Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
pa ge ~ The literature concerning stress is extensive and complex, extending through fields as varied as clinical applied psychology, anthropology, sociology, psychosomatic medicine, industrial relations, and epidemiology. Not included in this list are' of course, the extensive studies dealing with the biochemical and physiological of the responses to stress. These responses have been involved in mechanisms as basic as immunological function, metabolic function, and fundamental psychological processes, such as memory and learning. Since one of the primary problems in stress research is conceptual, and this problem takes many forms, there is a great deal of confusion in the field. Because stress researchers lack a common vocabulary, each writer must define his/her own terms, and the reader must scrutinize each article carefully in order to understand the writer's vocabulary The lack of a uniform and consistent vocabulary is a substantial impediment to progress and adds materially to the confusion in the field. Although the term rstressn is used throughout the literature, it is apparent that this term has multiple meanings, depending upon the particular field in which the concept is being investigated. Within the context of this report, we shall attempt to use one set of operational definitions to define stress, and at least to be consistent with our own definitions of the primary psychological variables that induce many of the profound long-term effects commonly attributed to stress. Stress can be approached from a purely behavioral perspective, and it effects studies on primarily behavioral outcomes. However, stress has also been viewed predominantly as a physiological and psychosomatic process, and the outcomes arestudies on either pathophysiological processes or basic biological processes. This report, however, will focus on an integration of these two perspectives and present a psychobiological view of stress.
Page 3 It is important to note that, historically, the concept of stress has been predominantly implicated with changes in the endocrine systems. Initially, the changes were specifically related either to increased secretion of catecholamines or to activation of the pituitary-adrenal system. The problems are best illustrated by examining the concept of stress beginning with Selye's (1936) early work in which he defined a general adaptation syndrome (GAS) in rodents. This nonspecific response occurred after diverse noxious agents, such as exposure to cold, surgical injury, spinal shock, and muscular exercise. The essential argument was that the response did not depend upon the type of agent that produced it; rather, like inflammation, it was deemed as nonspecific. GAS was divided into three stages: an alarm reaction, a stage of resistance, and a stage of exhaustion. The initial stage included activation of the pituitary-adrenal system and eventually resulted in adrenal hypertrophy, thymicolymphatic involution, and gastric ulceration if the noxious stimuli persisted. If the response to the aversive situation was sustained, physiological resistance ultimately developed, and it was hypothesized that stressed subjects would enter a third stage--exhaustion--which occurred 1-3 months after the initial exposure. Problems with this view have occurred at several levels. First, there was an early emphasis on the physical and chemical aspects of the stressful stimuli, and we now know that psychosocial stimuli are also potent elicitors of the stress response. Second, Selye has received much criticism for the "nonspecificity" view because, with modern hormone assays, it is now possible to detect differential endocrine responses to certain stimuli. Further, the importance of the final stage of exhaustion has been questioned. Diseases due to exhaustion of this syndrome are rare, and, with the exception of a few animal models (e.g.,
Page 4 intruders in wild rat colonies), have not been demonstrated as a response to psychosocial stimuli (Allen, 1972). Moreover, a number of physicians have studied moribund patients and have found that adrenal exhaustion did not occur even at death. Rather, there is usually increased adrenal output immediately before and after death (Sandberg et al., 1956~. The dramatic picture described by Selye is emphatically different from the present day concept of stress in the lay literature, which includes the daily troubles and anxieties of commuters and executives. The broader use of the term has resulted in an urgency to reduce or eliminate stress in both personal and professional arenas, even though Selye (1974) himself has minimized the significance of this type of stress and stated that its absence occurs only after death. This paradoxical situation reveals that we do not have a clear and generally accepted definition. As a consequence, there is a serious communication problem and increasing talk about a crisis in stress research (Wolf et al., 1979). At the very least, there is a growing impatience with the present state of vagueness in an area so vitally important for issues of health and quality of life. We believe that much of the controversy over stress theory can be eliminated through clarification of the Different limb,. that is, by focusing on the nature of the stimuli that provoke physiological responses rather than on the physiological responses themselves. This type of investigation requires an unusual integration of physiology and psychology--disciplines which have traditionally been separated--and puts the major emphasis on psychological variables.
Page 5 One of the purposes of this report is to examine the importance of psychological variables that have been determined to have profound endocrinological consequences both in animals and humans. In fact, the major conceptual framework pervading this report is that one of the primary aspects of stressful stimuli eliciting an endocrine response is psychological in nature. This perspective is derived from Mason's (1968, 1975a,b) review of psychoendocrine research, particularly involving the pituitary-adrenal cortical system. As mentioned above, much of the early stress research had emphasized the nonspecificity of the organism's response to a wide variety of physical stressors (Selye, 1950). However, even in the 1950s, it was becoming increasingly apparent that psychological factors were importantly involved. For example, in one study, Renold et al. (1951) examined the physiological response of participants in the Harvard Boat Race. Utilizing a traditional measure of that period, the decline in eosinophils following stress, they found that eosinophils in the crew members were markedly lower 4 hours after the race. This decline could have been attributed solely to the exercise and physical strain, but the investigators also discovered that the coxswains and coaches had similar eosinophil drops, even though their stress was purely psychological. In Mason's major review of the stress literature in 1968, he pointed out that much of the prior work, including the experiments on physical stimuli, shared one important characteristic, namely, that a typical aspect of the stressful experience was exposure to novel, strange, or unfamiliar environments. Therefore, the common threat that may have explained the animals' response was the psychological dimension of the stimuli, rather than the particular physical trauma to which they had been exposed. In subsequent research, Mason (1975a,b) was able to show that when animals are exposed to the
Page 6 stimuli in such a way that they do not experience distress or novelty, then typical stressors such as heat or fasting do not necessarily result in activation of the pituitary-adrenal system. The concept that psychological variables can activate, and inhibit, the endocrine system has subsequently received much support in experimental studies on both animals and humans. PITUITARY-ADRENAL SYSTEM Although the response to stress can best be defined as a syndrome, which includes many changes in neurochemical and metabolic processes, for the purposes of this report we will focus on the response of the pituitary-a/renal system. It is important to remember, however, that we are utilizing this as a model system. There is abundant evidence indicating that the hormone function of ocher endocrine systems--including insulin, growth hormone, and prolactin--can also be influenced by psychological variables. In addition, it has been demonstrated recently that the endorphins are also extremely responsive to stress. In fact, it appears that almost all of the stimuli capable of eliciting an ACTH response from the pituitary are also capable of releasing beta endorphins (Guillemin et al., 1977). There are two reasons for focusing on the pituitary-adrenal system in illness. First, there is an extensive data base showing the effects of psychological variables on the pituitary-adrenal system. Second, and perhaps more important, is the profound influence that adrenal hormones have on many basic functions related to health. There have been many attempts in recent years to resolve the issue of the primary stimuli that elicit the endocrine response, in particular pituitary-adrenal responses, which occur under conditions of stress. As Mason
Page 7 (1975a) pointed cut, when the psychologically threatening or arousing aspects of the situation were altered, classical stresses such as fasting and heat no longer activated the pituitary-adrenal system. In the case of heat, there was, in fact, a reduction in the corticoids when the mode of presentation was gradual. The importance of the rate of presentation of a particular stimulus was also demonstrated in another experiment that used a potent physiological insult to induce adrenocortical activity. Hemorrhage in the magnitude of 10 ml/kg at the rate of 6.6 ml/kg/min actively stimulates the adrenal cortex of the dog. In contrast, if the same ultimate volume of blood loss is achieved at a much slower rate of hemorrhage (i.e., 0.3 ml/kg/min), the pituitary-adrenal system does not activate (Gann, 1969). That rapid hemorrhaging induces adrenocortical activity, while slow rates of hemorrhaging do not, once again indicates that the rate of stimulus change is one important parameter for the induction of pituitary-adrenal activity. The fact that dexamethasone blocks the pituitary-adrenal response at a high rate of hemorrhaging clearly indicates that neuroendocrine systems are involved and that the effect was not mediated peripherally. Regardless of the specific explanation that accounts for these results, their general significance cannot be underestimated. Hare recent studies on psychoendocrine responses have indicated further that it may be possible to use adrenal activity as a measure of specific emotional responses, rather than simply as a reflection of undifferentiated arousal (Hennessy and Levine, 1979; Mason, 1975a,b). In addition, studies on psychological stress bring out one point quite clearly; the great individual differences typically observed in response to a given stressor can best be explained in terms of cognitive mechanisms. For example, a subject's perception
