Executive Monkey Study

Full text

E
Executive Monkey Study
Jay M. Weiss
Emory University, School of Medicine, Atlanta,
GA, USA
Historical Development of
“Psychosomatic Medicine”
The “executive monkey”study, which was
published over 60 years ago in 1958, remains in
all probability the most well-known experiment in
afield referred to as Psychosomatic Medicine.
Given its prominence in this field, we will there-
fore consider the definition and historical devel-
opment of the concept of Psychosomatic
Medicine before describing the details of the
“executive monkey”experiment and its findings.
That mental or psychological events could influ-
ence disease was known as far back as the early
Greeks, as Hippocrates and Galen proposed that
mind and body were inseparably linked and that
physical wellbeing was tied to psychological
wellbeing (e.g., Coxe 1846). Similar sentiments
were expressed during the centuries that followed,
so that by the eighteenth and nineteenth centuries
one finds definitive pronouncements that psycho-
logical factors profoundly influence disease. For
instance, Daniel Hack Tuke, the prominent British
physician and psychiatrist, published a volume in
1872 titled “Illustrations of the Influence of the
Mind Upon the Body in Health and Disease”
(Tuke 1872) in which he described thought pro-
cesses and emotions both causing/exacerbating
disease as well as ameliorating it (for a detailed
quotation from Tuke in this regard, see introduc-
tory section Weiss 1972). The relationship
between psychological processes and bodily func-
tion (symptoms and disease) soon received new
attention with the studies and writings of Sigmund
Freud at the end of the nineteenth century when he
hypothesized that psychic conflicts could find
expression as physical disabilities in hysteria
(Freud, 1890–1893; translated by Strachey,
1955).
The early to mid-twentieth century saw a
renaissance of interest in what came to be known
as psychosomatic medicine –the relationship of
psychological events to somatic disease. An
excellent recounting of this history can be found
in an article by Lipowski (1984). He finds the term
“psychosomatic medicine”to have been first used
in 1922. Important events thereafter were the pub-
lication in 1935 of Helen Flanders Dunbar’s book
titled, “Emotions and Bodily Changes: A Survey
of Literature on Psychosomatic Interrelation-
ships”(Dunbar 1935), followed by the founding
of the Journal “Psychosomatic Medicine”in
1939. Psychoanalytic analysis introduced by
Freud reached perhaps its highest development
in relation to somatic disease in the work and
writings of Franz Alexander, who saw a variety
of different diseases as resulting from specific
intrapsychic conflicts (Alexander 1950). In the
case of peptic ulcer, Alexander posited that this
#Springer International Publishing AG 2018
J. Vonk, T.K. Shackelford (eds.), Encyclopedia of Animal Cognition and Behavior,
https://doi.org/10.1007/978-3-319-47829-6_251-1

pathology resulted from the ulcer patient’s uncon-
scious desire to be dependent and taken care of
while at the same time adhering to the compulsion
to act independently and assertively, with the
unconscious desire resulting in the need to be
constantly fed causing chronic hypersecretion of
stomach acid and eventual ulceration. Needless to
say, it is difficult, if not impossible, to see how this
intrapsychic conflict could explain the duodenal
ulcers that were found in the “executive”mon-
keys, and as the result of challenges such as this as
well as other psychoanalytic explanations were
soon viewed as conjectural and unlikely to be
correct. Meanwhile, controversy raged within
Psychosomatic Medicine as to whether mental
events could actually cause disease (called “psy-
chogenesis”) or whether mind and body were
really inseparable and that disease resulted from
physiological processes necessarily involving
interplay of psychological and somatic factors.
While the field has settled on the latter view
(as evidenced in the Lipowski article), even at
the present time the term “psychosomatic”is com-
monly used to mean that psychological influences
lead to disease. For example, in 2016, author
Claudia Kalb described Charles Darwin’s con-
tinuing gastrointestinal problems, asking whether
these problems might have been caused by an
infectious tropical bug picked up in his travels
aboard the ship Beagle or “were Darwin’s lifelong
symptoms psychosomatic –physical manifesta-
tions of ongoing mental stress?”(Kalb 2016).
The “Executive Monkey”Experiment:
Description and Findings
Turning now to the “executive monkey”experi-
ment, this experiment utilized pairs of rhesus
monkeys, ultimately four pairs of monkeys. The
two monkeys of each pair were restrained in
chairs as shown in Fig. 1. The monkeys were
subjected to intermittent electric shocks that one
member of each pair, shown at the left side of the
figure, could control by pressing a lever in front of
it which postponed occurrence of the next shock,
so that shock only occurred when this monkey,
dubbed the “executive monkey,”failed to activate
its lever. The second member of the pair also had a
lever mounted at the front of its chair but the lever
was inactivated so that this monkey received the
same shocks as did the “executive”but did not
have any control over shock occurrence. Thus,
one monkey of the pair exerted control over the
occurrence of shock while the other monkey,
called the “yoked”monkey, had no control, with
both monkeys receiving the same shocks through-
out the experiment.
This experiment attracted so much attention
because, when the monkeys were maintained in
these experimental conditions for many days, the
“executive”monkeys eventually died apparently
as a result of gastrointestinal pathology including
lesions and perforating ulceration, while the
“yoked”monkeys developed no such pathology
and lived on. It thus appeared that being the
“executive”and exerting control caused pathol-
ogy and death, whereas not exerting control
avoided this detrimental outcome.
The history of how the experiment unfolded is
important for understanding and interpreting the
results. The first report of the procedure described
above appeared in the Journal “Psychosomatic
Medicine”in 1958 in an article for which
R.W. Porter was the first author (Porter et al.
