Abstract
Different conceptual perspectives converge to predict that if individuals are stressed, an encounter with most unthreatening natural environments will have a stress reducing or restorative influence, whereas many urban environments will hamper recuperation. Hypotheses regarding emotional, attentional and physiological aspects of stress reducing influences of nature are derived from a psycho-evolutionary theory. To investigate these hypotheses, 120 subjects first viewed a stressful movie, and then were exposed to color/sound videotapes of one of six different natural and urban settings. Data concerning stress recovery during the environmental presentations were obtained from self-ratings of affective states and a battery of physiological measures: heart period, muscle tension, skin conductance and pulse transit time, a non-invasive measure that correlates with systolic blood pressure. Findings from the physiological and verbal measures converged to indicate that recovery was faster and more complete when subjects were exposed to natural rather than urban environments. The pattern of physiological findings raised the possibility that responses to nature had a salient parasympathetic nervous system component; however, there was no evidence of pronounced parasympathetic involvement in responses to the urban settings. There were directional differences in cardiac responses to the natural vs urban settings, suggesting that attention/intake was higher during the natural exposures. However, both the stressor film and the nature settings elicited high levels of involuntary or automatic attention, which contradicts the notion that restorative influences of nature stem from involuntary attention or fascination. Findings were consistent with the predictions of the psycho-evolutionary theory that restorative influences of nature involve a shift towards a more positively-toned emotional state, positive changes in physiological activity levels, and that these changes are accompanied by sustained attention/intake. Content differences in terms of natural vs human-made properties appeared decisive in accounting for the differences in recuperation and perceptual intake.
Introduction
The growing interest in environmental stress has been accompanied by a rapid accumulation of evidence indicating that environmental stressors (e.g. crowding, community noise, air pollution) can elicit substantial stress in large groups of people (for surveys see Cohen et aL, 1986; Evans & Cohen, 1987). Much previous stress research has been concerned primarily with person-based variables in responses to situations, such as coping and perceived control. Many studies also reflect an increasing emphasis on physical properties of environments (e.g. high stimulation levels) that increase demands on coping resources (e.g. Evans et al., 1986). Whether concerned with person-based variables or environmental characteristics, the vast majority of research to date has focused on situations that challenge resources or threaten well-being and accordingly elicit stress. Also, most studies have been concerned with extreme or unusual environmental conditions such as heat stress or loud aircraft noise. A different, though complementary perspective on stress and environments is evident in the question of whether different everyday, non-extreme physical environments have different influences in terms of fostering or hampering recovery from stress (Ulrich & Simons, 1986). If individuals are experiencing uncomfortable stress, due to either environmental conditions or other factors (e.g. illness, bereavement), do encounters with some types of common environments have restorative effects, while other everyday settings hamper or even work against recovery? To help answer these questions, an understanding of the concepts of stress and restoration is required.
Stress is the process by which an individual responds psychologically, physiologically, and often with behaviors, to a situation that challenges or threatens well-being (Baum et al., 1985). The psychological component includes cognitive appraisal of the situation, emotions such as fear, anger, and sadness, and coping responses. The physiological aspect consists of activity responses in numerous bodily systems, such as the cardiovascular, skeletomuscular and neuroendocrine, that mobilize the individual for coping or dealing with the situation. This mobilization uses resources or energy and, if prolonged, contributes to fatigue. The behavioral component includes a wide range of manifestations for instance, avoidance, alcohol or cigarette use, and declines in cognitive performance on tasks such as proofreading (e.g. Cohen et aL, 1986). Also, after cessation of a stressor, after-effects may be observed such as a decline in frustration tolerance and lower task performance (e.g. Glass & Singer, 1972).
A second concept central to the theory and research described in later sections is 'stress recovery' or 'restoration'. These terms will be used here interchangeably, although restoration can be construed as a broader concept that is not limited to stress recovery situations, or to recovery from states characterized by excessive psychological
and physiological arousal, but could also apply to recuperation, for instance, from understimulation or excessively low arousal (Ulrich, 1981, 1983). 'Restoration' could
also pertain to recovery influences that extend to the anabolic recharge of energy expended in the psychophysiological mobilization involved in responding to a stressor. In contrast to a stress response, restoration or recovery from stress involves numerous positive changes in psychological states, in levels of activity in physiological systems, and often in behaviors or functioning, including cognitive functioning or performance. Central to the psychological component of restoration are positive changes in emotional states, i.e. reduced levels of negatively toned feelings such as fear or anger, and increases in positively-toned affects (Zuckerman, 1977; Ulrich, 1979). Because
responses to short-term as well as long-term fatiguing stressors sometimes involve declines in cognitive functioning or performance, recovery may be evident in gains in performance (Hockey, 1983; Hartig et al., 1987).
