated
more responses than did animals trained under the high
force condition.

Initial doseeffects versus re-determination

To examine non-contingent tolerance, the initial dose
effect and re-determination doseeffect data for PRE and
POST groups were combined (PRE/low group with
POST/low group and PRE/high group with POST/high
group) for all doses of each drug and then compared.
Chronic drug administration significantly shifted response forces to the right in a nonparallel fashion for
both DZ (F1,28=39.69, P<0.01) and MZ (F1,27=100.21,
P<0.01). Doseeffects were seen for both DZ
(F4,112=6.19, P<0.01) and MZ (F4,108=14.63, P<0.01).
The observation of a determination by low/high-force interaction (F1,27=5.17, P<0.05) and a determination
dose interaction (F4,108=2.55, P<0.05) for MZ indicates
that high-force animals were disrupted to a greater extent
on the initial dose-effect curve than on the re-determination curve, which was not seen under the low-force contingency. The reduction in the force disruptions from the
initial dose determination to the re-determination is sug-

332
Fig. 5 The chronic effects of
orally administered diazepam
(DZ) (left panels) and midazolam (MZ) (right panels). The
top panels depict the mean
peak forces, middle panels
show response rates, and the
lower panels represent the ratio
of reinforcements to responses.
Due to hardware problems, animals tested in chamber number
4 were not reinforced properly
on chronic days 2225. In order to maintain equal N values,
the data recorded for all animals were not included in the
analysis or graphics for those
days

gestive of the development of tolerance and explains the

interaction with dose.
While comparisons of initial versus re-determination
doseeffect curves revealed no significant alterations for
DZ on response durations or number of responses produced, chronic drug administration significantly shifted
mean response durations (F1,27=4.49, P<0.05) and response output (F1,27=12.38, P<0.01) to the right for MZ.
The reduction of responses with a concomitant increase
in reinforcers brought about a nonparallel shift in the
doseeffect to the right for the ratio of reinforcers/
responses for both DZ (F1,28=16.12, P<0.01) and MZ
(F1,27=27.91, P<0.01). The data on the ratio of reinforcers/responses (Fig. 4) indicate that DZ and MZ were not
as disruptive during the re-determination as seen during
the initial doseeffect determination.

Discussion
The oral administration of DZ and MZ produced very
similar patterns of behavioral disruption, which would be
expected with two drugs of the same class. Both drugs
induced dose-related increases in response rate, duration,
and force, which resulted in a marked decrease in the
proportion of reinforced responses. Both MZ and DZ are
very lipophilic and, as a result, have comparable brain
uptakes (Arendt et al. 1987). In addition, while MZ is a
short-acting BZ in both rats and humans, DZ is not. Although DZ is a long-acting BZ in humans with a half-life
between 20 h and 70 h (Greenblatt et al. 1982), it is
short-acting in the rat, being quickly taken up into the
brain (Hironaka et al. 1984). Indeed, a 5-mg/kg dose of
DZ has been reported to produce a plasma half-life of