Page 8 of a stressor as z th.ea_, or the coping responses that are available to the subject, may well determine the physiological response. It may be insufficient, therefore, to merely describe the stimulus operations involved in producing a stressor. A psychobiological approach to understanding endocrine function cannot escape making reference to cognitive processes. In his new description of arousal theory, Berlyne (1960, 1967) provides a framework for the description of the processes by which stimulators of arousal (and thus, activators of the pituitary-adrenal response) operate. Novelty, uncertainty, and conflict are considered primary determinants of arousal. These have been labeled by Berlyne as collative factors, because in order to evaluate them, it is necessary to compare similarities and differences between stimulus elements (novelty), or between stimulus-evoked expectations (uncertainty). The basic cognitive process involved in stimulation of the pituitary-adrenal system, then, is one of comparison. To a large extent, the cognitive processes of comparison can be best understood by adding the concept of uncertainty, although there are some differences between uncertainty and novelty. Uncertainty seems to be a major factor underlying many psychological responses. The processes involving neuroendocrine activation under conditions of uncertainty are best explained by a model elaborated by Sokolov (1960) to account for the general process of habituation. The pattern of habituation is familiar to most people. A subject is presented with an unexpected stimulus and shows an alerting reaction. Physiological components of this orienting reaction are well known--general activation of the brain, decreased blood flow into the extremities, changes in electrical resistance of the skin; and increases in both adrenomedullary and cortical hormones. If the stimulus is frequently repeated,
Page 9 all of these reactions gradually dim~nisn and eventually disappear, and the subject is said to be habituated. It does appear, however, that physiological responses may habituate more slowly than the observed behavioral reactions. Sokolov's model, in essence, is based on a matching system in which new stimuli or situations are compared with a representation in the central nervous system of prior events. This matching process results in the development of expectancies whereby the organism is either habituated or gives an alerting arousal reaction (Pribram and Melges, 1969). Thus, the habituated organism has an internal representation of prior events with which to deal with the environment--expectancies--and if the environment does not contain any new contingencies, the habituated organism no longer responds with the physiological responses related to the alerting reaction. Activation of the pituitary-adrenal system by any change in expectancy can also be accounted for by invoking the powerful explanatory capacity of the Sokolov model. NOVELTY AND UNCERTAINTY Exposure of an animal to novelty is one of the most potent experimental conditions leading to an increase in pituitary-adrenal activity. Novelty can be classified as a collative variable, since the recognition of any stimulus situation as being novel requires a comparison between present stimulus events and those experienced in the past. Increases in pituitary-adrenal activity in response to novelty have been demonstrated in humans as well as animals. For example, increased adrenocortical activity, as evidenced by elevated levels of circulating cortisol, are observed in individuals during their first exposure to procedures involved in drawing blood at a blood bank. However, if they have had
Page 10 prior blood bank experience, there are no such increases (Mendoza and Barchas, 1982). Further, in an experiment to be discussed in detail later, young adults experiencing their first jump off a tower during parachute training also show a dramatic elevation of adrenocortical activity, which is not observed on subsequent jumps from the tower (Levine, 1978). Studies on animals also indicate another important characteristic of the cognitive process which results in pituitary-adrenal activation, that is, the ability of the animal to discriminate similar vs. unfamiliar stimulus elements. In a series of experiments on rats and mice it was demonstrated that if novelty was varied along a continuum with increasing changes in the stimulus elements, there was a graded adrenocortical response according to the degree.to which the environment represented a discrepancy from the normal cage living environment of the organism (Hennessy and Levine, 1977 ; Hennessy et al., 1979~ . Thus, minor changes, such as placing the animal in a different cage, but one identical with its home cage, resulted in an elevation of plasma corticosterone, but one that was significantly less than when the animal was placed in a totally novel cage containing none of the elements of its familiar living conditions. This capacity to make fine discriminations resulting in graded elevations of pituitary-adrenal activity are clearly demonstrative of the remarkable capacity of the central nervous system to regulate the output of pituitary-adrenal response. Novelty, according to the theory presented by Sokolov, should indeed be one of the most potent variables that elicits increases in pituitary-adrenal activity. Insofar as an organism has no expectations about an unfamiliar environment, that environment should represent a degree of uncertainty that should lead to increases in neuroendocrine activity.
Page 11 Although novelty can be subsumed under the general concept of uncertainty, not all conditions which create uncertainty are novel. Uncertainty can also be evoked by insufficient information concerning the nature of upcoming events. Uncertainty can be seen to vary along the continuum from highly certain, predictable events to highly uncertain, unpredictable events. The presentation of a novel stimulus is likely to lead to an increase in uncertainty because, by definition, there is little information the organism can use to predict forthcoming events. However, uncertainty can also be defined in terms of contingencies between environmental events. Experimentally, the dimension of uncertainty can be controlled by limiting the amount of information available to the organism to predict the occurrence of a specific event. Thus, one would hypothesize that if an organism is given information about the occurrence of either an appetitive or an aversive stimulus, such predictability should lead to a reduction in the pituitary-adrenal response. Further, situations in which there is an absence of predictability should lead to a dramatic increase. There are many experiments that illustrate the value of predictability in modifying the pituitary-adrenal response to a variety of stimuli (Weinberg and Levine, 1980). One illustration of the effects of reducing uncertainty by" providing predictability can be seen in a study by Dess et al. (1983). Dogs were subjected to a series of electric shock which were either controllable or predictable. The predictable condition involved presenting the animal with a tone prior to the onset of shock. In the unpredictable condition, no such tone was presented. The adrenocortical response observed on subsequent testing of these animals clearly indicated the importance of reducing uncertainty by predictability. Animals that did not have the signal preceding the shock showed
. Page ;2 an adrenocortical response which was two to three times that observed in animals with previous predictable shock experiences. It should be noted that the procedures used in this experiment are typical of those utilized in experiments examining learned helplessness (Selig~an, 1975). Learned helplessness refers to the protracted effects resulting from prolonged exposure to unpredictable and uncontrollable stimuli of an aversive nature. It has been observed that organisms exposed to this type of an experimental regimen show long-term deficits in terms of their inability to perform appropriately under subsequent testing conditions. Further, these animals show ~ much greater increase in adrenocortical response when exposed to novelty (Levine et al., 1973) than do control animals. Thus, an organism exposed to an uncontrollable and unpredictable set of aversive stimuli not only shows a dramatic increase in adrenocortical activity while exposed to these conditions, but there is also a long-term effect in other unrelated test conditions. The concept of control is particularly important in understanding these long-term effects, and it will be dealt with later when the issue of coping is discussed. There is yet another series of experiments related to the issue of uncertainty. These do not utilize aversive stimuli typical of stress research, but are more directly related to a psychological response commonly described as frustration. Frustration can be evoked when the organism fails to achieve a desired goal following a history of successful fulfillment of these goals. Experimentally, the operations utilized to produce frustration involve either preventing an animal from making the appropriate response to achieve a desired object, or not reinforcing the animal for a response that has had a prior history of reinforcement. In a broader sense, frustration involves the failure of an
Page 13 animal to fulfill expectancies developed in previous experiences and, thus, car. be subsumed under the larger heading of uncertainty. For example, rats trained to press a lever for water, in which each lever-press delivered a small amount of water, showed a dramatic elevation of plasma corticosterone when the water was no longer available following the lever-press response (Coe et al., 1983b). Elevations of plasma corticosterone have been shown to be a robust and reliable phenomenon occurring under many experimental conditions in which reinforcement contingencies are altered. Thus, not only is an elevation of plasma corticosterone observed when reinforcement is eliminated, but if the animal receives less reinforcement than it has previously become accustomed to, then elevations of plasma corticosterone also occur. A similar phenomenon can be observed when using aversive stimuli. If an animal has learned to make an appropriate avoidance response that eliminates the occurrence of an electric shock, and this animal is then prevented from making the response, an increase in pituitary-adrenal activity occurs even when no electric shock is delivered (Coover et al., 1973). These experiments have led to the belief that one of the primary conditions to activate the neuroendocrine mechanisms leading to a subsequent adrenal response is a change in expectancies concerning well-established behaviors. In the case of the appetitive learning situation when reinforcement is eliminated, as well as in the avoidance experiment, activation of the pituitary-adrenal system occurred following disruption of ongoing behavior which had once led to a predictable set of outcomes. This can best be understood if one assumes that, under conditions whereby frustration is evoked, the absence of reinforcement represents a condition in which uncertainties are introduced.
