PUNISHMENT AND THE POTENTIAL FOR NEGATIV

JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR

2018, 109, 365–379

NUMBER

2 (MARCH)

PUNISHMENT AND THE POTENTIAL FOR NEGATIVE REINFORCEMENT WITH
HISTAMINE INJECTION
PAULO CÉSAR MORALES MAYER1, MARCUS BENTES DE CARVALHO NETO2, AND
JONATHAN L. KATZ3
1

UNIVERSIDADE ESTADUAL DO OESTE DO PARANÁ - UNIOESTE
2

UNIVERSIDADE FEDERAL DO PARÁ - UFPA

3

NATIONAL INSTITUTE ON DRUG ABUSE – INTRAMURAL RESEARCH PROGRAM

The present study examined punishment of responding with histamine injection, and its potential to
generate avoidance of punishment. Sprague–Dawley rats were trained under concurrent schedules in
which responses on one lever (the punishment lever) produced food under a variable-interval schedule,
and under some conditions intermittent injections of histamine, which suppressed behavior. Responses
on a second (avoidance) lever prevented histamine injections scheduled on the punishment lever. After
stabilization of punished responding, a variable-interval 15-s schedule of cancellation of histamine
(avoidance) was added for responding on the second/avoidance lever, without subsequent acquisition
of responding on that lever. Progressive decreases in the length of the punishment variable-interval
schedule increased suppression on the punishment lever without increases in response rates on the
avoidance lever. Exchanging contingencies on the levers ensured that response rates on the avoidance
lever were sufficiently high to decrease the histamine injection frequency; nonetheless response rates
on the avoidance lever decreased over subsequent sessions. Under no condition was responding maintained on the avoidance lever despite continued punishing effectiveness of histamine throughout. The
present results suggest that avoidance conditioning is not a necessary condition for effective punishment, and confirm the importance of empirical rather than presumed categorization of behavioral
effects of stimulus events.
Key words: punishment, suppression, avoidance, negative reinforcement, histamine, lever press, rats

Most experiments involving punishment use
electric shock as the punishing stimulus, which

has well-documented advantages such as its
quantifiable physical dimensions (shock intensity, period, and duration). Those dimensions of
electric shock are readily controlled, allowing its
use across a wide range of values, and
The authors thank Drs. Kennon A. Lattal, M. Jackson
Marr, Frans van Haaren, and John V. Keller for comments
on a previous version of the manuscript. All experiments
were approved by the NIDA IRP Animal Care and Use
Committee.
The data reported herein constituted part of the
requirements of Paulo César Morales Mayer for the
degree of Doctoral in Behavioral Theory and Research at
Universidade Federal do Pará - UFPA, with Marcus Bentes
de Carvalho Neto as mentor (CNPq Grant #311603/20165). The experiments were conducted at the Intramural
Research Program of the National Institute on Drug
Abuse. Dr. Mayer’s time at NIDA-IRP was funded in part
by scholarship from the CAPES (PDSE#3878/13-9), Ministry of Education, Brazil.
Address correspondence to Jonathan L. Katz, NIDA
Intramural Research Program – Molecular Neuropsychiatry Research Branch, 251 Bayview Blvd., Baltimore, MD
21224. Phone 001-443-812-1415; e-mail JKATZ@intra.nida.

nih.gov.
doi: 10.1002/jeab.319

facilitating replication and direct comparisons
across studies (Azrin & Holz, 1966; Barker et al.,
2010; Dinsmoor, 1998). Among previously
noted drawbacks of electric shock are its welldocumented eliciting functions (e.g., Flaherty,
1985), which may contribute to competition
between responses, and may complicate the
interpretation of its effects.
The generality of the behavioral principles
established with electric-shock punishment
can be tested or extended by using novel stimuli (Barker et al., 2010; Branch, Nicholson, &
Dworkin, 1977; Catania, 2013; Church, 1969;
Dinsmoor, 1998). Although infrequently
employed, examples of various punishing stimuli include: intense noise (e.g., Friedel,
DeHart, & Odum, 2017; Holz & Azrin, 1962),
air blasts (e.g., Carvalho Neto et al., 2005;
Spealman, 1978), bright light (e.g., Barker
et al., 2010), species-specific stimuli (e.g., Masserman & Pechtel, 1953), and time out from

positive reinforcement (e.g., Ferster, 1958;
Herrnstein, 1955; though see Leitenberg,
1965, for the varied effects of time out). Each
of these stimulus groups have their own
strengths and weaknesses, the latter often

Published 2018. This article is a U.S. Government work and is in the public domain in the USA.
365

366

PAULO CÉSAR MORALES MAYER et al.

including, but not limited to, a lack of precise
physical
dimensions,
habituation,
and
unscheduled avoidance.
Consequent histamine injection is an alternative punishing stimulus that has been studied with some frequency and with several of

the advantages of an ideal punishing stimulus,
as described by Azrin and Holz (1966). For
example, it can be used across a wide range of
magnitudes (see dose-effect curves in Goldberg, 1980; Katz & Goldberg, 1986; Podlesnik & Jimenez-Gomez, 2013; Sharpless, 1961).
Further, histamine has precise physical/pharmacological specifications; the actual contact
with the organism is constant from one occurrence to the next (there are no indications of
pharmacological chronic tolerance or tachyphylaxis), and options for escape or avoidance
are not possible unless specifically programmed (e.g., Takada, Winger, Cook, Larscheid, & Woods, 1986). Histamine has been
used as an alternative to electric shock as a
punisher in various schedule arrangements
(e.g., Goldberg, 1980; Holtz & Carroll, 2015;
Negus, 2005; Podlesnik, Jimenez-Gomez, &
Woods, 2010). Though other drugs have also
been used (Koffarnus & Winger, 2015; Prada,
Takada, Katz, Goldberg & Barrett, 1987;
Takada, Barrett, Allen, Cook, & Katz, 1992),
histamine may be best suited among drugs for
use as a punisher due to its fast onset and offset (Goldberg, 1980). Finally, histamine, at the
doses presently utilized, elicits no known skeletal responses (Kuraishi, Nagasawa, Hayashi, &
Satoh, 1995; White & Rumbold, 1988), and