1958). Titled “Some Experimental Observations
on Gastrointestinal Lesions in Behaviorally Con-
ditioned Monkeys,”the article describes that in
various behavioral studies being undertaken with
rhesus monkeys at Walter Reed Army Medical
Center in Washington, DC, gastrointestinal
pathology leading to death began to show up in
many of the animals. The procedures performed
on monkeys confined in chairs as shown in Fig. 1
involved conditioned anxiety (shock paired with a
signal) and punishment (shock given for
performing a response otherwise rewarded with
a sugar pellet), as well as a combination of such
procedures. Eleven of nineteen animals undergo-
ing such procedures died during the experiments,
and on autopsy gastrointestinal pathology was
discovered as the likely cause of death.
The paper by Porter et al. concludes with a
description of the first two pairs of monkeys that
eventually constituted the “executive”monkey
study. In an attempt to determine that behavioral
2 Executive Monkey Study

conditions led to the pathology seen in the 11 mon-
keys described above –which the investigators
strongly believed to be the case –and perhaps gain
insight into what conditions could do this, two
monkeys were subjected to a shock avoidance
paradigm, this being the lever press contingency
whereby these two animals were made to avoid/
control shock by pressing the lever at the front of
their chair. The avoidance procedure, however,
leads to animals receiving electric shocks, so that
any pathology eventually observed could be sim-
ply a result of the physical stressor of electric
shock being given to the monkeys and not because
of the behavioral condition of performing an
avoidance response. In order to rule out the pos-
sibility that electric shocks alone might produce
the pathology eventually seen, this aspect of the
experiment included a “yoked”animal for each
avoidance monkey, these yoked animals having
no control over shock and not engaging in avoid-
ance responding but receiving the same shocks as
the monkeys performing the avoidance response.
If these yoked animals did not develop pathology,
it would be evident that the electric shock itself
did not cause such pathology. It is of note, there-
fore, that the “executive”monkey experiment was
not designed to determine the effects of having
control versus not having control in a stressful
situation, as subsequent scientific and popular
perception seized upon, but included yoked ani-
mals (i.e., animals having no control) to rule out
electric shock as the cause of pathology. That the
yoked animal was included in the study to exclude
electric shock as being responsible for develop-
ment of pathology is explicitly stated in the
Executive Monkey Study, Fig. 1 A pair of rhesus mon-
keys in chairs as used in the “executive monkey”study.
The monkey on the left is controlling (avoiding) delivery
of shock to its feet by pressing the lever mounted at the
front of its chair, while the monkey on the right (yoked
monkey) receives the same shocks as the avoidance, or
“executive,”monkey at the right. The yoked monkey also
has a lever mounted at the front of its chair, but this lever is
inactive and the yoked monkey is not attending to it. As
shown in Brady (1958) Scientific American article
Executive Monkey Study 3

subsequent summary article that appeared in Sci-
entific American (Brady 1958).
A second important paper appeared later that
year in which findings from “executive”-yoked
pairs reached n=4 (eight monkeys); similar
results were seen with the addition of two pairs
of monkeys as reported in the first paper described
above. Thus, in the total of four pairs, which
constituted all of the animals eventually reported
for the “executive monkey”study, all four “exec-
utive”monkeys died of gastrointestinal pathology
whereas the four yoked animals showed no
behavioral impairment, and, when sacrificed,
showed no G.I. pathology (Brady et al. 1958).
Following this, the most famous article describing
these results –“Ulcers in ‘Executive’Monkeys”–
appeared in Scientific American and presented the
data from the four pairs of monkeys (Brady 1958).
That receipt of electric shock was not the cause
of the pathology seen in these monkeys becomes
even more evident when the schedule for delivery
of shock and the avoidance behavior it generated
is examined (Brady et al. 1958; Porter et al. 1958).
The avoidance schedule to which the “executive”
monkeys were subjected is called a “Sidman
avoidance schedule”named after its developer,
Dr. Murray Sidman. In this schedule, shocks
were not preceded by any warning signal, and
brief shocks (5.0 mA intensity, 0.5 s in duration)
were given every 5.0 s (shock delivered to the feet
of the animals) unless the active lever is pressed in
which case the next shock was delayed for 20 s.
Thus, unsignaled shocks occurred in a train of one
brief shock every 5.0 s unless an avoidance
response was made which then delayed the onset
of the next train of shocks for 20 s. When this
schedule was in place, “executive”monkeys (i.e.,
avoidance monkeys) proved remarkably profi-
cient at acquiring and practicing the lever press
response to avoid (i.e., postpone) shock. The
investigators report that within a few hours of
the beginning of training, stable avoidance
responding emerged, the avoidance monkey mak-
ing approximately 15–30 responses per minute.
This response rate remained essentially
unchanged during the entire experiment. As a
result, shocks received by the monkeys never
exceeded 2 per hour and typically averaged less
than 1 per hour. Thus, very few shocks were
received by the monkeys throughout the course
of the experiment; shocks occurred so infre-
quently it is not surprising that the results revealed
(by inspection of yoked animals) that this physical
stressor did not generate gastrointestinal
pathology.
Describing the course of the experiment and
the development of pathology, the animals were
confined in the chairs as shown in Fig. 1continu-
ously. Each day of the study was divided into
“time-in”periods lasting 6 h during which shocks
would occur (and the executive [avoidance] mon-
key therefore was going to respond to control
shock and the yoked monkey also received
shock if the avoidance monkey did not avoid
shock) and “time-out”periods also lasting 6 h
when no shocks would occur. These periods alter-
nated, so that a 24-h day consisted of two 6-h
“time-in”periods and two 6-h “time-out”periods.