Page 1.4 COPING Since the time of Bernard and Cannon, the importance of maintaining physiological homeostasis has been well known. The consequences of inappropriate adrenal secretion are evident from the effects of both hyper- and hypo-adrenal output. Organisms deprived of adrenal corticoids are clearly in jeopardy and unable to effectively deal with even minor stressors that are of little consequence to an intact organism, such as water or salt restriction. Conversely, excessive secretion of the pituitary-adrenal system is also maladaptive. Prolonged elevations of adrenocortical hormones can have high biological cost, leading to increased susceptibility to psychosomatic illnesses as well as other pathophysiological processes, through their effects on the immune system. It would follow, therefore, that there must have evolved a set of mechanisms available to the organism whereby it could regulate and modulate excessive output of adrenocorticoids. We believe that these mechanisms are predominantly psychological and should be classified under the general rubric of .coplng . Coping differs from habituation in one profound sense. In the case of habituation, it is presumed that the organism has changed its evaluation of the stimulus through repeated experience and has developed a set of expectancies concerning the benign characteristics of the stimulus or environment. Coping, on the other hand, is a more active process and can be defined in terms of the absence of a physiological response even under conditions in which the aversive stimulus continues to be present. In the case of coping, cognitive and behavioral processes are actively involved in determining whether an individual does or does
Page 15 not respond to a specific stressful situation. It is not just the aversive nature of the stimuli, per se, that determines the physiological response, but rather the individual's evaluation of these stimuli. We can regard this as a filter or Bating function. An organism can alter its evaluation of potentially threatening or aversive stimuli if it can avoid, alter, or master the stress-inducing aspects of the situation. In a previous context, we have discussed the importance of predictability. There are two other psychological processes also involved in the process of altering an organism's evaluation of stressful events: control over aspects of the situation, and feedback about the efficacy of its actions. Perhaps the most important single determinant of the ability of the organism to reduce its hormonal responses to aversive stimuli is control. Control can best be defined as the capacity to make active responses during the presence of an aversive stimulus. These responses are frequently effective in allowing the animal to avoid or escape from the stimulus, but they may also function by providing the animal with the opportunity to change from one set of stimulus conditions to another, rather than to escape the aversive stimulus entirely. Control, in and of itself, can reduce an organism's physiological response to such noxious stimuli as electric shock. It has been observed that rats able to press a lever to terminate shock show less severe physiological disturbances (e.g., weight loss and gastric lesions) than yoked controls which cannot respond, even though both animals receive the identical amount of shock. Similarly, animals able to escape from shock show a reduction in plasma corticosterone following repeated exposures of shock (Davis et al., 1977). The effects of control have also been demonstrated in an experiment by Hanson et al. (1976).
Page 16 These investigators studied rhesus monkeys that were exposed to another noxious stimulus--a loud noise. One group of monkeys were permitted to control the duration of the noxious stimulus by making a lever-press response to terminate the noise. A comparable group of monkeys were given the identical amount of noise, but were not permitted to regulate the duration. The animals that were allowed control procedures showed identical levels of plasma cortisol as those which were not exposed to the noise at all, compared to the yoked controls that showed extremely high levels of plasma cortisol. The effect of control on plasma cortisol levels was also demonstrated very clearly in a recent experiment on dogs (Dess et al., 1983). These animals were subjected to a standard procedure used to produce learned helplessness. They were placed in a hammock and given uncontrollable and unpredictable shock. Other dogs were allowed to control the shock and terminate it by making a panel-press response with their heads. We have previously discussed the role of predictability within the context of this experiment. The results further indicated that controllability also affected the magnitude of the cortisol response to the shock. Having neither control nor predictability elicited the maximum cortisol response; having both minimized the impact of the shock. However, the capacity to control the stimuli appeared to have a greater effect in reducing the cortisol response to shock. Rodin (1980) has presented an elegant series of studies with profound implications for the importance of control in human situations. As others had previously suggested (Birren, 1958; Gould, 1972), Rodin hypothesized that the transition from adulthood to old age may represent a loss of control. She further argued that the ability to sustain a sense of personal control in old age
Page 17 may be greatly influenced by societal factors, and that this, in turn, could affect the physical well-being of an aged individual. In order to investigate this problem, she introduced a program of "coping skill training" in a nursing home population, which was intended to reintroduce control to a group of individuals who felt a high degree of helplessness about their situation as a function of being institutionalized. There were two impressive aspects of the cortisol data presented in this study. First, among the aged in institutions, cortisol levels appear to be chronically elevated, similar to those observed in depressed individuals. Perhaps more important, however, was that individuals who were trained in certain aspects of coping skills (including control) showed a marked and significant reduction of their elevated cortisol levels. The concept of control in reducing stress in humans can be observed in other situations as well, such as the work environment. Rose and associates (Rose et al., 1982a,b,c) investigated a variety of endocrine parameters, including growth hormone and cortisol, in a large group of air traffic controllers during and after the work day. The job demands placed upon air traffic controllers have been considered to be extremely stressful. Blood samples were obtained automatically at 20-minute intervals over a 5-hour period from a large group of working air traffic controllers in their routine occupational environment. The data indicated that cortisol and growth hormone levels were not appreciably elevated and that both hormones showed a lack of consistency across repeated studies in terms of average levels or measures of episodic secretory activity. Thus, there appeared to be little in the way of an increased stress physiology under working conditions which were presumed to be stressful. It is important to note, however, that the population selected for study was composed of highly
Page 18 experienced individuals who had been on the job an average of 11 years. One could conclude, therefore, that as a consequence of their extensive work experience these individuals had developed adequate coping mechanisms, particularly in their ability to exercise control over the environment, which enabled them to minimize the psychological consequences of their stressful occupation. FEEDBACK Although it is clear that control is a major factor involved in coping, the ability or efficiency with which an organism can cope is also dependent upon another factor--feedback. Feedback refers to stimuli or information occurring after a behavioral response has been made in reaction to an event. These stimuli may be used to convey information to the responding organism indicating that it has made the correct response to a noxious event, for example, or that the aversive event is terminated for at least some interval of time. According to Weiss (1971a,b,c), the amount of stress an animal actually experiences when exposed to noxious stimuli depends upon the number of coping attempts the animal makes (control) and the amount of relevant information the coping response produces (feedback). As the required number of coping responses increase and/or the amount of relevant feedback decreases, the amount of stress experienced increases. In an extensive series of studies, Weiss demonstrated that if two groups of rats were subjected to the same amount of electric shock, the severity of the ulceration was reduced if the animal could respond--avoid and escape--and if the situation had some feedback information, i.e., a signal following the termination of shock. Although feedback information usually occurs in the