the known elicited cardiovascular effects are
pharmacologically independent of its punishing effects (Goldberg, 1980; Podlesnik &
Jimenez-Gomez, 2013).
The generality of the behavioral principles
established with electric shock have been
extended with the use of histamine. Studies
have established that as with electric shock,
suppression of behavior is related to the magnitude of the stimulus (e.g., histamine dose:
Goldberg, 1980; Katz & Goldberg, 1986; Podlesnik et al., 2010), injection delay (Woolverton,
Freeman, Myerson, & Green, 2012), and various scheduling conditions of the procedures
employed (e.g. Negus, 2005; Podlesnik &
Jimenez-Gomez, 2013). Additionally, the
effects of drugs on punished responding with
histamine appear to be largely similar to those
with electric-shock punishment (e.g. Goldberg,

1980; Katz & Goldberg, 1986). Finally, as with
electric shock, histamine has also been shown
to function as a negative reinforcer (e.g.
Takada et al., 1986).

Azrin, Hake, Holz and Hutchinson (1965)
studied punishment and concurrently, the
potential of that punishment to generate negative reinforcement of another response.
Pigeons were trained to peck one key with
food reinforcement under a fixed-ratio
(FR) 25-response schedule. After electricshock punishment was introduced as a consequence of each response, another contingency
was introduced in which a response produced
a change in overall illumination in the chamber and eliminated punishment for the
remainder of the fixed ratio. As expected,
punished response frequency was an inverse
function of punishment magnitude (shock
intensity), whereas the escape response frequency was a direct function of punishment
magnitude. These relations were obtained
with different schedules of food reinforcement, and the escape response showed
schedule-appropriate patterns when reinforced under fixed-interval and fixed-ratio
schedules. In a latter discussion of that finding, Azrin and Holz (1966) stated that “a
major effect of punishing a response is to generate a strong tendency on the part of the subject to escape from the punishing situation
entirely” (p. 408).
Lattal and Cooper (1969) also trained
pigeons to peck a key under an FR 25 schedule of food presentation and punished this

responding on an FR1 schedule of electricshock presentation. The punishment contingency could be avoided if any two responses
were interspaced by a specified interval
(15, 30, or 45 s), a contingency which
resourcefully exploits its compatibility with
operant response suppression. As in the
study by Azrin et al. (1965), meeting the
response omission requirement produced a
change in the color of the response key
allowing the completion of the FR 25 schedule of food presentation without punishment. Avoidance responding was maintained
most prominently at the lower omission time
requirements.
Arbuckle and Lattal (1987) examined
responding of pigeons trained to key peck
with food reinforcement delivered on a VI
180-s schedule. Subsequently, a VI 30-s

HISTAMINE PUNISHMENT AND NEGATIVE REINFORCEMENT
schedule of electric-shock presentation was
added to the schedule of food presentation
which resulted in a small punishing effect.

Once performances stabilized, an avoidance
contingency was added such that the electric
shock could be avoided if the time between
two successive responses was greater than
5, 10 or 30 s (during different phases of the
study). In contrast to the previous studies,
meeting the avoidance contingency produced
no exteroceptive stimulus change. With the
introduction of the avoidance contingency the
shock frequency initially decreased six-fold or
greater. Further, the response rate was an
inverse function of the avoidance interval, as
would be expected for avoidance responding
(e.g. Sidman, 1953). Those results suggested
that a punishment contingency can be a sufficient predisposing condition for the establishment of avoidance responding.
The present series of experiments extended
those by Arbuckle and Lattal (1987) further
investigating the relation between avoidance
and punishment with histamine injection as
the punishing stimulus. In the Arbuckle and

Lattal study, the avoidance response was a particular spacing of responses on the same operandum as the punished response. As the
duration of interresponse times among punished responses can be selectively modified by
punishment contingencies alone (Everly &
Perone, 2012; Galbicka & Branch, 1981), the
present study extended the analysis to assess
the effects of punishment and avoidance with
contingencies arranged on different levers
(topographical tagging, Catania, 1973).
Method
Subjects
Seven male Sprague–Dawley rats (Taconic
Farms, Germantown, NY), weighing approximately 300 g at the start of the study served as
subjects. One subject (PM06) died prematurely and was replaced with another (PM16).
The subjects were acclimated to a
temperature- and humidity-controlled vivarium for at least 1 week with a 12-hr light/dark
cycle (lights on at 07:00 a.m.) during which
food (Scored Bacon Lover Treats; Bio-Serv,
Frenchtown, NJ) and tap water were available
at all times. After acclimation, body weights
were maintained at approximately 325 g by