The presence of the “time-in”phase was signaled
by illumination of a red light in full view of both
animals of the pair. Regarding development of
gastrointestinal pathology and death of the “exec-
utive”animals, one died after 9 days of this pro-
cedure while the other three died after 23, 25, and
48 days. Pathology found in the “executive”mon-
keys was gastric hemorrhage and erosions as well
as duodenal ulcers that perforated in two
instances. When the yoked monkeys were
sacrificed shortly after the death of their matched
“executive”partners, yoked monkeys showed no
evidence of any gastric pathology.
In the Scientific American article by Brady
(1958), a brief description is given of other testing
conditions (described as being “still in progress”),
such as prevention of social interaction between
the two monkeys and exposure to different time-
in, time-out intervals for responding. However,
the details of these studies were not given nor,
more importantly, were these procedures and their
outcome described in any subsequent formal pub-
lication, so the number of monkeys that com-
prised the “executive monkey”study remains
four pairs of animals.
4 Executive Monkey Study

Subsequent Research: Being Able to
Control Found to Reduce Stress and
Pathology
Popular press and scientific researchers quickly
realized that the “executive”monkey study com-
pared animals that had control over a stressor (i.e.,
monkeys able to perform the avoidance response)
with yoked animals that did not have any such
control, and so the effects observed were viewed
as the consequences of exerting control versus not
having that ability, or, as the results suggested,
that responsibility/burden. The study was thence-
forth seen in this light, despite the fact that, as
explained above, the investigators who carried out
the experiment did not intend to study effects of
having control versus no control but, instead,
included yoked monkeys to set up what is called
a“control condition”to rule out the possibility
that shock alone produced pathology. Regardless,
the “executive monkey”study was thereafter seen
as demonstrating the effect of having control ver-
sus no control.
While the pathological findings of the “execu-
tive monkey”study seemed plausible given the
amount of responding by the avoidance monkeys
versus the nonresponding of the “yoked”mon-
keys, other data looking at the consequences of
having control versus noncontrol, though
extremely limited in amount and scope, did not
point to the same conclusion. In 1948, O. Hobart
Mowrer and his student P. Viek had found that rats
able to control (terminate) an electric shock by
jumping into the air showed less subsequent fear
(i.e., were more willing to approach and eat food)
than were rats that received equal shocks but
could not terminate them (Mowrer and Viek
1948). Though it was obvious that the conditions
of the “executive monkey”study and the experi-
ment by Mowrer and Viek were very different, the
contrasting findings suggested that the conse-
quences of having control versus no control
might not be described comprehensively by what
was seen in the “executive monkey”study. Fol-
lowing publication of the “executive monkey”
findings, however, no additional experimental
results looking at consequences of having control
versus no control appeared for approximately
10 years, but then new findings were reported at
this time.
In 1968, 10 years after publication of the
“executive monkey”study, the findings of my
Ph.D. dissertation research were published in
which rats were the subjects (Weiss 1968). These
data also pointed to a conclusion opposite to that
suggested by the “executive”monkey study, indi-
cating that control over a stressful event lessened
consequences of stress and stress-induced pathol-
ogy. This study examined effects of having con-
trol versus no control over electric shocks on
(a) body weight and food/water intake,
(b) fearfulness as measured by willingness of the
rats, when thirsty, to drink in the control versus no
control environment, and (c) development of gas-
tric pathology (gastric erosions). Using matched
triplets of rats in which one rat could avoid and
escape electric shocks (i.e., control occurrence
and duration of any shocks received), a “yoked”
animal that received the same shocks but had no
control over them, and a nonshock subject that
simply remained in the apparatus and received all
stimuli as the other two animals but did not
receive any shocks, differences between these
three types of animals were observed. The exper-
imental design for a triplet is shown in Fig. 2,
although the particular apparatus shown in this
figure was used in a later experiment and differs
somewhat from the apparatus employed in Weiss
(1968) (apparatus used in Weiss 1968 is shown in
that article).
A highly important detail in these studies,
which can be seen in Fig. 2, was that electric
shocks received by the rats were delivered
through electrodes affixed to the rat’s tail, a pro-
cedure prompted by the insight of my brilliant Ph.
D. advisor, Professor Neal E. Miller. (Neal Miller
was a member of the National Academy of Sci-
ences and first recipient of the National Medal of
Science in the field of Social and Behavioral Sci-
ence [1964], receiving this award in the third year
after the Medal was established by President John
F. Kennedy.) Miller had criticized the original
Mowrer and Viek experiment on the basis that
the animals were given shocks via a grid floor.
Miller pointed out that on a grid floor, rats can
change their bodily posture to lessen shock and/or
Executive Monkey Study 5

jump completely off the floor to momentarily
avoid shock, so that using this method to deliver
shock cannot insure that the two matched animals
(i.e., avoidance-escape animal and yoked animal)
will get exactly the same shocks. Consequently,
differences seen between these conditions might
be due to the animals learning different postural
maneuvers to consistently receive different
amounts of shock. When I originally proposed to
Miller to study control versus noncontrol for my
dissertation research, he called attention to this
problem and urged me to try to develop shock
electrodes that could be affixed to the animal’s
body to eliminate the ability of the rat to make
postural changes that would alter shock. Fortu-
nately, I was able to accomplish this during the
next few months, and the method used has been
published (Weiss 1967). Consequently, all exper-
iments using the triplets as described and shown
above employed tail electrodes affixed to the rats
for delivery of shocks. With each matched pair of
avoidance-escape and yoked animals having their
shock electrodes wired in series into the electrical
circuit for all of the experiments, shocks received
by these two animals were necessarily exactly the
same in number, duration, and intensity.