Page 19 contexc of cc;~Lrol, namely, information about the efficacy of a response, it has been reported that feedback information, per se, even in the absence of control, can reduce the pituitary-adrenal response to noxious stimuli. Hennessy et al. (1977) reported that the presence of a signal following the delivery of shock resulted in a reduced adrenocortical response even in the absence of control. In contrast, the pituitary-adrenal response of animals given a random signal was not significantly different from those animals that had no signal at all. Although the human data concerning the role of feedback are not abundant, there is one large study (Ursin et al., 1978) that investigated the coping process in humans following repeated experience with parachute training. In that study, the hormonal and behavioral responses of a group of Norwegian paratroop trainees were examined following repeated exposure to jumping off a 10-meter tower on a guide wire. After the first jump experience, there was a dramatic elevation of plasma cortisol, but as early as the second jump, there was a significant drop to basal levels; thereafter, basal levels persisted on subsequent jumps. It is also important to note that the fear ratings changed dramatically following the first and second jumps, so that there was very little fear expressed after the second jump, even though there had been a very high rating of fear prior to the first jump. We believe that both aspects of the coping model presented by Weiss can be applied to this situation. The individuals were able to make appropriate responses; that is, after the first experience they had already improved their performance of the task. However, since the individuals were Jumping on a guide wire, performance was probably not the critical factor. The second aspect of the coping model, feedback, may be more important in this context. Although the situation was potentially dangerous
Page 20 and threatening, the trainees had gone through the experience and suffered no bad consequences. Thus, a maximum amount of feedback about the absence of danger in 8 potentially threatening situation became quickly obvious. When one closely examines the three factors--control, predictability, and feedback--that have been shown to be involved in the coping process, they all have common elements which can be viewed within the framework of the determinant of the stress response proposed earlier--uncertainty. Each of these factors, either acting alone or more probably in concert, appear to have the capacity to reduce uncertainty. Control provides the organism with the capacity to eliminate, or at least to regulate, the duration of the stimuli. Thus, the uncertainty involved in unpredictable and uncontrollable situations is reduced. We have discussed predictability in a previous section of this report, and by definition, it serves to reduce uncertainty. Feedback can also be viewed in terms of reducing uncertainty, since feedback provides information to the organism about the efficacy and success of the response being emitted. We can therefore speculate that any set of operations that reduces uncertainty, whether they be passive such as habituation, or active such as the utilization of control, predictability, or feedback, can lead to an amelioration or elimination of adrenal activation. SOCIAL SUPPORT There are now ample behavioral and physiological data which point to the importance of social relationships in determining an individual's ability to cope with stress. Hamburg and Adams (1967) emphasized the great need for the continuity of personal relationships when individuals are involved in crisis.
Page 21 There have been several studies th.'t have pointed directly to social support as a major determinant in specific health-related issues. Nuckolls et al. (1972) studied the relationship between social stress, as measured by the cumulative life-change scores of Holmes and Rahe (1967), and psychological assets relating it to the prognosis of pregnancy. These studies showed that women with high life-change scores before and during pregnancy had an excessive incidence of complications if their social support was low, while a high level of social support appeared to be protective against adverse outcomes. Cobb (1976) discussed further evidence showing that social support affected the length of hospitalization and the rate of recovery from illness, minimized the effects of retirement and bereavement, and helped one to endure the threat of catastrophe. Cobb also studied the effects of social support on the effects of job termination. One hundred men whose jobs were abolished were visted by public health nurses before and after their jobs had terminated, and periodically up to 24 months after being unemployed. The results showed a tenfold increase in arthritis (with two or more swollen joints) in men who had low social support, as compared to men who had a great deal of social support. The effects of social support on the pituitary-adrenal system have been inferred in the now highly cited studies of Bourne (1970, 1971) conducted during the Viet Nam war. There are two studies that are relevant to this discussion. In the first study, Bourne (1970) measured urinary levels of corticoids in helicopter medics who were involved in medical evacuation flights. The striking finding was that when 17-OHCS levels on flight days were compared to output on those days when the medics remained on base, there were no significant differences. In fact, their overall levels tended to be lower than comparable
Page 22 levels of recruits and of the general population of men in the United States. Although this study was not experimental, so that social factors could not be manipulated, Bourne hypothesized that a strong support group was an important social asset, and that one of the defenses being used by these helicopter medics was a deep belief that their task was worthwhile, as observed by the gratitude of the men they rescued. This feeling, combined with a sense of personal worth, was socially validated by the frequent medals and the additional merit pay they received. In the second study (Bourne, 1971), a group of men in the Special Forces who had intercepted a message that their isolated camp was to be attacked by the Viet Cong in a few days, were evaluated. Again, 24-hour urine samples were obtained from the men in camp, both before the expected attack and several days following. Although the attack actually did not occur, anticipation of the attack resulted in different levels of urinary 17-OHCS excreted in some of the men. Most of the Special Forces enlisted personnel were enthusiastic about the impending attack and spent much of their time in task-oriented activities, such as bringing extra ammunition, fortifying their defenses, etc. Their behavior and attitude was in sharp contrast to the captain and the radio operator, who spent much of their time communicating with the commanding officer at battalion headquarters, the captain under considerable pressure to perform well. Both the radio operator and the captain showed sharp increases in corticoid excretion on the day of the expected attack, compared to the days preceding and following this day. In contrast, most of the enlisted men actually showed a suppression of adrenocorticoids on the expected day of attack, which returned to basal levels in the following days. According to Bourne' the problem confronting the radio
_ Page 23 operator and the captain was that the orders being issu~d~might not be relevant to the present situation, so that the officer had to compromise between satisfying his superiors and not disturbing the experienced soldiers with orders that they would regard as inappropriate. Bourne noted that the officer's duty did not permit him to keep busy with emotionally satisfying manual chores of the camp. In addition, the demands made upon the young officer to maintain his role as an authority figure over the older and experienced enlisted men was a tension-inducing influence that could have elevated his adrenal output. Once again, according to Bourne, the most important asset is the group consensus as to how a stimulus should be perceived. Although Bourne implicated social variables in the modulation of the adrenocortical response under these extremely stressful conditions, it is clear that the data are open to other interpretations. One could argue that in the case of both the helicopter medics and the enlisted men, they were able to exercise a large degree of control, and that this control may be the more important determinant of their low level of adrenal activity under these circumstances. There are available experimental data that clearly implicate the role of social factors both in accentuating and modulating the pituitary-adrenal - response. In order to study the influence of social behavior on hormonal systems, it is of course essential to investigate an animal whose predominant way of life is social. There are many forms of social organization and social behavior among animals, but it is in the primate species that the most striking feature of adaptation appears to be living in groups. Hans Kummer, a leading primatologist, has stated, "...primates seem to have only one unusual asset in coping with their environments: a type of society which, through constant
Page 24 association of young and old and through a long life duration, exploits their large brains to produce adults of great experience. One may, ther=t'ore, expect to find some specific primate adaptations in the way primates do things as groups" (1971, pp. 37-38). In several experiments using a small South American primate, the squirrel monkey, we have found that group living can serve to reduce the individual's hormonal responses to events outside the group (Vogt et al., 1981). Animals in groups communicate, and communication about available resources facilitates the survival of all group members. In addition, one of the primary adaptive functions served by groups is that gregarious animals are better at discovering, mobbing, and chasing a predator than single individuals. Further, by existing in a stable social situation in which it is possible to rely on the reactions of familiar social partners, the individuals may also show a reduced stress response. Nonhuman primates generally exhibit strong behavioral reactions indicative of fear when exposed to a snake or snake-like object (Vogt et al., 1981). In order to determine whether a group could serve as an effective modulator of stress responses, we exposed squirrel monkeys to a live boa constrictor, which was presented above their cage in a wire mesh box. Monkeys living in social groups consisting of males and females were exposed to the snake for 30 minutes while in a group, and also after being removed from the group and placed in an individual cage. An empty box, similar to the one containing the snake, was placed on top of the monkeys' cage on different days to control for the effects of general disturbance. All of the monkeys showed increased vigilance, agitated activity, and total avoidance of the stimulus box when the snake was-presented in either the group or individual condi~cion. However, while the presence of the snake