The fact that both stress symptoms and restoration appear in different response modes has led Baum et al. (1985), among others, to advocate strongly the use of research approaches that assess responses in more than one mode. An outcome of general synchrony between data obtained from different modes (e.g. psychological and physiological) would suggest convergent validity, and justify greater confidence in the findings. At the most general level, the purpose of the present study was to investigate, using a multi-modal combination of measures--physiological and verbal--the extent to which exposure to different everyday outdoor environments may foster or hinder recovery from stress. A second major objective was to test the notion that exposures to natural settings may promote greater stress recovery than contacts with urban environments, and that such differences in recovery effects should be evident in emotional states and physiological indicators. The physiological measures used in this research can objectively indicate effects of environmental exposures on bodily systems, and yield insights concerning attention or intake responses to settings during stress recovery (e.g. Lacey & Lacey, 1970). Compared with other types of stress measures, physiological procedures offer the important advantage of continuous monitoring of responses during an environmental encounter. Also, the use of physiological measures addresses a conspicuous gap in environment-behavior research; physiological methods have been neglected by investigators, and very little is known about the physiological correlates of experiences with everyday physical settings.
In investigating recovery effects of outdoor environments, this research distinguished broadly between natural and urban environments. This is justified by different theories of human-environment interactions, and by evidence from many studies indicating that natural vs human-made visual properties elicit different patterns of affective responses in unstressed individuals, and have a central role in influencing perception and categorization of outdoor settings (Kaplan et al., 1972; Wohlwill, 1983).
Investigations of groups in different countries have employed multivariate procedures such as multidimensional scaling to show that natural vs built groupings of landscape
scenes emerge as prominent dimensions when affective ratings are obtained for diverse samples of views (e.g. Kaplan et al., 1972; Bernaldez & Parra, 1979). Visual
environments tend to be categorized broadly as 'natural' by American and European groups if the content is predominantly vegetation and/or water, and if human-made
features such as buildings and cars are absent or inconspicuous (Ulrich, 1983). Accordingly, we selected for study examples of natural settings that were dominated either by vegetation or a water feature. Consistent with the general natural/urban distinction, various urban settings were selected that had little or no vegetation and lacked a water feature.
Previous findings concerning stress reducing influences of natural settings. Apart from the perceptual categories implicit in individuals' responses to outdoor environments, a consistent finding in well over 100 studies of recreation experiences in wilderness and urban nature areas has been that stress mitigation is one of the most important verbally expressed perceived benefits (Driver, 1976; Knopf, 1987; Schroeder, 1989). Recreation experiences are often complex (Hull, 1990), and apart from possible restorative influences of viewing or experiencing nature, other mechanisms foster stress recovery, e.g. physical exercise and achieving sense of control and other coping
advantages through 'temporary escape' (Driver & Knopf, 1975). Although this body of research offers convincing evidence that stress recovery occurs in recreation experiences, the contribution of natural content and configurations per se has not been clearly isolated. However, the notion that simply viewing unthreatening nature tends to foster recovery from stress has received empirical support from a study by Ulrich (1979) of students who were experiencing mild stress because of a final exam. This study used self-ratings to assess recovery produced by viewing color slides of either everyday nature scenes or unblighted city views lacking nature such as vegetation. Findings suggested that the natural scenes held attention more effectively and fostered greater recovery as indicated by higher levels of positive affects and greater reductions in fear. Honeyman (1990) replicated this study with the addition of a recovery condition consisting of urban scenes containing prominent vegetation. Her results suggested that the urban settings with nature produced more recovery than the urban scenes lacking nature. In a study performed in Sweden using unstressed subjects (Ulrich, 1981), findings from self-ratings suggested that slides of unspectacular natural scenes sustained attention much more effectively through a lengthy viewing session,
and produced more positively toned feeling states, than did urban scenes. These verbal results were broadly consistent with recordings of brain electrical activity in the alpha frequency range that suggested individuals were more wakefully relaxed during the nature exposures (Ulrich, 1981). These studies suggest that restorative effects of natural, in contrast to urban, scenes involve positive changes in emotional states accompanied by attention.