adjusting daily food rations. Animals had free

367

access to water at all times in the home cages.
Care of the subjects was in accordance with
the guidelines of the National Institutes of
Health and the National Institute on Drug
Abuse, Intramural Research Program, Animal
Care and Use Program, which is fully accredited by Association for Assessment and
Accreditation of Laboratory Animal Care
International.
Surgical preparation and postsurgical care.
Under anesthesia (ketamine 60.0 mg/kg and
xylazine 12.0 mg/kg, i.p.) chronic indwelling
catheters were surgically implanted in the
right external jugular vein. Catheters exited at
the midscapular region of the subject through
a back mount which allowed connection with
the subject to the injection pump. After surgery subjects received heparinized saline
(50 IU/ml, i.v.) and subcutaneous antibiotic
(enrofloxacin, 5.0 mg/kg in saline) daily to
minimize the likelihood of infection and the
formation of clots or fibroids. Catheter
patency was checked about every 10 days with
methohexital injection (1.5 mg/kg). If the
catheter failed, it was removed and after at
least 7 days another vein (either left jugular,
right femoral or left femoral) was catheterized.
All subjects were allowed to recover from surgery for approximately 7 days before returning
to behavioral procedures.
Apparatus
Daily sessions were conducted using six
operant-conditioning chambers (modified
ENV-203; Med Associates, St. Albans, VT) that
measured 25.5 × 32.1 × 25.0 cm. The chambers were enclosed within sound-attenuating
cubicles that were equipped with a fan for ventilation and were provided with white noise to
mask extraneous sounds. The front wall of
each chamber contained a house light (28 V,
100 mA) at the ceiling and two response
levers, mounted 5.0 cm from the midline and
4.0 cm above the grid floor. A downward displacement of a lever with a force approximating 0.20 N defined a response and always
activated a relay mounted behind the front
wall of the chamber producing an audible
“feedback” click. Six light-emitting diodes
(LEDs) were located in a row above each
lever, three green and three yellow. A receptacle for the delivery of 45-mg food pellets via a
pellet dispenser (Model ENV-203-20; Med

368

PAULO CÉSAR MORALES MAYER et al.

Associates) was mounted on the midline of
the front wall between the two levers and
2.0 cm above the floor. The three green LEDs
above the lever on which the response was
emitted flashed on for 0.1 s with each food
pellet delivery. An injection pump (Model 22;
Harvard Apparatus, Holliston, MA, set at a rate
of 4.22 ml/min.) placed above each cubicle
delivered injections of specified volumes from
a 20-ml syringe. The syringe was connected by
Tygon tubing to a single-channel fluid swivel
(375 Series Single Channel Swivels; Instech
Laboratories, Plymouth Meeting, PA) that was
mounted on a balance arm above the chamber. Tygon tubing from the swivel to the subject’s catheter completed the connection to
the subject and was protected by a surrounding metal spring. The three yellow LEDs above
the lever on which the response was emitted
flashed on for 0.1 s with each histamine or
saline injection.
Drugs
Histamine dihydrochloride (Sigma-Aldrich,
St. Louis, MO) was dissolved in sterile saline
(0.9% NaCl) solution (4.28 mg/ml). Heparin
solution was used to flush the catheters, before
and after the sessions. Methohexital was
administered on occasion to test catheter
patency.
Procedure
The house light was turned on at the start
of each experimental session. Subjects were
initially trained to press both levers with food
reinforcement (45-mg food pellets; Bio-Serv)
and performances were allowed to stabilize
under concurrent variable-interval (VI) schedules on the two levers. The VI schedule consisted of 20 interval values in a mixed
progression (Fleshler & Hoffman, 1962). Subsequently, the schedule for the lever on which
most responding occurred was changed to
extinction (EXT) and the schedule on the
alternate lever was set at VI 120-s. When
response rates were again stable the venous
catheters were implanted and after 7 days,
subjects were placed back on the concurrent
VI 120-s EXT schedule. Additionally, a VI 15-s
schedule of saline injection (consisting of
12 intervals arranged as those for food reinforcement) was introduced on the lever on

which responses continued to produce food
according to the VI 120-s schedule. All sessions
lasted 1 hr and were conducted at about the
same time daily in the afternoon. This training
phase lasted between 23 and 71 sessions (see
Table 1).
Histamine punishment. After three sessions
on the above schedule, response-produced histamine injections were introduced. Under this
condition, i.v. histamine injections (0.3 mg/
kg/injection) replaced saline and were delivered according to the VI 15-s schedule (conjoint VI 120-s [food], VI 15-s [histamine]). The
dose was selected based on pilot studies and
previously published papers (e.g. Podlesnik &
Jimenez-Gomez, 2013). This lever is hereafter
referred to as the punishment lever. If both a pellet and an injection were simultaneously scheduled for a response, the pellet was delivered
following that response and histamine was
delivered following the next response. The
schedule for responses on the alternate lever
remained EXT.
Punishment avoidance. Once performances
with histamine injections were stable from one
session to the next, a VI schedule of punishment avoidance replaced EXT on the alternate lever. Under this schedule, a single
response on the alternate lever (hereafter
referred to as the avoidance lever) canceled the
next histamine injection scheduled according
to the VI 15-s schedule on the punishment
lever. The response on the avoidance lever
cancelled only the presentation of the next
histamine injection scheduled on the punishment lever. Additional responses on the avoidance lever had no scheduled consequences
until the start of the next interval within the
punishment VI schedule. All scheduled histamine injections could therefore be avoided if
the subject made at least one response on the
avoidance lever within every interval of the VI
15-s schedule. This procedure was similar to
that used for electric shock (e.g., de Villiers,
1974), and has been referred to as a schedule
of deletion as distinct from a schedule of postponement (e.g. Perone & Galizio, 1987). The
sequence and number of sessions at which the
above conditions were executed are shown in
Table 1.
In subsequent sessions, the scheduled frequency of histamine injections was increased
systematically by decreasing the VI schedule
parameter from 15 s to 1 s, in order to

369

HISTAMINE PUNISHMENT AND NEGATIVE REINFORCEMENT
Table 1
Sequence of procedures during the first histamine punishment condition
Lever 1

Lever 2

Schedule 2

Histamine
dose
(mg/kg)

VI 180 (pellet)

-

-

VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)

VI 15 (inj)
VI 15 (inj)
VI 15 (inj)
VI 15 (inj)
VI 15 (inj)
VI 15 (inj)