The results reported in Weiss (1968) were
briefly described as follows: in the first experi-
ment, animals were exposed to avoidance-escape
versus yoked conditions for a session lasting
2.5–3.0 h (i.e., avoidance-escape rats could
avoid tail shock or terminate it by jumping onto
a shelf while yoked animals received the same
shocks but could not control them). Twenty-four
hours later, avoidance-escape rats were found to
be significantly heavier in body weight than the
yoked animals, having eaten more food during the
night after the stress session than did yoked
Executive Monkey Study, Fig. 2 The experimental
design paradigm for matched triplets of animals as used
in Weiss (1968) and Weiss (1971a,b,c). The Avoidance-
escape animal shown at left could control shock by
performing a response (wheel-turn shown in this figure);
this is noted as connection of wheel “To shock control.”
Both the Avoidance-escape and the Yoked animal shown at
center which is unable to perform a response to affect
shock received exactly the same shocks via tail electrodes,
with their tail electrodes wired in series into the same
electrical circuit; electrode connection is noted as “To
shock source.”The third animal shown at right (“No
shock”) receives all stimuli throughout the procedure as
the other two animals of the triplet except that it receives no
shocks as its tail electrodes are not connected (“No con-
nection”). Apparatus shown here was used in Weiss
(1971a,b,c), while apparatus and responses to control
shock were different in Weiss (1968) (see text), but all
aspects of triplet design shown here were the same in all
of these experiments
6 Executive Monkey Study

animals. Yoked animals also weighed less than
matched no-shock animals, while avoidance-
escape animals were only slightly lighter than
the no-shock subjects. In the second part of this
study, all rats were made thirsty by water depriva-
tion and placed into the shock environment where
a water tube was now present. Yoked animals
were considerably more reluctant to approach
the drinking tube than were avoidance-escape or
nonshock subjects, thereby showing that yoked
rats, as a result of exposure to shocks over which
they had no control, were now more fearful in that
environment than the other two groups. Finally, to
assess gastric pathology as was measured in the
“executive monkey”study, a second experimental
situation was used in which animals were mildly
restrained in a tube wherein avoidance-escape rats
could control shock to its tail by a nose-poke at the
front of its tube, “yoked”rats could nose-poke but
without effect on shocks and received equal tail
shock as their matched avoidance-escape part-
ners, and no-shock rats did not receive shocks in
this condition. When gastric pathology was
assessed after rats had been in these conditions
for 48 h, “yoked”rats developed clearly more
gastric erosions than were seen in the other two
groups.
Thus, the first foray into examining effects of
control versus noncontrol on various physiologi-
cal indices after appearance of the “executive
monkey”study yielded results consistently oppo-
site to its findings, including what was observed
when development of gastric pathology was
assessed. However, it was the next series of exper-
iments that led to the most illuminating results;
these findings will now be described.
Addressing the “Executive Monkey”
Findings: Assessing the Importance of
Signals Before Shock
The next experiments attempted to determine why
the findings in Weiss (1968) differed from what
was seen in the “executive monkey”study. As
described earlier, the “executive monkey”proce-
dure required the animals controlling shock to
respond in a “Sidman”avoidance procedure in
which all shocks occurred without any signal pre-
ceding them; responses simply postponed a train
of shocks before which no warning signal had
occurred. In contrast to this, in Weiss (1968),
shocks had always been preceded by a warning
signal. In those experiments, therefore, a warning
signal informed the rats that shock onset was
imminent, and also responses made during the
warning signal not only avoided shock (i.e.,
caused shock not to occur) but terminated the
warning signal to convey additional external
information for the animal exercising control.
Was a key factor in producing the difference the
presence of a warning signal before shock as used
by Weiss versus simply postponing shocks that
had no warning signal preceding them as was the
case in the “executive monkey”study?
To test this, the importance of warning signals
was examined (Weiss 1971a). In this large study,
rats were exposed to electric shocks that were
(1) unsignaled, or (2) preceded by a warning sig-
nal (“beeping”tone auditory signal beginning 20 s
before shock), or, additionally, (3) preceded by a
series of auditory warning signals (different tones)
even more elaborate than a usual single warning
signal (i.e., the “beeping”tone). As previously,
matched triplets of rats were used; the design as
well as the apparatus for this study is shown in
Fig. 2, and an animal in an enclosure with the
wheel mounted at the front is shown in Fig. 3.In
all three warning-signal conditions, a response by
the avoidance-escape rat –turning of the wheel at
the front of its apparatus –postponed the onset of
shock, which consisted of a train of brief, rapidly-
occurring shock pulses, for 200 s (i.e., this was an
avoidance response). If the shock train had begun,
the response immediately terminated shock and
shock did not occur again for 200 s (i.e., this was
an escape response). Thus, the avoidance-escape
rat exerted control over shock, avoiding/postpon-
ing shock or terminating shock if it had begun,
and the effect of a response by this animal on
shock –postponement of the next shock for
200 s –was exactly same in the three warning-
signal conditions.
What therefore differed in the three warning-
signal conditions was (a) how much external
information rats got as to when shock would
Executive Monkey Study 7

occur, and (b) how much external information rats
got when the avoidance-escape rat responded. In
the unsignaled shock condition, these rats, like the
“executive monkeys,”got no external feedback
informing them when shock would occur, and
also they did not experience any stimulus change
in their environment when they made an avoid-
ance response postponing shock. For these ani-
mals, only an escape response produced any
external change because in this case the shock
train terminated. For signaled shock animals, rats
knew when shock was imminent (i.e., the
“beeping”tone warning signal sounded), and if
the avoidance-escape animal responded during
this warning signal (avoidance response), an
external stimulus change occurred because the
warning signal immediately terminated. For the
third warning-signal condition, even more
information was provided; this condition essen-
tially provided an external clock leading up to
shock. In this condition, called Progressive signal,
a succession of warning signals was initiated 30 s
after a shock had occurred, and consisted of five
tones that became higher in frequency pitch and
louder every 30 s until the “beeping”tone warning
signal then began 20 s before shock. Note that the
yoked and nonshock animals of each triplet heard
exactly the same external tone signals as the
avoidance-escape animal, and yoked animals
experienced the same shocks and shock termina-
tions as well. Animals were exposed to these
conditions for 48 h, at which time they were
sacrificed and the presence of gastric pathology
(gastric erosions) was determined and quantified.