These positive emotional states elicited by viewing unthreatening nature may be a mechanism underlying the finding that hospital patients recovering from surgery had more favorable recovery ~ourses, including shorter hospital stays, lower intake of potent narcotic pain drugs, and more favorable evaluations by nurses, if their windows
overlooked trees rather than a brick building wall (Ulrich, 1984). A questionnaire study of patients who were severely disabled by accidents or illness (and presumably stressed) found that a highly preferred category of window views included scenes of natural content such as trees (Verderber, 1986). In research by Heerwagen and Orians on patient anxiety in a dental fears clinic (Heerwagen, 1990), self-ratings and heart rate data suggested that patients felt less stressed on days when a large mural depicting a natural scene was hung on a wall of the waiting room, in contrast to days when the wall was blank. These findings from studies of health facilities are paralleled by results from prison research suggesting that cell window views .of nature are associated with lower frequencies of prisoner stress symptoms such as digestive illness and headaches, and with fewer sick calls (Moore, 1982; West, 1985).
Summary and Discussion
The findings from the physiological and verbal measures converge to indicate that different everyday outdoor environments can have quite different influences on stress recovery. The results strongly support the conclusion that recuperation was faster and more complete when subjects were exposed to the natural settings rather than the
various urban environments. Regarding the physiological findings, there was impressive consistency across the stress measures (PTT, SCR, EMG) in indicating greater recovery influences of nature. Also, the heart period data revealed directionally different responses during recovery to the natural vs urban environments; this pattern of variation suggests that intake/attention was significantly higher during the natural exposures, a finding consistent with previous results obtained from verbal measures (Ulrich, 1979, 1981). The four physiological measures, including heart period, reflect activity in different bodily systems that are only weakly coupled or interrelated. The consistency across physiological indicators raises the possibility that differential influences of natural vs urban settings may tend to be widespread in bodily systems. In this regard, a previous study of unstressed subjects found that natural and urban scenes had different influences on activity in the electrocortical system (Ulrich, 1981).
Along with suggesting that nature elicited greater intake/attention, the overall pattern of heart period data raises the possibility that there was a salient parasympathetic component to responses to the natural but not the urban environments, especially during the initial minutes of recovery. (It will be recalled that the autonomic nervous system can be subdivided into the parasympathetic and sympathetic branches.) Because parasympathetic influences have a central role in attention and restoration, whereas sympathetic activation is central in the taxing mobilization involved in responding to stressors, future research in environmental psychology on restoration following stress should attempt to disentangle the influences of these different systems. Distinguishing more clearly between parasympathetic and sympathetic influences will require more complex physiological recordings (Papillo & Shapiro, 1990).
Results from the ZIPERS self-ratings suggested that the natural settings had more restorative influences than the urban environments on three affective dimensions:
Positive affects, Anger/Aggression, and Fear. The most salient difference was the comparatively much greater restoration in Positive affects associated with the natural exposures, a finding consistent with results from an earlier study of mildly stressed subjects (Ulrich, 1979). Traditional arousal theory accurately predicts that the surge in Positive affects produced by nature should occur in association with the observed sharp reduction in physiological arousal indicators. As Berlyne (1971) and others have
postulated, a lowering of arousal produces an increase in pleasurable feelings if an individual's initial state is stress or excessive arousal. Concerning this issue, a noteworthy finding was the pattern of significant correlations during stress recovery among the autonomic measures and every item comprising the Positive affects factor, indicating that Positive affects increased as autonomic, but not somatic (EMG), arousal lowered. Although only the Positive affects self-ratings correlated significantly with
physiological recordings, there was nonetheless broad synchrony between the verbal and physiological findings. In general, individuals exposed to the natural settings both
reported improved feeling states and evidenced lower stress levels in physiological indicators. Compared with the nature groups, subjects exposed to the urban environments experienced less recovery as evidenced both by self-ratings and physiological responses. The general pattern of convergence across data from different
response modes suggests that the main conclusion regarding differential recovery warrants considerable confidence. The findings indicate that restorative effects of everyday outdoor environments have, at the least, psychological and physiological components. Because stress often has behavioral manifestations, it seems possible that restorative influences of environments may also be expressed in behaviors or enhanced functioning. Future research should investigate, for instance, whether many natural
settings might have restorative effects that include increased performance levels on tasks requiring attention and cognitive processing (Hartig et al., 1987; Kaplan & Kaplan, 1989). In view of the finding here that restoration was evident in central nervous system indicators, future research should examine whether such effects are also expressed in the endocrine system as changes in stress hormones, and in indicators of enhanced functioning in the immune system (Kennedy et al., 1990). Additional studies are also needed to reveal the relative contributions of visual and auditory stimulation to restorative influences of everyday outdoor settings.