333

approximately 0.88 h (Friedman et al. 1986). With the

reported serum half-life of 0.92 h for a 3-mg/kg dose of
MZ in the rat (Tang et al. 1988), it should be of no surprise that these two BZs produce comparable effects on
the force differentiation task in the present studies.
The acute behavioral effects of MZ and DZ on a task
requiring the discrimination of response force to obtain
reinforcement were observed over the same range of
doses (0.330.0 mg/kg) as those reported in previous
studies in rodents which examined various behavioral effects (File 1985; Griffiths and Goudie 1987). Results obtained by Canon and Lippa (1977) have shown that orally administered DZ (515 mg/kg) dose dependently decreased response rates maintained by a FR schedule and
increased burst responding as measured by a DRL schedule. Similar results have also been reported for orally administered triazolam, an ultra-short-acting BZ, on
FR30DRL20 responding (Fowler et al. 1993). The increases in average response rates seen here with DZ and
MZ may have been a by-product of the disrupted force
emission (overshooting the band) produced by the drug.
One possible interpretation of this is that, because the
drugs reduced force accuracy, reinforcers were delivered
less frequently, resulting in the rats spending relatively
more time striking the operandum. The hypothesis that
the BZs may directly stimulate response rates is not supported by previous reports which have shown BZs to
predominantly decrease response rates maintained by FR
schedules (Canon and Lippa 1977; Fowler et al. 1993).
The present results with DZ and MZ support previous
reports of the acute effects of sedative-hypnotic compounds on tasks evaluating response forces. For example,
during a procedure in which rats were required to hold a
lever coupled to a force transducer with a paw so that it
remained between upper and lower limits of a force-band
for a continuous 1.5-s period, MZ (Tang et al. 1988) and
DZ (Culberson et al. 1990) were reported to impair indices of motor performance as well as work session rates.
Previous investigations utilizing the same force operandum that was used in the present experiment have shown
that chlordiazepoxide dose-dependently increased response forces under a variety of response topographies,
including a FR20 operant schedule (Fowler et al. 1977),
an extinction schedule (Fowler et al. 1974), and a schedule in which one group of rats was required to exert
downward responses on a force-sensing circular disk and
in a second group, which was required to grasp and pull a
wire bail attached to a force transducer (Fowler et al.
1983).
The addition of two different sets of force-band requirements (low vs high) in the present study to evaluate
the effects of kinetic requirements as a possible modulator of behavioral drug effects proved to have only a
small effect on the levels of behavioral disruption produced by drug administration and on the degree of tolerance that developed to this disruption. While the results
indicate that there was some evidence for drug disruptions dependent on base levels of force training, these
disruptions were primarily limited to MZ, with the initial

acutely administered drug producing somewhat greater

increases in the low-force group. The fact that the greater disruption occurred in the low-force condition suggests that the difference between low and high was possibly related to greater perceptual motor difficulty of the
low condition rather than to effort requirements per se.
The increased response forces observed for DZ and
MZ in the present study and for chlordiazepoxide in the
previous investigations may be a result of the anticonflict and/or anxiolytic qualities of BZs, which are
known to increase behavioral output in aversive situations. Fowler et al. (1977) postulated that the aversive situation could be extended to force contingent situations
where the aversiveness was created by high levels of exertion (i.e., force requirements) instead of by traditional
punishment procedures. Spinally-mediated disinhibitory
mechanisms have also been postulated to play a role in
BZs force-increasing effects (Fowler et al. 1983). A third
explanation for these force increases comes from work
completed by Arnone and Dantzer (1980) and Fowler
(1974). Arnone and Dantzer (1980) reported that DZ produced elevations in extinction-induced aggression in pigs.
These elevations are opposite to what would be expected
for an anti-anxiety/anti-frustration drug. In addition,
Fowler (1974) reported that chlordiazepoxide further increased peak force that was already increased by extinction. These results suggest that the BZs may directly
stimulate behaviors induced by extinction situations.
With regard to tolerance, re-determination of acute effects of DZ and MZ after chronic dosing provided evidence for the role of both pharmacological and behavioral factors in tolerance development. The strongest evidence for learned tolerance was seen for DZ- and MZmediated changes in the dependent measure of response
peak force. Re-determination of acute drug effects produced lower response forces for rats dosed both before
and after the test session than was observed when these
doseeffect curves were determined before chronic exposure. Animals trained to produce low forces tended to
display greater shifts in the doseresponse curve than animals trained to produce high forces, regardless of the
drug chronically administered. Initial peak force elevations produced by a 10.0-mg/kg dose of DZ fell to 72%
of initial values for the low-force group, while the comparable value for the high-force group was 84% of initial
values. For an identical dose of MZ, peak force elevations were reduced to 64% of initial values for the lowforce group and 82% for the high-force group. These
findings indicate that drug experience while performing
the task significantly reduced the effects of both BZs
tested, confirming the results of two previous studies in
which tolerance developed to the performance decrements of MZ (Tang et al. 1988) and DZ (Culberson et al.
1990) in a procedure requiring rats to hold a lever between upper and lower limits of a force band for a continuous 1.5-s period. In addition, the present results demonstrated that the degree of tolerance may be affected by
the task requirements, since low-force task demands
were correlated with a greater degree of tolerance than