Page 25 consistently produced a behavioral response, it did not elicit adrenal activation in the monkeys when they remained in the group. Thus, it did appear that the social group could reduce the physiological response to a potentially threatening stimulus. Of equal importance was the fact that when the monkeys were separated from the group and placed in an unfamiliar cage, they showed striking elevations in plasma cortisol which were as great as those produced by exposure to the snake while alone. Whether these elevations in plasma cortisol following removal from the social group represent a response to loss of social partners or a response to novelty was difficult to specify in this experiment, but it is clear that the individual is much more prone to stress following disruption of group integrity. In additional experiments, we have subjected animals living in social groups to a variety of potentially stressful stimuli, such as a strange mobile robot, an unfamiliar Nonspecific, loud noise, etc. In none of these conditions while the animals were living as a group in their home environment were we able to elicit a pituitary- adrenal response . In 8 further study, social buffering, the apparent capacity of group membership to reduce adrenocortical responses to stress, was investigated in squirrel monkeys using a conditioning paradigm which involved classical conditioning or cortisol secretion. Adult males were assigned to two groups: one group received pairing of conditioned stimulus (CS) with footshock; a control group received the conditioned stimulus without shock. All animals were then tested with presentations of the conditioned stimulus without shock under three social housing conditions. In four successive phases of the experiment they were tested either individually, as dyeds, or within a social group, and then retested when housed individually. Neither group showed a cortisol response to the CS
Page ~5 prior to training. Following training, CS-evoked elevations of cortisol were found only in individually-housed animals. Animals housed in a group did not show the cortisol response elicited when individually housed. These results are consistent with the findings discussed above that the presence of familiar conspecifics can ameliorate a neuroendocrine response to psychological stress (Stanton et al., 198S). One of the important aspects of social support systems which differs from ordinary coping mechanisms is that these systems do not require the organism to deal directly with the aversive event, per se. How then is it possible to fit the data on social support within the general context of the propositions that uncertainty leads to an activation of the pituitary-adrenal response, and that reduction of uncertainty reduces or eliminates this response? Cobb (1976) discusses social support as a moderator of life stress and sees it as providing information that falls into three categories. The first leads the subject to believe that he is loved and cared for. The second leads a person to have higher self-esteem as a result of public expression of approval. The third is the perception of social congruity derived from a shared network of information and mutual obligation in which each member participates, and the common knowledge is shared and accepted by all. It is possible to speculate that all three types of information derived from social support serve to minimize an individual's level of uncertainty about the situation. Thus, the availability of stable and familiar social relationships can provide a set of predictable outcomes due in part to the long history of previous interaction and experience.
Page 27 These hypotheses would lead to the prediction that an unfamiliar social group would provide none of these beneficial features and, in fact, would constitute a condition of high uncertainty and therefore lead to an elevation of pituitary-adrenal activity. In adult organisms, we have also found that one of the most reliable elicitors of increased pituitary-adrenal activity is the formation of new social groups. We have demonstrated in several experiments that there is a a striking elevation of plasma cortisol when animals are placed in a new social group and, in fact, this elevation of cortisol can continue for several months (Coe et al., 1983a; Gonzalez et al., 1981~. CATECHOIAMINES Thus far, we have emphasized the importance of psychological factors in regulating one of the major stress responses to the pituitary-adrenal system. However, there is extensive literature on changes in peripheral and central noradrenergic systems during exposure to various stressors. The purpose of this report is not to review the data on catecholamines, but to attempt to further develop the hypothesis that various physiological systems respond specifically to psychological conditions, and that the psychological conditions required for the elaboration of one endocrine response are not the same as those required for other endocrine responses. The only way to examine specificity is when several hormone systems are examined within the same experimental paradigm. Thus, it becomes possible to determine whether both hormone systems are responding to the same stimuli or whether there are different responses depending on the psychological conditions.
Page 28 Several experiments have examined both plasma catecholamine and corticosterone changes. In a recent study, Swenson and Vogel (1983) studied these systems in rats that were exposed to footshock under two conditions. In the first condition the animals were able to escape the footshock, whereas in the second condition the shock was inescapable. Several measures were taken in this study, which included peripheral and central catecholamines as well as corticosterone. For this report, however, we will examine only the corticosterone response and the plasma epinephrine response during shock and only for the 30 minutes the animals remained in the apparatus after the shock was terminated. In this particular experiment, both plasma epinephrine and corticosterone were elevated after footshock. The only significant difference between these two measurements was that, while the catecholamine response was diminished 30 minutes after the termination of shock, plasma corticosterone tended to remain elevated. Based on this experiment, catecholamines and plasma corticosterone appear to reflect, with only a minor difference, similar response characteristics to the presence of footshock. However, there are several studies indicating that the pituitary-adrenal response and the catecholamine response are indeed very different. In the study of men going through paratrooper training (Ursin et al., 1978), both plasma cortisol and urinary epinephrine were examined during the course of the training. The plasma cortisol was rapidly reduced to basal levels after the first Jump off the tower. However, although the magnitude of the epinephrine response was also diminished after repeated jumps, there was still a significant elevation of epinephrine even at the last jump' whereas the response of the pituitary-adrenal system had long ceased to be significant. Utilizing different
Page 29 experimental paradigms, Frankenhaeuser (1980) studied urinary catecholamines and cortisol in subjects who performed a choice reaction task at the individual's preferred pace. The data indicated that under these conditions, both epinephrine and norepinephrine were significantly elevated above baseline, whereas cortisol was significantly reduced. Thus, as in the study on paratroopers, a presumed dissociation between the pituitary-adrenal response and the response of the catecholamine system is evident. Finally, Dantzer and Mormede (1985) demonstrated once again a dissociation between indices of catecholamine activity and the activity of the pituitary-adrenal system. Thus, animals exposed to footshock were compared with animals that were permitted to exhibit aggressive behavior when the footshock was presented. Weinberg and Levine (1980) previously reported that the ACTH response in animals that were shocked and fought, compared with those only shocked, was significantly reduced in the fighting animals. Dantzer and Mormede report that plasma corticosterone was also significantly reduced in animals tested after a repeated series of electric shocks when fighting was permitted. The animals not permitted to fight (controls) showed no such reduction. However, when catecholamine activity was examined via tyrosine hydroxylase, the tyrosine hydroxylase was significantly elevated in fighting animals compared with control animals. Although Dantzer and Mormede present these data as an apparent paradox, I believe this can be resolved if the catecholamine response is based on a set of psychological events different from those regulating the corticosterone response. Thus the catecholamine response seems to be more in tune with processes related to attention and vigilance; i.e., when the task requires the subject (human or animal) to be vigilant and attentive to specific stimulus