Findings from This Analysis: Exerting
Control Reduces Stress and Ulceration in
All Signal Conditions. Subsequent
Analysis Leads to a Promising General
Formulation
The results of this study did not indicate that the
simple presence or absence of warning signals
was the key to understanding the “executive mon-
key”findings. In all three conditions –Unsignaled
shock, Signaled shock, and the more elaborate
warning signal condition called Progressive
signal –the avoidance-escape rats able to exert
control over shock developed less pathology and
evidence of stress than did their matched yoked
animals that could not control shock (Weiss
1971a). Differences in gastric ulceration between
avoidance-escape and yoked rats were statistically
significant in all three signal conditions, favoring
avoidance-escape animals over yoked animals.
On its face, therefore, the “executive monkey”
study appeared to be an anomaly, and writings
between 1958 and the beginning of the 1970s
that emphasized the negatives of having the bur-
den of being in control were left lacking empirical
support.
However, detailed analysis of the large amount
of data generated by this study (20 matched trip-
lets were included in each signal condition) led to
a theoretical formulation that to this day appears
Executive Monkey Study, Fig. 3 The Plexiglas enclo-
sure used in Weiss (1971a) and also Weiss (1971b,c). Each
rat was placed in this Plexiglas chamber where it was able
to turn the wheel mounted at the front of the enclosure.
Electric shock could be delivered to the rat’s tail protruding
through the rear end of the enclosure via electrodes
attached to the tail. Each of these enclosures was placed
into a separate soundproof chamber so that each animal
was isolated and separated from other two animals of its
triplet. The white tubing leading from the top of the wheel-
turning enclosure is an air exhaust that provided constant
ventilation of the enclosure. The graduated cylinder seen at
the rear contains water; a spout from the cylinder protruded
into the enclosure through a small hole thereby providing
the rat access to drinking water throughout the procedure
8 Executive Monkey Study

to be valid and subsequently led to a potential
explanation for the “executive monkey”findings.
Examining the wheel-turning behavior of the rats,
particularly the animals in the unsignaled shock
condition, revealed that rats which made the most
responses tended to develop the most severe gas-
tric ulceration. This led to the first leg of the
hypothesis: As coping attempts (i.e., responses)
by the animal increases, stress and pathology
tends to increase. However, this relationship
between the amount of responding and develop-
ment of pathology was observed to be much
weaker in avoidance-escape animals of the Sig-
naled shock condition, and, in particular, was
weakest in avoidance-escape rats of the Progres-
sive signal condition. Why? The idea emerged
that the response rate of animals could be high
but if the amount of information they received
about the success of their responses was also
high, pathology would be prevented from devel-
oping. What was the relevant information? In a
stress situation, the relevant information that the
animal wants is that the danger/stress situation is
over, it has ended, it has gone away. In other
words, it seeks stimuli not associated with dan-
ger/stress/electric shock; i.e., it wants a stimulus
condition that represents safety. When shock is
signaled, and even more so in the Progressive
signal condition, wheel-turning responses by the
avoidance-escape animal often terminate warning
signals to immediately bring about silence which
is a safety period insofar as silence for these ani-
mals is present when shock is not occurring or
imminent. Consequently, warning signal termina-
tions produce a large amount of relevant informa-
tion that the animal seeks. This “good
information”from a response was called “relevant
feedback.”This led to the second leg of the
hypothesis: As relevant feedback from coping
attempts (i.e., responses) increases, stress and
pathology tends to decrease.
Thus the consequences of being in a stressful
situation are a function of two factors: adverse
effects increase as the number of coping attempts
(responses) increases, but then these adverse
effects are held in check by the outcome of the
coping attempts, decreasing (or failing to
increase) as the relevant feedback resulting from
coping attempts (responses) increases. The com-
bination of these two functions has been presented
graphically, which is shown in Fig. 4(from Weiss
1971a). This figure shows the prediction of
adverse effects generated by the convergence of
the two functions (adverse effects in this case
labeled “ulceration”insofar as this is what was
measured in the experiment described above). The
figure includes a hypothetical example indicating
how much pathology will occur when a certain
number of responses are made in a given situation
and these responses produce a certain amount of
relevant feedback.