334

was observed for the high-force task when drug exposures were repeated for 36 days.
Regarding learned tolerance, both DZ and MZ rats
dosed pre-session displayed a greater attenuation of drug
effect than those animals receiving drug post-session.
Initial versus re-determination differences were most
pronounced for the response rate and reinforcement ratio
measures for DZ, while MZ-treated animals displayed
some degree of tolerance on all of the dependent measures. Interestingly, no indication of sensitization to the
effects of the BZs was seen in this study for the rats
dosed after the session or for any of the non-forcedependent variables. Figure 2 and Figure 3 reveal that, as
the proportion of reinforced responses declined with increasing doses (decreased in-band efficiency), the rate of
response correspondingly increased, resulting in an almost constant rate of reinforcement (Fig. 4). Thus, rats
were able to compensate for their overshoot errors by responding more frequently per unit of time. In effect, the
variable ratio of reinforcement became leaner as the
acute dose increased and as the frequency of overshoot
or out-of-band errors increased. This reasoning implies
that the ratio of effort to reward is a factor in determining tolerance effects. Consequently, the reinforcementloss hypothesis (Schuster et al. 1966) can be extended to cover conditions in which the initial drug effect produces a greater effort-to-reward ratio without affecting the rate of reinforcement. Apart from these considerations, the results suggest that practicing a skilled
motor task in the drugged state affords some additional
tolerance beyond that engendered by equivalent drug
treatment, which is not correlated with task performance.
Although these findings tend to support a reinforcementloss hypothesis, there are studies in the literature
that do not support this hypothesis. Branch (1979) reported tolerance to drug-induced increases in rates of responding maintained by a fixed-interval schedule that
did not produce reductions in reinforcements. In a later
study, Dworkin and Branch (1982) reported the development of morphine tolerance in monkeys trained to respond on a continuous shock-avoidance schedule. Tolerance to the rate suppression produced by morphine did
occur, but only sufficiently to re-establish baseline levels
of shock frequency, while response rates continued to remain somewhat depressed. Clearly, other variables such
as dose of drug delivered chronically, behavioral effect
examined, frequency of drug exposure, and degree of behavioral alteration may all play a role in the degree and
topography of learned tolerance that is displayed.
In summary, DZ and MZ produced graded increases
in response forces in rats under two different force-band
conditions, with the acute effects of both drugs being
slightly greater on the low-force requirement than on the
high-force requirement. Repeated pre-session or postsession treatment with either drug produced a shift to the
right in the doseeffect function for response forces.
These effects, like similar effects reported previously for
chlordiazepoxide, may reflect the central nervous system
muscle relaxant and ataxic properties of BZs. Addition-

ally, these results demonstrate the utility of quantitative

measures of motor control (i.e., response force and duration) and should prove useful for future investigations of
tolerance to BZs effects.
Acknowledgements This research was supported by DA05253
awarded to MJK, and these data were submitted as partial fulfillment of the requirements for the doctoral degree for SEB. All animal experiments were conducted in compliance with the principles
of laboratory animal care as stated in the Guide for the Care and
Use of Laboratory Animals (National Research Council 1996).