Page 30 events, the c.atecholamine response appears to be elicited. On the other hand, due to the reduction of uncertainty in response to the information gained by repeated exposures, the cortisol response tends to be diminished. How then do we resolve the results of the above-described experiments with the results reported by Swenson and Vogel (1983)? It can be argued that during the initial phases of any stressful experience (those closely related to the alarm reaction of Selye, 1936), many systems operate simultaneously. Thus, arousal is increased as a result of uncertainty and, in the case of catecholamines, attention and vigilance are hypothetically increased. However, when the subject is given repeated experiences in which feedback information is presented and coping mechanisms are evoked, the response of the pituitary-adrenal system either diminishes or ceases. In contrast, it is assumed that these systems still require some attention and vigilance on the part of the subject, and therefore the catechol~mine system continues to be responsive. STRESS AND PERFORMANCE Any attempt to discuss the broad range of effects of stress on performance, both in animals and humans, would clearly represent a limited view of what is an extensive area ranging from performance in laboratory animals to industrial productivity. Any discussion of the effects of stress and performance must discriminate between two classes of events: those which involve the proactive effects of stress on performance, that is, the effects of some prior stressor on subsequent performance characteristics, and those which are inherent in the task itself. Although we can arbitrarily view these as different phenomena, they do have many characteristics in common. The notion that continued exposure to a
Page 31 stressor may produce effects that appear only after the stimulation is terminated have been prevalent in the stress literature for many years. This assumption is derived principally from an adaptive cost hypothesis which suggests that, although humans and animals can often adapt to extreme conditions, there are cumulative costs to that adaptation. An early form of this hypothesis which emphasizes the biological cost of these adaptive processes was offered by Selye (1956), who proposed that after prolonged exposure to a stressor, one's adaptive reserves are drained, resistance breaks down, and exhaustion sets in. Others (Basowitz, et al., 1955; Dubos, 1965; Milgram, 1970; Uohlwill, 1966) make-similar points with regard to the poststressor effects on behavior. Dubos states, Although man is highly adaptable and can therefore achieve adjustments to extremely undesirable conditions, such adjustments often have...indirect effects that are deleterious" (1968, p. 139). Consistent with our previous views concerning the importance of control in reducing uncertainty, and thus resulting in effective coping, one common characteristic which pervades the literature on the aftereffects of stress is that the stress normally involves prior exposure to uncontrollable noxious events. Thus, if control is biologically important to organisms, its absence may be viewed as deleterious. The most important contribution to this hypothesis has derived from relatively recent literature which has been conducted within the framework of the Learned helplessness hypothesis. In this section, we will attempt to review the effects upon subsequent learning and performance proactive effects. which have been pursued within the context of testing learned helplessness theory. In the classic proactive interference experiment, Overmier and Seligman (1967) studied the effects of exposure to inescapable shock on subsequent avoidance
Page 32 learning in dogs. They began by placing naive animals in a restraining hammock enclosed in a soused- attenuated cubicle . The dogs received 64 5 - second electric shocks of moderately high intensity delivered on the average of 1/minute. After 24 hours the animals were then placed in a standard two-way shuttle-box apparatus (designed to test conditioned avoidance learning) and given signal avoidance training. The basic observation was that the dogs pretrained in the hammock with inescapable shocks later failed dramatically in this avoidance task, a task which is learned relatively easily by animals not preexposed to the uncontrollable stress. There were three features prominent in this failure to learn: (1) in addition to not learning to avoid, these animals also did not learn to escape a moderately intense shock; (2) even when the occasional escape response or avoidance response did occur, it did not lead to an increased probability of future escape-avoidance behavior, contrary to a pattern observed in normal dogs; (3) the dogs appeared passive, showing much less persistence in struggling and vocalization in the continued presence of the shocks than did naive dogs. However, it became very clear that this was a robust phenomenon and one which not only was specific to the failure to avoid an electric shock, but appeared to have much more general effects on performance. What was important in these studies was that the prior exposure to shock was uncontrollable. Seligman and Mater (1967) demonstrated that uncontrollability of the pretreatment shocks caused the interference effect by comparing two groups. One group received a series of escapable shocks; the second group received a series of inescapable shocks, the duration of which was determined by the first group, i.e., this group was ~yoked. to escape subjects. The addition of a third group, naive with respect to shock, into the treatment phase constituted what is now called the triadic design. It
Page 33 was determined that the group whic',, received inescapable shock, even though the duration of the shock was identical to those that escaped, were the group which showed the most profound proactive interference. Thus far the discussion has been limited to instances of exposure to inescapable shock in one situation and disrupting ahock-produced performance in another. One concern is how general such phenomena might be with respect to other combinations of both aversive and positive events. We can observe that four classes of interaction may exist. The shock-shock interaction is a special case of one of these; in particular, of the interaction between the experience of one aversive event, and learning based on the second aversive event. If one substitutes events of two different hedonic qualities, positive or negative, for either the first or the second event, four classes of interaction can be obtained in which the exposure to uncontrollability might modulate subsequent performance. These are shown in Table 1. Furthermore, within the aversive-aversive category, similar effects of uncontrollability can be found when stimuli other than electric shock are utilized. This is clearly relevant to the human performance field, where most of the stressful events would indeed not be restricted to the laboratory model of stress-shock. An example of the substitutability of uncontrollable aversive events on performance under different aversive conditions was provided by Altenor et al. (1977). In this experiment, a group of rats were initially exposed to either footshock or submerged in very cold water. For half the rats exposed to each aversive condition, the event was either escapable or inescapable. Ultimately, half of each of these groups were tested in an avoidance task for shock-escape learning, while the other half were tested in an underwater maze for water-escape
Page 34 learning. The subjects from each of the inescapable treatment conditions were significantly impaired for performance on the test task than the corresponding escapable pretreatment groups. Of more importance is the fact that the magnitude of the aftereffects appeared no smaller when the aversive stimuli were different in the pretreatment tests than when they were the same. Rosellini (1978) has reported an experiment which demonstrates the aversive-appetitive interaction. In two experiments, prior exposure of rats to inescapable shock interfered with the subsequent learning to press a lever to earn food in the second chamber. With respect to the appetitive-appetitive class of interactions, Uelker (1976) and Wheatley et al. (1977) have presented Two examples, the first with pigeons, and the second with rats. In both experiments, three groups differing in their initial treatments were used: (1) a response-dependent reinforcement group; (2) a matched group which received reinforcers independently of behavior; and (3) a naive control group. After several sessions with the initial treatment, all groups were required to perform a different task to obtain reinforcers. In both experiments the group initially exposed to the response-independent uncontrollable food presentations were very much slower to learn the required response. Even recognizing the powerful effects that stem from exposure to noncontrollable events, the class of interactions that has produced the most surprising results is the appetitive-aversive one. Rosellini ant Bazerman (1976), for example, found that rats exposed to 17 days of noncontingent food deliveries were significantly impaired relative to control groups in shock-escape responding in a shuttle-box. Others have obtained results congruous with these (Goodkin, 1976; Might and Katzev, 1977~. The effects of experience on controllability must be powerful indeed for initial appetitive treatments to result in interference with escape from life-threatening stimuli.