Executive Monkey Study, Fig. 4 The three-
dimensional plane describing the proposed relationship
between the two independent variables –“Responses”
(coping attempts) and “Relevant Feedback”from such
coping attempts –and how this is related to the severity
of stress and/or pathology (labeled here “Ulceration”). As
Responses increase and Relevant Feedback decreases,
stress/pathology increases. For ease of reading, the axes
for Responses and Relevant Feedback are labeled in the
foreground; customarily, these labels are placed on these
axes in the background. Also presented in this figure is a
hypothetical example showing the amount of pathology
resulting from (A) a low number of responses being made
(B) when these responses generate a low amount of rele-
vant feedback
Executive Monkey Study 9

Figure 5shows how the results of exposure to
the three different warning signal conditions in
Weiss (1971a) conformed to the formulation
shown in Fig. 4. Wheel-turn responses were
counted, so this was known for all groups. Regard-
ing relevant feedback, this could be estimated as
well. For the avoidance-escape groups, those in the
Signal condition received a moderate amount of
relevant feedback for responding, resulting from
signal terminations and escape responses, whereas
those in the Progressive signal condition received
even more relevant feedback because of the pres-
ence of more auditory signals that responses termi-
nated. The lowest amount of relevant feedback was
experienced in the unsignaled (No Signal) condi-
tion. Here avoidance responses postponed shock
but changed nothing in the external environment;
only escape responses (shock terminations) pro-
duced relevant feedback in this condition. These
three conditions are represented appropriately in
Fig. 5. The most interesting and informative
groups, however, are the yoked groups. This for-
mulation tells us why it is so detrimental to have no
control over a stressor. For these animals, relevant
feedback from any response is always zero –all
responseshavenoeffectontheenvironmentorthe
stressor. That is, in fact, what it means to be without
control –relevant feedback from any coping
attempt is zero. In this condition, adverse effects
of stress are a direct function of the number of
coping attempts made –the higher the response
rate in this condition, the more the pathology. This
relationship can be seen in the three yoked groups
shown in Fig. 5.
Tests of the Formulation and Explaining
the “Executive Monkey”Findings
The validity of the formulation presented in Fig. 4
was then tested in two experiments. First, it is
noted above that having no control is a condition
where relevant feedback from any coping attempt
is zero. But what would happen if the value of
relevant feedback might fall below zero, becom-
ing negative? This could occur if a coping attempt
(avoidance-escape response) produced some
aspect of the stressor rather than leading to a
stimulus condition indicating absence of the
stressor. In the first test experiment, avoidance-
escape animals were required to perform a wheel-
turning response to avoid and/or terminate a train
of shocks but then negative relevant feedback for
such responding was introduced by giving one
brief pulse of shock immediately following any
wheel-turning response. The results of this study,
reported in Weiss (1971c), made clear that this
condition was very detrimental; so much pathol-
ogy developed when animals were forced to
respond when relevant feedback from coping
attempts becomes negative that animals could
hardly survive this condition for 24 h. Next, as
noted above, avoidance-escape responding in an
unsignaled shock condition produces a low level
of relevant feedback. Therefore, a manipulation
was preformed to greatly increase relevant
Executive Monkey Study, Fig. 5 Results found in
Weiss (1971a) for the Avoidance (Avoidance-escape) and
Yoked groups of the three different signal conditions. For
each group, the amount of gastric ulceration observed
(height of bar) is shown at the point where the amount of
responding observed in that group and the amount of
relevant feedback from responding in that condition inter-
sect. Note that for Yoked groups, amount of relevant feed-
back in all conditions is zero, which is necessarily the case
when animals have no control over the stressor
10 Executive Monkey Study

feedback in this condition. To do this, a brief
auditory signal was added after every avoidance-
escape response; because this signal is a stimulus
always removed in time from the occurrence of
shock, it represents a safety signal and greatly
increases relevant feedback from responding in
the unsignaled shock condition. The results of
this study, reported in Weiss (1971b), showed
that simply adding this auditory signal for relevant
feedback essentially eliminated gastric pathology,
reducing it almost to the level seen in no-shock
subjects that received no shocks at all. In sum-
mary, findings of these two experiments strongly
supported the formulation shown in Fig. 4and the
underlying reasoning that led to it. Incidentally,
these studies established that the rat’s receiving
uncomfortable or even painful electric shocks
throughout the procedure had almost no influence
on the development of stress/gastric pathology;
the severity of stress and pathology was shown
to depend almost entirely on behavioral and psy-
chological factors attending the shock situation
and did not result from the shocks themselves.
The formulation shown in Fig. 4can be used to
account for the results seen in the “executive
monkey”study. As described earlier, the
“Sidman”avoidance schedule used in that exper-
iment administered unsignaled shocks to the mon-
keys, and, for the monkey controlling the shock
by pressing its lever, the relevant feedback from
responses was therefore low. In fact, because
shock in that experiment consisted of a train of
very brief shock pulses each pulse given 5.0 s
apart, the animal controlling shock could not
even perform a clearly-defined escape response
that demonstrably terminated shock –in contrast,
every lever press by this monkey just postponed
shock without producing any stimulus change in
its environment, a condition of very low relevant
feedback. Thus, a high rate of responding in such
a condition is prone to high stress and pathology.
Additionally, an important detail regarding
how the “executive monkey”study was
conducted adds to its pathogenic potential. As
also discussed earlier, the “executive monkey”
study was not designed to examine the effect of
having control versus no control, but, rather, was
conducted to establish that preforming an
avoidance response could lead to gastric pathol-
ogy. Consequently, no attempt was made to
equate the avoidance and yoked monkeys of
each pair; instead, a yoked subject was simply
paired with each avoidance monkey to insure
that an animal not performing the avoidance
response would receive an equal number of
shocks in order to rule out the electric shock as
having produced the gastric pathology. Because
of this, the experimenters used an unfortunate
procedure: they set up both animals of each pair
with an active lever at the outset of training, and
let the animal that became the avoidance subject
self-select itself by taking over avoidance
responding. Within the first hours of the experi-
ment, one monkey of the pair responded at a much
higher rate than the second monkey, and so the
high-rate responder was made the avoidance sub-
ject and the other monkey’s lever was deactivated
so that it became the yoked subject. This anomaly
is crucial –as a result of this procedure, the “exec-
utive monkey”wasdesignatedassuchbecauseit
responded at a high rate, and higher than its yoked
subject. Thus, the monkeys that were made the
“executive”were, because of their response ten-
dencies, more prone to develop pathology than the
yoked monkeys from the outset. And then these
“executive monkeys”were placed into an ongoing
situation where their responding produced very
low relevant feedback. Consulting Fig. 4,thisset
of circumstances –high response rate, low relevant
feedback from responding –is precisely the kind of
situation that will result in pathology, and this is
what indeed was observed in the “executive mon-
key”study.