0
0.3
0.3
0
0.3
0

Schedule 1

Sessions

Schedule 1

PM01

PM02

PM03

PM04

PM05

PM06

PM16

VI 180
(pellet)
Extinction
Extinction
Extinction
Avoidance
Extinction
Extinction
Extinction

1-45

01-46

1-39

1-41

1-36

1-44

1-13

46-71
72-74
75-77
78-83
84-89

47-59
60-62
63-67
68-73
74-79
80-86
87-89

40-68
69-71
72-75
76-81
82-86
87-90

42-67
68-71
72-73
74-79
80-84
85-88

37-67
68-70
71-74
75-79
80-84
85-89

45-64
65-67
68-72
73-78
79-83
84-87

14-23
24-26*

Note. The table displays a description of contingencies on each lever, the number of sessions conducted with each subject
on those conditions, and the dose of histamine used. A dose of “0” indicates saline injections.
* For this subject a VI 1-s schedule of saline injections was used.

increase the degree of potential reduction in
histamine injection frequency produced by
avoidance responses. It was anticipated that a
greater change in histamine injection frequency produced by a response would facilitate acquisition of responding on the
avoidance lever (e.g., Herrnstein & Hineline,
1966). Initially sessions with the VI schedule of
histamine cancellation alternated with sessions
in which responses on the avoidance lever had
no effect (EXT; Table 2). The EXT conditions
were later eliminated in an attempt to facilitate acquisition of responding on the avoidance lever. As the longer punishment VI
values proved ineffective in establishing avoidance responding, some subjects were studied

only at the shorter punishment VI-schedule
parameters (Table 2).
Once performances stabilized with the VI 1s schedule of histamine cancellation, the
scheduled contingencies on the two levers
were reversed. This change ensured that
response rates on the newly designated avoidance lever were sufficiently high to actually
reduce the frequency of histamine injections.
As a result, the relatively high response rate
on the avoidance lever decreased the frequency of histamine injections. Because results
following the reversal in lever contingencies
varied among subjects (see Results) reversals
were conducted repeatedly with several subjects (Table 3). As responding appeared to be

Table 2
Sequence of histamine-punishment conditions with variations in VI schedule parameter
Lever 1

Lever 2

Sessions

Schedule 1

Histamine
Schedule 2 dose (mg/kg) Schedule 1 PM01

VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)
VI 120 (pellet)

VI 7.5 (inj)
VI 7.5 (inj)
VI 4 (inj)
VI 4 (inj)
VI 2 (inj)
VI 1 (inj)

0.3
0.3
0.3
0.3
0.3
0.3

Extinction
Avoidance
Extinction
Avoidance
Avoidance
Avoidance

PM02
90-93
94-98

PM03

91-93
94-96
97-99
90-93
100-105
94-97 99-100
106-110
98-109 103-107* 111-124*

PM04
89-91
92-94
95-97
98-103
104-107
108-119

PM05
90-93
94-98
99-103
104-124*

PM06 PM16
88-90
91-93

27-33

Note. Details are as described for Table 1.
* Sessions 101-102 for PM02, 117-118 for PM03, and 114-115 for PM02, PM03, and PM05, respectively, were sessions following catheter replacement in which saline was used in place of histamine for the reintroduction of the subject to the
experiment and are not displayed in the table.

370

PAULO CÉSAR MORALES MAYER et al.
Table 3
Sequence of procedures for the exchange of contingencies between the levers
Lever 1

Schedule 1

Lever 2

Sessions

Histamine
dose
Schedule 2 (mg/kg) Schedule 1 Schedule 2

VI 120 (pellet) VI 1 (inj)
Avoidance
-

0.3
0.3/0

VI 120 (pellet) VI 1 (inj)

0.3/0.56

Avoidance

0.3/0.56

-

VI 120 (pellet) VI 1 (inj)
Avoidance
-

0.3/0.56
0.3

Avoidance
VI 120
VI 1 (inj)
(pellet)
Avoidance VI 120
VI 1 (inj)
(pellet)
Avoidance VI 120
VI 1 (inj)
(pellet)

PM01

PM02

PM03

PM04

107-109 105-107 122-124 117-119
110-118 108-123 125-135 120-128
119-124*
136-146 129-134
135-138*
125-131
147-152 139-143*

PM05

PM16

122-124 31-33
125-127 34-35
128-132 37-41
133-140 42-49

153-168 144-146* 141-149 50-56
150-157

Note. Details are as described for Table 1.
* indicates sessions with the alternative dose listed in the Histamine dose column.

maintained after the exchange in contingencies, particularly in one subject (PM01), saline
injections were substituted for histamine injections to assess whether those response rates
were maintained by the cancellation of histamine injections.
Results
During the VI 120-s schedule of food presentation (with the VI 15-s schedule of saline
injection) response rates for most subjects stabilized at about 10 to 15 responses per min.
Response rates for PM02 stabilized at about
6 to 7 and for PM16 at about 25 responses per
min (Fig. 1, squares on the first panels labeled
“Sal”) and generally occurred at a stable rate
across the session. Response rates on the eventual avoidance lever stabilized at about 0.5 to
1.5 responses per min (Fig. 1, triangles).
With the introduction of response-produced
histamine injections on the punishment lever,
response rates decreased to between 50%
(PM02 and PM04) and 82% (PM06) of rates
obtained with response-produced saline injections (Fig. 1, histamine panels). During subsequent sessions, response rates recovered for
every subject to different extents, though suppression remained between 30% (e.g., PM02)
and 63% (e.g., PM01) of the response rates
prior to the introduction of histamine injections. With these different amounts of
response suppression there were no substantial differences in the overall rate of food