Page 3 5 STRESS AND PERFORMANCE IN HUMANS That the effects of uncontrollable events are obtainable across all classes and combinations of both positive ant aversive events attests to the broad generality ant biological significance of uncontrollability as a behaviorally debilitating experience. It is not surprising, therefore, that a phenonemon which has such a broad generalization within the animal literature should also have received substantial verification in the human literature under both experimental and naturalistic conditions. However, it is relatively recently that the first experimental studies of the post-exposure effects of stress on behavior were reported (Glass and Singer, 1972). Based on previous research which demonstrated that the ability (or perceived ability) to predict or escape an aversive event reduced both the aversive quality of the stimulus (Corah and Boffa, 1970; Pervin, 1963) and the resultant physiological response (Champion, 1950; Corah and Boffa, 1970; Stotland and Blumenthal, 1964), these workers hypothesized that performance following stress exposure may be similarly mediated by stressor predictability and perceived control over stressor determination. Glass and Singer's original work was influenced by the adaptive cost hypothesis. They suggested that detrimental effects on performance following exposure to unpredictable uncontrollable stressors, (which occur due to the substantial effort required to adapt to these aversive events) would leave one less able to cope with subsequent demands and frustrations, because predictable and controllable stresses were viewed as less aversive adaptations of these stresses, would presumably require less effort and, therefore, would be less likely to impair post-stimulation performance. Although the adaptive cost hypothesis did
Page 36 not prove to be tenable, Glass and Singer did conclude on the basis o. their work that exposure to unpredictable uncontrollable stress produces post-st.muiation deficits in performance on a number of tasks, and that the ability to predict and control stressors ameliorates these deficits. It is not the purpose of this report to comprehensively review all of the studies that have shown deficits in some aspect of performance as a consequence of prior exposure to uncontrollable and unpredictable stress. However, some examples of the type of research conducted in this area and a review of the findings would suffice. Using noise as a stressful stimulus, Glass and Singer (1972) reported a number of studies that examined post-stimulation effects after exposure to unpredictable uncontrollable noise (pp. 47, 50, 52, 55, 80). These studies typically involve approximately 25 minutes exposure to 108-110 dB random in~cermit~cent bursts of broad band conglomerate noise made up of a nuder of typical urban sounds. During noise exposure, the subjects worked on simple cognitive tasks. Autonomic response was monitored during stressor exposure. Immediately after noise exposure, one or more of three measures were administered to the subject: tolerance for frustration tasks (Feather, 1961), a proof-reading task (Glass and Singer, 1972), and the Stroop (1935) Color-Word task. The Feather test requires a subject to work on two soluble and two insoluble line puzzles for 15 minutes. The subject can only work on one puzzle at a time and cannot return to a puzzle after moving on to the next. The puzzles are presented so that the first and third are insoluble and the second and fourth are soluble. The criterion measure for amount of tolerance of frustration is the number of trials, puzzle cards, or amount of time spent on insoluble puzzles. The proof-reading task involves correcting misspellings, grammatical mistakes,
Page ~7 incorrect punctuations, transpositions, and typographical errors. Each subject is usually given 8-15 minutes, and the quality of performance is measured as the percentage of errors not found out of the total number of errors that could have been detected at the point the subject was told to stop. In the Stroop task, the stimuli are names of four colors--green, red, orange, and blue--each of which is printed in one of the other three colors. For example, the word preen may be printed in red, orange, or blue. The four color words are randomly presented over a series of trials, and the subject is asked to name the color and the word which is printed. The control version of the task, in which the subject is required to name the colors of a set of asterisks or zeros, is also administered to each subject. Stroop interference scores on accuracy and speed are obtained by subtracting the subject's score on the control stimuli from the subjects's Stroop scores. Post-stimulation deficits in performance occurred in all of the studies in all three of the tasks. The data presented by Glass and Singer appear to be highly reliable and have been replicated by several other authors (Gardner, 1978; Rotton et al., 1978; Wohlwill et al., 1976). Although the Glass and Singer research suggests that post-stimulation effects occur only following unpredictable noise, at least two studies reported similar deficits following exposure to high intensity steady-state continuous noise. Such steady-state continuous noise presumably does not have unpredictable components (Broadbent, 1979; Hartley, 1973). Thus, the data on post-stimulation effects of noise producing deficits on performance appear to be consistent for variable continuous and steady-state continuous noise, as well as for unpredictable noise. .
Page ;8 Although noise has received a considerable amount of attention in the experimental literature (due to the ability to control this particular variable more closely than others), there is evidence that other stressful conditions do produce post-exposure effects. Studies have indicated that subjects who have experienced high taskload perform more poorly following task completion than those who have been exposed to low taskload. Cohen and Spacapan (1978) found in a forced reaction time experiment that those required to respond to 100 lights/minute had less tolerance for frustration following task completion than those responding to 50 lights/minute. Hartley (1973) reports that those required to perform a serial reaction time task for the first 20 minutes of an experiment perform more poorly on the same task in the last 20 minutes than those who simply read during the initial phase of the study. Rotton et al. (1978) found that the subjects who expected to be required to recall a speech, even though never actually required to do so, showed lower tolerance for frustration following task completion than those subjects not expecting a recall test. Crowding also appears to effect subsequent performance and to produce post-exposure deficits. Studies on the effects of crowding on human behavior have found it necessary to distinguish two kinds of crowding conditions: spatial density and social density. Spatial density is manipulated by varying available space but keeping the number of people constant; social density is manipulated by varying the number of people occupying a fixed quantity of space. As an example of the aftereffects of spatial density, She rood (1974) had groups of female high-school students perform a number of tasks in either ~ large or a small room. After 1 hour of exposure the subjects were moved into a larger area. Each student was then tested at her own desk on tolerance for frustration at a
Page 39 proof-reading task. Those subjects who had been working in a high-density small room showed less tolerance for frustration than did their low-density large room counterparts. However, there were no differences on the proof-reading task. Some of the post-crowding deficits varying spatial density have been reported by Evans (1979) and Aiello et al., (1977). The existing studies in which spatial density was varied indicate that exposure to high-density conditions produced post-exposure deficits on the limited number of tasks investigated. However, the few studies which have varied social density provide evidence indicating that, although social density may produce post-stimulation deficits on performance, these effects interact with a number of other variables. Saegert et al. (1975) required male and female subjects to do a number of tasks in a railway station during crowded and uncrowded times of the day. Whereas females who had been exposed to high levels of density performed more poorly on the S troop test than did their low-density counterparts, males performed better after high than after low density. Glass and Singer also have reported that there are post-exposure deficits in performance following a variety of other stressful conditions, which include electric shock, a frustrating experience with a bureaucracy, and an experience of arbitrary or sex discrimination. The previously cited studies provide evidence for both the reliability and generality of the post-exposure effects of stress on performance. These effects have appeared in the vast majority of studies and these studies have used a wide range of stressors. The data suggest that the effect is most likely to occur when the stressor is clearly unpredictable and when a sensitive aftereffects measure is used. Moreover, the factors that might mediate the stressfulness of the situation, i.e., subject gender, perception of
Page 40 control, individual need for personal space, all appear to be important determinants of whether a particular manipulation will produce a significant deficit in subsequent performance. Within the context of coping theory, control, predictability, and feedback are considered to be important psychological dimensions which lead to a reduction in the psychological and biological responses to stress. The data in the stress and performance field are almost unanimous in supporting the role of both perceived and implemented control in ameliorating the post-exposure effects of the stressful experience. In some cases, groups with control performed as if they were not exposed to a stressor (Gardner, 1978; Glass et al., 1973); whereas in others, Sherrod et al. (1977) control provided some improvement in the post-stress task performance but did not completely ameliorate the effect. What does appear to be particularly cogent is the control, or perceived control, over the termination of the stress. There is a single study that indicates that initiation control, that is, the ability to control the onset of the stress, similarly produces an amelioration of the post-stress performance deficits. Providing someone with more than one kind of control does appear to be more effective than only having control over termination of the stressors. The role of predictability in reducing the aftereffects of stress has not received much attention, although the existing evidence does indicate (Glass and Singer, 1972) that the post-exposure deficits in performance are more likely to occur following exposure to an unpredictable, rather than to a predictable stress. In view of the abundance of data produced experimentally indicating post-exposure effects on subsequent performance, it is not surprising that investigators have also studied the effects of these stressful experiences in
Page 41 more naturalistic settings. Several investigators have studied the effects of long exposure to community noise on the performance of elementary school students. In one exemplar study, Cohen et al. (1973) tested children living in an apartment building built on bridges spanning a busy expressway. When tested in a quiet setting, children who lived in noisier apartments showed greater impairment of auditory discrimination and reading ability than those who lived in quieter apartments. The length of time in residence increased the magnitude of the correlation between noise and auditory discrimination. A subsequent study of children attending school in the air corridor of a busy metropolitan airport (Cohen et al., 1980) indicated that children living and attending school in the air corridor were poorer on both a simple and a difficult problem-solving task and were more likely to give up the task than their quiet-neighborhood controls. There are several studies which have also investigated the naturalistic effects of crowding. Rodin (1976) reported that 6- to 9-year-old children from high-density apartments of a low-income housing project were less likely than children from less dense homes in the same project to control (choose) their own outcomes on a reward task. In a subsequent study, 8th grade children from high-density apartments were more adversely affected by a learned helplessness pretreatment (insoluble puzzles) than were their low-density counterparts. These effects persisted even after statistical control for social class and race were used. Although these naturalistic studies on crowding do not specifically address the issue of stress on performance, they do indicate chronic effects of crowding on certain aspects of human behavior.