More than 45 years have now elapsed since
publication of the Scientific American article in
1972 titled “Psychological Factors in Stress and
Disease”(Weiss 1972) which was relevant to the
“executive monkey”experiment that had been
published in the same Journal in 1958. Since
then, a number of significant developments have
occurred. First, the principles proposed in that
article –i.e., that stress and pathology increases as
(a) responses increase and (b) relevant feedback
from responses decreases –has been embraced,
applied, and reiterated in a number of fields, includ-
ing, notably, industrial psychology (e.g., Karasek
Executive Monkey Study 11

1979;Siegrist1996). Second, a large amount of
research has established that having control over a
stressful situation is more beneficial –i.e., results in
less stress –than having no control, which is oppo-
site to what the “executive monkey”study
suggested (e.g., Gray 1998;Payne2013; Sherman
et al. 2012). This is now accepted as an accurate
general conclusion. Third, it is often said that the
“executive monkey”results now have been falsi-
fied, debunked, as a result of the selection factor
used to choose the “executive”monkey. For exam-
ple, this is what was stated recently by Sherman
and colleagues in Proceedings of the National
Academy (Sherman et al. 2012). I respectfully dis-
agree. The results of the “executive monkey”study
have not been debunked; they are not erroneous.
They are highly lawful and exactly what one would
predict given the conditions that generated them.
These findings are clearly explained in accord with
the formulation shown in Fig. 4.Whatiscorrectto
say is that the executive monkey experiment does
not describe what one would expect to be the
normal consequence of exerting control versus no
control, but, rather, represents a highly unusual
situation where exerting control was detrimental –
because of the resulting high response rate, aug-
mented by the avoidance monkeys being selected
because of their high response proclivity, and the
low relevant feedback from responding. Conse-
quently, we now understand the findings in the
“executive monkey”experiment and why negative
consequences occurred for animals “in control,”
and therefore these findings can be appropriately
integrated into our knowledge.
Concluding Comments
First, with the discovery, in the 1980s, of the
pathogen Helicobachter pylori as an important
factor in peptic ulcer disease, and the successful
treatment of the disorder by antibiotics (Marshall
and Warren 1984; Marshall et al. 1985), initially it
was thought that any association of this disorder
with stress was now disproven, and decades of
previous study and theorizing in this regard were
erroneous. However, (1) it was subsequently dis-
covered that ulcer could develop and also recur in
the absence of H. pylori (e.g., Laine et al. 1998;
Tsuji et al. 1999), (2) it has been proposed that
H. pylori is not the primary cause of ulcer but,
rather, is responsible for a secondary infection that
accounts for chronicity of ulcer disease (Tovey
et al. 2001; Vikram et al. 2013), and (3) while
prevalence of H. pylori infection is high across the
world (for example, present in 35% of the popu-
lation in the U.S. [Hooi et al. 2017]), peptic ulcer
disease is found in a much lower percentage of
people (for example, present in 1.1–1.5% of the
population in U.S. [Sung et al. 2009]), so other
factors clearly must be present for peptic ulcer
disease to occur. Thus, H. pylori infection is nei-
ther necessary nor sufficient for development of
peptic ulcer. Also, stress has been shown to
decrease bodily defenses against pathogens by
decreasing protective immune responses includ-
ing phagocytosis that controls gram-negative bac-
teria such as H. pylori (e.g., Glazer and Kiecolt-
Glazer 2005), so stress may influence infection by
H. pylori. Research throughout this period has
continued to show that development of peptic
ulcer disease requires hypersecretion of stomach
acid. Fifty years earlier, Wolf and Wolff had
shown that emotional upset led to gastric engorge-
ment and increased stomach acid secretion in
humans (Wolf and Wolff 1947). In summary, dis-
covery of the significance of H. pylori for ulcer
disease doesnot rule out an important role for stress
in development of gastric and duodenal ulcer.
Second, while the writings of psychoanalysts
on the topic of ulcer have been pushed aside for
many years, this writer suggests that perhaps the
work of perceptive psychoanalytic colleagues
should not be summarily dismissed. Franz Alex-
ander, in theorizing that many ulcer patients were
hard striving individuals whose unconscious
desire was, instead, to be passive and nourished,
describes a high response rate, low relevant feed-
back condition leading to gastric pathology.
Experimental research on animals described in
this chapter provides a framework indicating that
this particular condition is one of high stress and
conductive to development of gastric pathology.
That framework may integrate the data from these
different sources; perhaps there is a surprising
agreement here.
12 Executive Monkey Study

References
Alexander, F. (1950). Psychosomatic medicine. New York:
W. W. Norton.
Brady, J. V. (1958). Ulcers in executive monkeys. Scientific
American, 199(4), 95–98.
Brady, J. V., Porter, R. W., Conrad, D. G., & Mason, J. W.
(1958). Avoidance behavior and the development of
gastroduodenal ulcers. Journal of the Experimental
Analysis of Behavior, 1,69–72. https://doi.org/10.
1901/jeab.1958.1-69.
Coxe, J. R. (1846). The writings of Hippocrates and Galen.
Philadelphia: Lindsay and Blakiston.
Dunbar, H. F. (1935). Emotions and bodily changes:
A survey of literature on psychosomatic interrelation-
ships. New York: Columbia University Press.
Freud, S. (1955). The standard edition of the complete
works of Sigmund Freud. Vol. 2, Studies of hysteria
(1893–1895) (trans: Strachey, J.). London: The
Hogarth Press.