pellet delivery. The average of 29 (0.4) pellets per session decreased by 1.83 (1.7) when
histamine
injections
stably
suppressed
responding.
With the introduction of the avoidance contingency, the rate of responding on the punishment lever did not change systematically
across subjects (Fig. 1). Response rates on the
avoidance lever remained at low levels for all
subjects throughout this phase with the exception of some transient responding during the
first two sessions for PM04 and PM06. The
mean frequency of histamine injections did
not substantially change from that obtained in
the previous phase without the avoidance
contingency.
Replacing histamine with saline increased
responding on the punishment lever for all
subjects except PM04 (Fig. 1, second saline
panels). With PM02 the initial increase was
not sustained in subsequent sessions. The reintroduction of histamine injections decreased
response rates on the punishment lever
(Fig. 1, second histamine panels). The
decrease was only transient with PM02, which
was followed by a small increase when saline
was introduced once again (Fig. 1; third saline
panel). Throughout all of these subsequent
conditions rates of responding on the avoidance lever remained low and not appreciably
different across conditions (Fig. 1).
In an attempt to increase the rate of responding on the avoidance lever, the punishment
and avoidance VI parameter was systematically

HISTAMINE PUNISHMENT AND NEGATIVE REINFORCEMENT

371

Fig. 1. Rates of responding on the punishment lever (squares), on the avoidance lever (triangles), and rates of injections (circles) under various schedule conditions as described in Table 1. Sal – Punishment lever: Conjoint VI 120-s (pellets), VI 15-s (saline injections), Avoidance lever: EXT; Hist – Punishment lever: Conjoint VI 120-s (pellets), VI 15-s
(0.3 mg/kg histamine injections), Avoidance lever: EXT; Avoid – Punishment lever: Conjoint VI 120-s (pellets), VI 15-s
(0.3 mg/kg histamine injections), Avoidance lever: VI 15-s (histamine injection cancellation).

varied in the next series of conditions (Fig. 2).
The left-most disconnected points show
response rates on the two levers and numbers
of injections when saline injections were scheduled with and without the avoidance contingency under a VI 15-s schedule. Rates of
responding on the punishment lever were
decreased below saline levels with histamine
injections scheduled under a VI 15-s schedule
providing a nominal rate of four injections/
min (Fig. 2, top panel). Rates decreased further
at a nominal rate of nine injections/min
(VI 7.5-s), with further increases in scheduled
frequency having no greater effect on response
rates. Response rates on the punishment lever
were not appreciably different with or without
the avoidance contingency (Fig. 2, top, compare squares and circles).

Response rates on the avoidance lever were
uniformly lower across all VI parameters with
histamine injections than with vehicle injections (Fig. 2, middle panel). Rates of responding on the avoidance lever were not different
with or without the histamine avoidance contingency (Fig. 2, middle panel) and did not
vary appreciably with changes in the nominal
injection rate. The frequency of histamine
injection (Fig. 2, bottom panel) was decreased
compared to the frequency of vehicle injections, though that frequency was not different
with or without the avoidance contingency
(Fig. 2, bottom panel).
Because response rates on the avoidance
lever were uniformly low at the introduction
of the avoidance contingency, there were
fewer opportunities for behavior to come in

372

PAULO CÉSAR MORALES MAYER et al.

contact with the avoidance contingencies than
there would have been had those rates been
higher. To assess the potential of a greater initial incidence of actual avoidances of histamine punishment on the maintenance of
responding on the avoidance lever, the contingencies on the two levers were exchanged.
Response rates on the two levers during the
first session after their contingencies were
exchanged were similar (Fig. 3). Additionally,

Fig. 2. Legend on next coloumn.

the frequencies of histamine injections during
the first sessions were uniformly reduced compared to the frequencies in the sessions before
the exchange. The similar rates of responding
on both levers in the first session were followed
by a progressive decrease in response rates on
the avoidance lever across the next several sessions, which, except for PM01 and PM02,
reached asymptotes that approximated those
before the exchange of contingencies, though
above zero. Further, the frequencies of histamine injections uniformly increased across sessions following the change to those obtained
before the exchange of contingencies. Similar
effects were obtained with PM03, PM04, PM05
and PM16 with subsequent exchanges of the
contingencies scheduled on the two levers. In
all of the 19 instances across subjects in which
the contingencies on the two levers were
exchanged, response rates on the avoidance
lever were increased, and in all but three
instances the number of injections delivered in
the first session was below that obtained in the
immediately preceding session.
For PM01 the response rate on the avoidance lever stabilized at a rate above that
obtained before the reversal of contingencies
on the two levers. To assess if avoidance of histamine was maintaining the low response rates
on the avoidance lever, histamine injections
were replaced with saline for this subject

Fig. 2. Rates of responding on the punishment lever
(top panel), on the avoidance lever (central panel), and
rates of injections (bottom panel) with variations in the
scheduled rate of histamine injection (punishment VI
schedule parameter) with (squares) or without (circles)
the VI 15-s histamine cancellation contingency in effect.
The various schedule conditions are as described in
Table 2. Disconnected points above “Veh” represent rates
when saline was injected rather than histamine (0.3 mg/
kg/injection). Data points correspond to the average of all
subjects exposed to the condition, with the values for individual subjects obtained from averages of the last three sessions from each condition and error bars corresponding
to the standard error of the mean. Statistical analysis of
the rate of punished responses indicated significant effect
of scheduled injection rate (F = 21.0; p < .001), but no
effect of the presence of the avoidance contingency
(F = 0.018; p = .894). Statistical analysis of the rate of
avoidance responses indicated significant effect of scheduled injection rate (F = 9.96; p < .001), but no effect of
the presence of the avoidance contingency (F = 2.82;
p = .111). Statistical analysis of the frequency of injection
indicated significant effect of scheduled injection rate
(F = 4.45; p = .005), but no effect of the presence of the
avoidance contingency (F = 0.171; p = .684).