References
Arendt RM, Greenblatt, DJ, Liebisch DC, Luu MD, Paul SM
(1987) Determinants of benzodiazepine brain uptake: lipophilicity versus binding affinity. Psychopharmacology 93:72
76
Arnone M, Dantzer R (1980) Effects of diazepam on extinction induced aggression in pigs. Pharmacol Biochem Behav 13:
2730
Branch M (1979) Consequent events as determinants of drug effects on schedule-controlled behavior: modification of effects
of cocaine and d-amphetamine following chronic amphetamine administration. J Pharmacol Exp Ther 210:354360
Canon JG, Lippa AS (1977) Use of drl in differentiating anxiolytic
and neuroleptic properties of CNS drugs. Pharmacol Biochem
Behav 6:591593
Chen C (1968) A study of alcohol tolerance to ethanol as a function of reinforced practice in rats. Psychopharmacologia 63:
258288
Culberson JW, Tang M, Lau CE, Falk JL (1990) Diazepam and
discriminative motor control: acute, chronic and withdrawal
effects. Pharmacol Biochem Behav 35:419427
Dworkin SI, Branch MN (1982) Behavioural effects of naloxone
following chronic morphine administration. Psychopharmacology 77:322326
Ettenberg A, Koob GF, Bloom FE (1981) Response artifact in the
measurement of neuroleptic-induced anhedonia. Science 213:
357359
File SE (1981) Rapid development of tolerance to the sedative
effects of lorazepam and triazolam in rats. Psychopharmacology 73:240245
File SE (1985) Tolerance to the behavioral actions of benzodiazepines. Neurosci Biobehav Rev 9:113121
File SE (1990) The history of benzodiazepine dependence: a review of animal studies. Neurosci Biobehav Rev 14:135146
Fowler SC (1974) Some effects of chlordiazepoxide and chlorpromazine on response force in extinction. Pharmacol Biochem
Behav 2:155160
Fowler SC, Filewich RJ, Leberer MR (1977) Drug effects upon
force and duration of response during fixed ratio performance
in rats. Pharmacol Biochem Behav 6:421426
Fowler SC, Lewis RM, Gramling SE, Nail GL (1983) Chlordiazepoxide increases the force of two topographically distinct operant responses in rats. Pharmacol Biochem Behav 19:787790
Fowler SC, Gramling SE, Liao RM (1986) Effects of pimozide on
emitted force, duration and rate of operant response maintained at low and high levels of required force. Pharmacol
Biochem Behav 25:615622
Fowler SC, Bowen SE, Kallman MJ (1993) Practice-augmented
tolerance to triazolam: evidence from an analysis of operant
response durations and interresponse times. Behav Pharmacol
4:147157
Friedman H, Abernethy DR, Greenblatt DJ, Shader RI (1986) The
pharmacokinetics of diazepam and desmethyldiazepam in rat
brain and plasma. Psychopharmacology 88:267270
Greenblatt DJ, Shader RI, Abernathy DR, Ochs HR, Divoll M,
Sellers EM (1982) Benzodiazepines and the challenge of pharmacokinetic taxonomy. In: Usdin E, Skolnick P, Tallman JF Jr,

335
Greenblatt D, Paul SM (eds) Pharmacology of benzodiazepines. Macmillan, London, pp 257269
Griffiths JW, Goudie AJ (1987) Analysis of the role of behavioral
factors in the development of tolerance to the benzodiazepine
midazolam. Neuropharmacology 26:201209
Herberg LJ, Montgomery AMJ (1987) Learnt tolerance to sedative
effects of chlordiazepoxide on self-stimulation performance,
but no tolerance to facilitatory effects after 80 days. Psychopharmacology 93:214217
Hironaka T, Katsuhiro F, Fuji T (1984) Absorption of diazepam
and its transfer through the blood brain barrier after intraperitoneal administration in the rat. J Pharmacol Exp Ther
229:809815
Kallman MJ, Fowler SC (1994) Assessment of chemically induced alterations in motor function. In: Chang L (ed) Handbook of neurotoxicology. I. Basic principles and current concepts. Dekker, New York, pp 373396
Lau CE, Falk JL, Tang M (1990) Motor performance decrement
by midazolam: antagonism by RO151788 and CGS8216.
Pharmacol Biochem Behav 36:139143
Lau CE, Ma F, Wang Y, Smith C (1996) Pharmacokinetics and
bioavailability of midazolam after intravenous, subcutaneous,
intraperitoneal and oral administration under a chronic foodlimited regimen: relating DRL performance to pharmacokinetics. Psychopharmacology 126:241248

Margules DL, Stein L (1968) Increase of antianxiety activity

and tolerance of behavioral depression during chronic administration of oxazepam. Psychopharmacologia 13:7480
McMillan DE, Leander JD (1978) Chronic chlordiazepoxide and
pentobarbital interactions on punished and unpunished behavior. J Pharmacol Exp Ther 207:515520
National Research Council (1996) Guide for the care and use of
laboratory animals. National Academy Press, Washington,
D.C.
Notterman JM, Mintz DE (1962) Exteroceptive cueing of response
force. Science 135:10701071
Sanger DJ, Zivkovic B (1987) Investigation of the development of
tolerance to the actions of zolpidem and midazolam. Neuropharmacology 26:15131518
Schuster CR, Dockens WS, Woods JH (1966) Behavioral variables affecting the development of amphetamine tolerance.
Psychopharmacologia 9:170182
Sepinwall J, Grodsky FS, Cook L (1978) Conflict behavior in the
squirrel monkey: effects of chlordiazepoxide, diazepam and ndesmethyldiazepam. J Pharmacol Exp Ther 204:88102
Tang M, Lau CE, Falk JL (1988) Midazolam and discriminative
motor control: chronic administration, withdrawal and modulation by the antagonist RO151788. J Pharmacol Exp Ther
246:10531060