Page 42 1!} summary, the following conclusions can be generated from the material whoosh we have just reviewed on the post-exposure effects of stress on performance. (1) The post-exposure effects of unpredictable uncontrollable stress on performance have been replicated in many different laboratories and with a variety of subject populations. They occur as a result of a wide range of stressors, such as noise, electric shock, social density, etc. Interventions that increase control and predictability are effective in reducing these effects. However, it is important to note that the laboratory research has used a limited number of tasks with which to measure the post-exposure effects on performance. (2) Post-exposure effects of environmental stresses occur following prolonged exposure in naturalistic settings. These studies have suggested that the effects are mediated by helplessness. However, the studies in naturalistic settings--particularly those related to crowding--are not specifically directed toward performance deficits. (3) There is evidence that various forms of control have ameliorative effects similar to those of perceived control over the termination of the stressor. These include termination control, in which one actually performs a coping response, as well as initiation, choice, and information control. There is some evidence to suggest that combining more than one form of control will further improve post-exposure performance. This improvement, of course, cannot exceed the performance level reached by the no-stress conditions. Thus far, most of the data we have presented in this report on the relationship between stress and performance indicates that stress has a negative effect and results in performance decrements. However, it would be erroneous to conclude that there is always a negative relationship between stress and
Page 43 performance. Hennessy and Levine (1979) presented a comparison between the concepts of stress and arousal and concluded that the concept of stress might be subsumed under the umbrella of arousal theory. Both stress and arousal can be considered as representing a "state phenomenon" referring to the tonic nature of the effects. A distinction has been drawn between stimulus response (SR) or motor systems with direct pathways through the brain, and estates or arousal systems with diffuse central nervous system connections (Groves and Thompson, 1970; Thompson and Spencer, 1966). This proposition was derived from the results of studies of response plasticity within SR systems; when a motor response is evoked, SR and arousal systems are activated. The two types of systems appear to be independent, yet interaction is possible. One of the most pervasive findings in the literature with relationship to arousal and performance is that these relationships are curvilinear--that the relationship between arousal and performance is characterized by a U-6haped function, rather than being monotonic. One of the earliest studies to show an inverted U-shaped curve between stress and performance was that of Yerkes and Dodson (1908) who found that when animals are exposed to electric shock of medium intensity, they made fewer errors in learning than they did when exposed to weaker or stronger shocks. Finan (1940) also found that equated groups of rats deprived of food for 1, 12, 24 and 48 hours showed differences in conditioning strength in a Skinner apparatus when conditioning strength was measured in terms of extinction. The optimal deprivation interval was 12 hours. Conditioning strength was less for intervals shorter and longer than the optimal. Levine (1966) also reported a study in which, as electric shock increased in an avoidance learning paradigm, the ability of the animals to learn declined beyond
Page 44 optimal levels of the unconditioned stimulus electric shock. Within the human literature, Freeman (1940) studied reaction time in palmer conductance, an index of autonomic activity. Over a period of many days, 100 measures of palmer conductance and reaction time were taken. When the pairs of values were plotted, an orderly function appeared. Low palmer conductance (thus, less autonomic activity) was associated with slow reaction times, and high conductance was associated with fast reaction times. However, beyond the optimal range of palmer conductance, reaction time once again became slower. Thus, we have the classic demonstration of the relationship between arousal and performance, in this instance reaction time, resulting in the U- or inverted U-shape function. Reviewing all of the literature on U- or inverted U-shaped functions would represent another volume. The inverted U-function has been one of the most robust in the psychological literature. Classical activation theorists, such as Duffy, Lindsley, and Hebb have all stressed the importance of this curvilinear function. It is of interest that Pavlov in his writings contained reference to this phenomenon: for every one of our animals [dogs] there is a maximum stimulus, a limit of harmless functional strain' beyond which begins the intervention of inhibition [the rule of the limit of intensity of excitation]. A stimulus, the intensity of which is beyond that maximum, instantly elicits inhibition, thus distorting the usual rule of the relationship between the magnitude of the effect and the intensity of excitation; a strong stimulus may produce an equal and even smaller effect, than a weak one.... (Pavlov, 1930, p. 213, quoted by Malmo, 1958). Given the robust relationship between arousal (stress) and performance, it would certainly be erroneous to assume that stress invariably leads to a decrement in performance. There is clearly some optimal
Page 45 level of arousal which is required for organisms to perform. What is also apparent, however, is that beyond this optimal level of arousal, performance decrements are indeed the rule. However, perhaps the most difficult issue concerning the relationship between stress and performance emerges from the U-shaped relationship between arousal and performance. Individual differences in response to stress, both physiological and behavioral, emerge throughout all of the studies on stress and performance. It would appear, therefore, that what would be an optimal arousal level for one individual may indeed be detrimental to another. The number of variables which contribute to the marked individual differences in optimal arousal levels are perhaps the most difficult to specify. As mentioned previously, there are gender differences. In addition, prior experiences with either controllable or uncontrollable stressful events certainly appear to affect the relationship between stress and performance. There is a large literature which indicates that experiences occurring early in development can also markedly affect the relationship between stress and performance. Further, there is recent evidence that indicates that there may be strong genetic factors determining the response characteristics of individual organisms to stressful stimuli. Perhaps the only generalization that one can make is that individual differences pervade every step in the process of stress arousal/reduction. Lazarus (1974) has repeatedly stated that at the level of a specific individual, the problem is to determine what kind of stress evokes what kind of stress response in what kind of person. What does emerge from the theoretical considerations presented here and from the experimental data on the biological and behavioral consequences of stress, is that control is a major mechanism by which organisms can effectively deal with
Page 46 stressful conditions; and that the absence of control or loss of control can indeed have profound and permanent effects on individual performance on a wide variety of tasks.