HISTAMINE PUNISHMENT AND NEGATIVE REINFORCEMENT

373

Fig. 3. Effects of exchanges of contingencies between the levers on response rates and injection frequencies over consecutive sessions. Schedules: Punishment lever: Conjoint VI 120-s (pellets), VI 1-s (0.3 mg/kg histamine injections),
Avoidance lever: VI 15-s (histamine injection cancellation). BL: baseline, corresponds to the last three sessions before
the first exchange of contingencies. Saline panel for PM01 indicates sessions in which histamine was replaced by saline.
Under all other conditions, except as noted, histamine injections were 0.3 mg/kg/injection. For PM04 the histamine
dose was 0.56 mg/kg/injection for sessions shown in the panel labelled “0.56.” That dose was used during subsequent
sessions shown in the following panels, with the asterisk indicating a single session at a dose of 0.3 mg/kg/injection.

(Fig. 3, PM01 “Saline” panel). Response rates
on both the punishment and avoidance levers
increased with saline substitution for histamine. Response rates decreased again when
histamine was reintroduced.
For PM04 after the second exchange of contingencies the histamine dose was increased to
0.56 mg/kg/injection. Although the increase
in dose reduced responding on the punishment lever, the low rates on the avoidance
lever were not appreciably affected. Further
exchanges of contingencies, with the increased
dose, did not result in acquisition of responding on the avoidance lever (Fig. 3, PM04, three
rightmost panels). A single session decrease in
dose for PM04 (Fig. 3, session denoted with an
asterisk) increased rates of responding on the
punishment lever and the frequency of injections. During that session response rates on the
avoidance lever decreased.
Cumulative records of performances shown
in Figure 4 are from one subject (PM04) that
showed a progressive decrease in response
rates on the avoidance lever after the

exchange and one subject (PM01) that
showed a continued maintenance of responding over successive sessions after the exchange
in contingencies on the two levers. The top
panel for each column shows records from the
last session before the exchange of contingencies, with high rates of responding on the punishment lever (the record with tick marks)
and low rates on the avoidance lever (the
record without tick marks). The middle panel
shows the first session with the exchanged contingencies and the increase in rates of
responding on what was in that session for the
first time the avoidance lever. The increase in
responding on the avoidance lever was most
pronounced initially and decreased as the session progressed. The lower density of tick
marks on the event line for the records displayed in the middle panel shows the reduction in the rate of histamine injection
produced by the increased responding on the
avoidance lever. The bottom panels show performances in the ninth (PM01) and the third
sessions after an exchange of contingencies

374

PAULO CÉSAR MORALES MAYER et al.

Fig. 4. Cumulative records of responding of subjects PM01 (left) and PM04 (right) during sessions before and after
the exchange in contingencies. Cumulative curves with tick marks correspond to responses on the punishment lever,
cumulative curves without tick marks indicate responses on the avoidance lever. Panels A and D show data from the last
session before the exchange of contingencies between the levers; Panels B and E show data for the first session after the
exchange of contingencies; and Panels C and F show data from the last session of the exchange condition. Tick marks
on the bottom, event line indicate the occurrence of histamine injections, tick marks on the cumulative response line
indicate pellet deliveries. Schedules were those described for Figure 3.

(PM04) when response rates for PM04 on the
avoidance lever approached those observed
before the exchange in contingencies, whereas
those for PM01 remained substantially above
those before the exchange in contingencies,
particularly at the beginning of the session.

Discussion
As in previous studies (e.g. Goldberg, 1980;
Podlesnik et al., 2010; Sharpless, 1961) the
introduction of response-dependent histamine
injections suppressed behavior that produced

HISTAMINE PUNISHMENT AND NEGATIVE REINFORCEMENT
the injections. Further, when saline replaced
histamine injections, rates of responding
recovered to those similar to, or approaching,
those obtained prior to the introduction of
histamine injections. For instances in which
the recovery of responding was incomplete,
the reintroduction of histamine produced an
immediate suppression that was either transient or sustained over subsequent sessions.
Further, the number of food pellets received
under the VI 120-s schedule did not substantially change, suggesting that the suppression
of responding was due to punishment by histamine injections and not due to a decrease in
frequency of reinforcement.
One possible problem when using drugs as
punishing stimuli is the potential for “direct”
effects of the accumulated drug to interfere
with continued responding. Direct effects
refer to those effects of drugs on behavior that
occur independently of the contingency relation between the response and the drug injection.
These
effects
may
complicate
interpretations of the effects of the contingency between responses and injections, and
traditionally have been of concern in studies
of the reinforcing effects of drugs (for a
detailed discussion see Katz, 1989), but may
also apply in the present studies.
Absent a comparison of response-dependent
and -independent scheduling of histamine
injections, several considerations argue that
the effects of histamine in this and previous
studies were due to the contingent relation
between responses and injections rather than
a direct effect of histamine. In the initial studies (Goldberg, 1980; Katz & Goldberg, 1986)
of punishment with histamine injections using
monkeys, intercalated periods in which
responding was not punished were scheduled,
with little to no evident effect of accumulated
injections on nonpunished responding. Further, in situations in which histamine suppressed food-reinforced responding in rats,
concurrent nonpunished responses maintained by food reinforcement were emitted
reliably showing no evidence of suppression
(Podlesnik & Jimenez-Gomez, 2013; Podlesnik
et al., 2010). A study of histamine injected
intraventricularly found a decrease in rates of
conditioned avoidance responding (Tasaka,
Kamei, Akahori, & Kitazumi, 1985). However,
a study of the histamine precursor, L-histadine
injected systemically, found no effect on