Glazer, R., & Kiecolt-Glazer, J. K. (2005). Stress-induced
immune dysfunction: Implications for health. Nature
Reviews: Immunology, 5, 243–251.
Gray, R. (Producer). (1998, August 14, 2017). Workplace
stress. Retrieved from http://rodericgray.com/
workplacestress.pdf
Hooi, J. K. Y., Lai, W. Y., Ng, W. K., Suen, M. M. Y.,
Underwood, F. E., Tanyingoh, D., Malfertheiner, P.,
Graham, D. Y., Wong, D. W. S., Wy, J. C. Y., Chan,
F. K. L., Sung, J. J. Y., Kaplan, G. G., & Ng, S. C.
(2017). Global prevalence of Helicobacter pylori infec-
tion: Systematic review and meta-analysis. Gastroen-
terology, 153(2), 420–429.
Kalb, C. (2016). Andy Warhol was a hoarder: Inside the
minds of history’s great personalities. Washington, DC:
National Geographic Society. (excepted in National
Geographic, May 2017, 231(5), 154).
Karasek, R. A. (1979). Job demands, job decision latitude,
and mental strain: Implications for job redesign.
Administrative Science Quarterly, 24(2), 285–308.
Laine, L., Hopkins, R. J., & Girardi, L. G. (1998). Has the
impact of Helicobacter pylori therapy on ulcer recur-
rence in the United States been overstated? A meta-
analysis of rigorously designed trials. The American
Journal of Gastroenterology, 93(9), 1409–1415.
Lipowski, Z. J. (1984). What does the word “psychoso-
matic”really mean? A historical and semantic inquiry.
Psychosomatic Medicine, 46(2), 153–171.
Marshall, B. J., & Warren, R. M. (1984). Unidentified
curved bacilli in the stomach of patients with gastritis
and peptic ulceration. Lancet, 16, 1311–1315. https://
doi.org/10.1016/S0140-6736(84)91816-6.
Marshall, B. J., Armstrong, J. A., McGechie, D. B., &
Glancy, R. J. (1985). Attempt to fulfill Koch’s postu-
lates for pylori campylobacter. Medical Journal of
Australia, 142(8), 436–439.
Mowrer, O. H., & Viek, P. (1948). An experimental ana-
logue of fear from a sense of helplessness. Journal of
Abnormal Psychology, 43(2), 193–200.
Payne, B. K. (2013). The myth of executive stress. Scien-
tific American, Mind Matters Column (24 Sept).
Porter, R. W., Brady, J. V., Conrad, D., Mason, J. W.,
Galambos, R., & Rioch, D. M. (1958). Some experi-
mental observations on gastrointestinal lesions in
behaviorally conditioned monkeys. Psychosomatic
Medicine, 20(5), 379–394.
Sherman, G. D., Lee, J. J., Cuddy, A. J., Renshon, J., Oveis,
C., Gross, J. J., & Lerner, J. S. (2012). Leadership is
associated with lower levels of stress. Proceedings of
the National Academy of Sciences of the United States
of America, 109(44), 17903–17907. https://doi.org/10.
1073/pnas.1207042109.
Siegrist, J. (1996). Adverse health effects of high-effort/
low-reward conditions. Journal of Occupational
Health Psychology, 1(1), 27–41.
Sung, J. J. Y., Kuipers, E. J., & El-Serag, H. B. (2009).
Systematic review: The global incidence and preva-
lence of peptic ulcer disease. Alimentary Pharmacol-
ogy and Therapeutics, 29, 938–946.
Tovey, F. I., Hobsley, M., & Holton, J. (2001). Helicobacter
pylori virulence factors in duodenal ulcer: A primary
cause or a secondary infection chronicity. World Journal
of Gastroenterology, 96(5), 1409–1416.
Tsuji, H., Kohli, Y., Fukumitsu, S., Morita, K., Kaneko, H.,
Ohkawara, T., Minami, M., Ueda, K., Sawa, Y.,
Matsuzaki, H., Morinaga, O., & Ohkawara, Y. (1999).
Helicobacter pylori-negative gastric and duodenal
ulcer. Journal of Gastroenterology, 34(4), 455–460.
Tuke, D. H. (1872). Illustrations of the influence on the
mind upon the body in health and diseases. London:
J. & A. Churchill.
Vikram, K., Ananthakrishan, N., & Tovey, F. I. (2013). Is
Helicobacter pylori infection the primary cause of duo-
denal ulcer or a secondary factor? A review of the
evidence. Gastroenterology Research and Practice,
2013.https://doi.org/10.1155/2013/425840.
Weiss, J. M. (1967). A tail electrode for unrestrained rats.
Journal of the Experimental Analysis of Behavior, 10,
85–86.
Weiss, J. M. (1968). Effects of coping responses on stress.
Journal of Comparative and Physiological Psychology,
65(2), 251–260.
Weiss, J. M. (1971a). Effects of coping behavior in differ-
ent warning signal conditions on stress pathology in
rats. Journal of Comparative and Physiological Psy-
chology, 77(1), 1–13.
Weiss, J. M. (1971b). Effects of coping behavior with and
without a feedback signal on stress pathology in rats.
Journal of Comparative and Physiological Psychology,
77(1), 22–30.
Weiss, J. M. (1971c). Effects of punishing the coping
response (conflict) on stress pathology in rats. Journal
of Comparative and Physiological Psychology, 77(1),
14–21.
Weiss, J. M. (1972). Psychological factors in stress and
disease. Scientific American, 226(6), 104–113.
Wolf, S., & Wolff, H. G. (1947). Human gastric function.
London: Oxford University Press.
Executive Monkey Study 13