375

avoidance responding (Gerald & Maickel,
1972). Another study found that, unlike
punishing effects, the effects of centrally
administered histamine were insensitive to
administration of histamine antagonists of
either subtype (Calcutt & Reynolds, 1976),
indicating that the effect was a by-product of
central administration and, in contrast to punishing effects (e.g. Goldberg, 1980; Katz &
Goldberg, 1986; Podlesnik & Jimenez-Gomez,
2013), was not mediated by histamine receptors. Taken together, the results to date suggest virtually no direct effects of systemically
administered histamine.
It is interesting to note that in the present
study histamine was effective as a punishing
stimulus in suppressing behavior but ineffective in establishing an avoidance response.
The absence of negative reinforcing effects in
this specific situation highlights that consequences that suppress behavior do not by
necessity function to maintain behavior that
avoids their presentation. This finding is
consistent with an emphasis on the situationspecific nature of the effects of behavioral consequences. As noted by Morse and Kelleher
(1977), there is a tendency to emphasize the
consequent event and to ignore the importance of the contingency relationships that
determine subsequent behavior. Effects of
many routinely studied environmental events
are often presumed, though it is preferable to
base categorizations of stimulus events on the
manner in which they change behavior.
The outcomes of the present study are relevant to historical concepts of punishment in
which two factors are hypothesized to be
involved. In one version of the hypothesis
(e.g. Estes, 1944; Skinner, 1938), responses elicited by the punishing stimulus or stimuli
paired with it, interfere or otherwise compete
with continued emission of the punished
response, resulting in its suppression. As noted
above however, studies of histamine have provided little to no evidence of overt elicited
behavioral responses (Kuraishi et al., 1995;
White & Rumbold, 1988). Absent some as yet
undocumented histamine-elicited response
competing with the punished response, this
version of the two-factor hypothesis is broadly
untenable. However, as eliciting functions of
electric shock are well documented (Flaherty,
1985), it remains possible that this version of
the hypothesis applies more narrowly to those

376

PAULO CÉSAR MORALES MAYER et al.

punishing stimuli, such as electric shock, used
in situations in which responses are reliably
elicited.
In another variant of a two-factor hypothesis, responses other than the punished response are
negatively reinforced if this emission competes
with the punished response, thereby decreasing the frequency of the punishing stimulus,
or stimuli paired with it (e.g. Dinsmoor, 1954;
Skinner, 1953; Solomon, 1964). Despite punishing effects observed on introduction of histamine injections in the present study,
avoidance responding on a second lever was
not stably maintained at levels greater than
those obtained with vehicle. With the initial
introduction of the avoidance contingency the
rate of responding on the avoidance lever was
exceedingly low, with no acquisition of that
response evidenced over subsequent sessions.
With the contingencies on the levers
exchanged, response rates on the avoidance
lever were initially sufficient for contact with
the avoidance contingencies, and the frequency of histamine injection initially
decreased. However, over the course of successive sessions, responding on the avoidance
lever extinguished. Further, repetitions of the
exchange were no more effective in producing
a sustained increase in avoidance responding,
and an increase in histamine dose for PM04
did not increase rate of responding on the
avoidance lever. Thus, the conditions which
were sufficient to produce suppression were
not sufficient for the maintenance of avoidance responding on a second lever, suggesting
that the observed punishment suppression was
not dependent upon the acquisition of the
avoidance response.
It remains possible that some unobserved
avoidance response was conditioned before
the introduction of the explicitly programmed
avoidance contingency, conserving a twofactor interpretation of the punishing effects
of histamine. Casual observations did not indicate any such responses, with that identification being an essential first step for testing this
version of the hypothesis. When the other
response is unspecified, or is as ill-defined as
“any behavior other than the punished
response,” the hypothesis both lacks parsimony and is untestable. If a prior avoidance
response had been established it would
unlikely be as efficient in reducing the frequency of histamine injections as responses on

the avoidance lever, especially when contingencies on the two levers were exchanged. As
shown, those elevated response rates on the
avoidance lever, rather than being maintained,
extinguished over successive sessions.
The present procedure by design omitted a
change in exteroceptive stimulus conditions
contingent on effective avoidance responses. A
similar tack was taken by Arbuckle and Lattal
(1987) in which punishment cancellation was
contingent on a period of response omission,
which resourcefully exploits its compatibility
with operant response suppression. Consequently, no specific training was necessary for
maintenance of the avoidance response, and
the results were consistent with a two-factor
hypothesis of punishment. The present study
similarly programmed a punishment avoidance contingency for a specific response,
though on a separate lever (cf. Catania, 1973).
It is possible that some as yet unidentified
aspect of the experimental situation specifically interfered with acquisition of lever-press
avoidance of punishment. Mitigating against
this interpretation is that several previous studies have demonstrated the maintenance of
one response with avoidance of consequences
arranged for a separate response (e.g.,
Morse & Herrnstein, 1956; van Haaren &
Zarcone, 1994; Thomas, 1968). Further,
escape from histamine injections previously
has been successfully trained and maintained
in rhesus monkeys under different conditions
(Takada et al., 1986). Possibly most important
is that when the contingencies on the levers
were exchanged in the present study, responding occurred on the avoidance lever at rates
sufficient to decrease the frequency of histamine injections, indicating that the two
responses were not incompatible.
Though rates of responding on the avoidance lever for most subjects extinguished with
the exchange in contingencies on the two
levers, those of PM01 and PM02 decreased to
a lower rate, but were not entirely eliminated,
an effect most prominent with PM01. That
outcome suggested the possibility that the
residual low rates of responding on the avoidance lever were indeed maintained by avoidance of histamine injections. However,
substitution of saline for histamine with PM01
resulted in an increase in response rates on the
avoidance lever, and reintroduction of histamine injections again suppressed response

HISTAMINE PUNISHMENT AND NEGATIVE REINFORCEMENT
rates. Had responding been maintained by
avoidance of histamine injections, a decrease in
response rates would be expected when saline
was substituted for histamine. These results
suggest the possibility that the responding of
PM01 on the avoidance lever was adventitiously maintained, possibly through chaining
of the responses on the two levers.
Further evidence contrary to the two-factor
hypothesis would have been provided if avoidance of histamine injections had been trained
alone prior to the implementation of punishment, and not subsequently maintained solely
by avoidance of histamine punishment. On
one hand, that procedure could demonstrate
only that histamine punishment and avoidance can occur concurrently, but would not
speak to the necessary and sufficient conditions for punishment. On the other hand,
punishment in the absence of avoidance is
one true logical test of the avoidance account
of punishment.
In summary, the present study found effective punishment by histamine injection without evidence of avoidance of the punishing
stimulus. As in previous studies (Azrin et al.,
1965; Lattal & Cooper, 1969) an added stimulus change may have facilitated acquisition of
the avoidance response under the current
conditions. The present study, however,
addresses whether the decrease in frequency
of the punishing stimulus alone is sufficient
for the maintenance of avoidance responding.
The present results indicate that conditions
sufficient for histamine punishment were not
sufficient for the maintenance of an explicit
punishment-avoidance response, an outcome
indicating the absence of a necessary general
relation between avoidance and histamine
punishment.
References
Arbuckle, J. L., & Lattal, K. A. (1987). A role for negative reinforcement of response omission in punishment. Journal of the Experimental Analysis of Behavior,
48(3), 407–416. https://doi.org/10.1901/jeab.1987.
48-407
Azrin, N. H., Hake, D. F., Holz, W. C., & Hutchinson, R. R.
(1965). Motivational aspects of escape from punishment. Journal of the Experimental Analysis of Behavior,
8(1), 31–44. https://doi.org/10.1901/jeab.1965.8-31
Azrin, N. H., & Holz, W. C. (1966). Punishment. In
W. K. Honig (Ed.), Operant behavior: Areas of research
and application (pp. 380–447). New York: AppeltonCentury-Crofts.

377

Barker, D. J., Sanabria, F., Lasswell, A., Thraikill, E. A.,
Pawlak, A. P., & Killeen, P. R. (2010). Brief light as a
practical aversive stimulus for the albino rat. Behavior
and Brain Research, 214(2), 402–408. https://doi.
org/10.1016/j.bbr.2010.06.020
Branch, M. N., Nicholson, G., & Dworkin, S. I. (1977).
Punishment-specific effects of pentobarbital: Dependency on the type of punisher. Journal of the Experimental Analysis of Behavior, 28(3), 285–293. https://doi.
org/10.1901/jeab.1977.28-285
Calcutt, C. R., & Reynolds, J. (1976). Some behavioural
effects following intracerebroventricular (ICV) injection in rats of histamine H 1- and H 2-receptor agonists and antagonists. Neuroscience Letters, 3(1), 82–83.
https://doi.org/10.1016/0304-3940(76)90112-9
Carvalho Neto, M. B., Maestri, T. C., Tobias, G. K. da S.,
Ribeiro, T. C., Coutinho, E. C. N. N.,
Miccione, M. M., … Moreira, D. (2005). O jato de ar
quente como estímulo punidor em rattus norvegicus.
Psicologia: Teoria e Pesquisa, 21(3), 335–339. https://
doi.org/10.1590/S0102-37722005000300010
Catania, A. C. (1973). Self-inhibiting effects of reinforcement. Journal of the Experimental Analysis of Behavior,
19(3), 517–526. https://doi.org/10.1901/jeab.1973.
19-517
Catania, A. C. (2013). Learning (5th ed.). Cornwall-onHudson, NY: Sloan Publishing.
Church, R. M. (1969). Response suppression. In
B. A. Campbell & R. M. Church (Eds.), Punishment
and aversive behavior (pp. 111–156). New York: Appleton-Century-Crofts.
de Villiers, P. A. (1974). The law of effect and avoidance:
A quantitative relationship between response rate and
shock-frequency reduction. Journal of the Experimental
Analysis of Behavior, 21(2), 223–235. https://doi.
org/10.1901/jeab.1974.21-223
Dinsmoor, J. A. (1954). Punishment: I The avoidance
hypothesis. Psychological Review, 61, 34–46. https://doi.
org/10.1037/h0062725
Dinsmoor, J. A. (1998). Punishment. In W. T. O’Donohue
(Ed.), Learning and behavior therapy (pp. 188–204).
Needham Heights, MA: Allyn & Bacon.
Estes, W. K. (1944). An experimental study of punishment.
Psychological Monographs, 57, 1–40. https://doi.org/10.
1037/h0093550
Everly, J. B., & Perone, M. (2012). Suppressive and facilitative effects of shock intensity and interresponse times
followed by shock. Journal of the Experimental Analysis of
Behavior, 98(3), 311–340. https://doi.org/10.1901/
jeab.2012.98-311
Ferster, C. B. (1958). Control of behavior in chimpanzees
and pigeons by time out from positive reinforcement.
Psychological Monographs: General and Applied, 72(8),
1–38. https://doi.org/10.1037/h0093787
Flaherty, C. F. (1985). Animal learning and cognition.
New York: McGraw-Hill, Inc.
Fleshler, M., & Hoffman, H. S. (1962). A progression for
generating variable-interval schedules. Journal of the
Experimental Analysis of Behavior, 5(4), 529–530.
https://doi.org/10.1901/jeab.1962.5-529
Friedel, J. E., DeHart, W. B., & Odum, A. L. (2017). The
effects of 100 dB 1-kHz and 22-kHz tones as punishers
on lever pressing in