Apart from differences in levels of recovery, a noteworthy finding obtained from the continuous physiological recordings was the rapidity of recuperation during natural exposures. During the first four minutes of recovery, all groups experienced at least some degree of recuperation, but the nature groups achieved recovery approaching base-line levels in both autonomic and somatic activity. By five to seven minutes, significantly greater recovery for the nature groups, compared with subjects exposed to urban settings, was evident for all three physiological stress indicators (SCR, PTT, EMG). After the 10 rain recovery videotapes, self-ratings data for the nature groups evidenced restoration to the point that, broadly speaking, feeling states were slightly more positively-toned than during baseline. The quickness of recovery during the nature conditions raises the possibility that these laboratory findings might be found to apply in many real contexts characterized by short-term contacts with nature. In urbanized countries, the great majority of encounters with nature elements probably are short episodes lasting only several seconds or a few minutes. Common types of nature contacts for urbanities may include, for example, viewing trees through a window in a workplace or residence, lunching in a park, or driving through an urban fringe area where roadsides are undeveloped. The findings of the present study justify the speculation that these and other short duration nature exposures might have an important function for many urbanities in facilitating recovery from such stressors as daily hassles or annoyances. The results cannot be generalized directly to longer term nature exposures that involve active participation such as a wilderness back-packing trip. Nonetheless, the findings may have relevance for research that seeks to understand benefits of wilderness recreation, including why most wilderness users report that reduced tension or stress, or tranquility, are very important benefits of their experiences (e.g. Driver, 1976).
Implications for theory
Some aspects of the findings are consistent with the conceptual speculations of Olmsted, Kaplan and Kaplan and others, in the sense that strong intake/attention was a concomitant of restorative influences of the nature settings. However, the overall pattern of findings indicates that a theoretical perspective emphasizing attention or 'fascination' is inadequate for explaining restorative influences of nature. In this regard, the present findings showed that the stressor film also elicited strong 'involuntary' or automatic attention. The stressor film depicted simulated blood and mutiliation, stimuli which arguably should elicit involuntary attention or fascination because of their survival relevance during evolution (Hare et aL, 1970). This combination of findings is consistent with Ulrich's theoretical position (1983) that attention holding properties of nature work both ways; that is, involuntary attention can be a salient component of non-restorative and even stressful reactions to certain natural stimuli, and can also be a prominent component of restorative responses to unthreatening natural settings. Put differently, the findings concerning attention indicate that stress reducing or restorative responses to nature must involve more than involuntary attention or fascination. In this regard, the present physiological and Verbal findings are consistent with the predictions of Ulrich's psycho-evolutionary theory (1983) that, following a stressor, the restorative influences of exposure to nature involve, among other responses, a broad shift in feelings towards a more positively-toned emotional state, positive changes in activity levels in different physiological systems, and that these changes are accompanied by moderately high levels of sustained attention. As was emphasized in an earlier conceptual section, there are theoretical grounds for suggesting that a critical element in restorative effects of natural scenes is a quick-onset positive affective reaction which may have a central role in shaping the positive character of changes in psychological and physiological states, and functioning or behavior (Ulrich, 1983). This theoretical position is consistent with growing scientific evidence that the initial levels of responding to natural stimuli involve quick-onset affective responses that appear to be closely linked with attention and subsequent conscious processing (e.g. Ohman et aL, 1989; Dimberg, 1990). With respect to stress recovery or restoration, a positive initial affective response, comprised of liking and moderate to high interest, should motivate and sustain prolonged attention/intake, should produce higher levels of positive feelings, reduce negatively-toned or stress related feelings such as fear and anger, and suppress stressful or extraneous thoughts (Ulrich, 1981, 1983). This combination of positive influences should be associated with reductions in taxing physiological mobilization, and perhaps with involvement of the parasympathetic nervous system that would function to restore energy. However, if the initial affective response is negative (e.g. dislike, fear), as many studies have shown occurs for such natural stimuli or configurations as snakes and heights, the ensuing emotional/ cognitive/ physiological responding probably will be non-restorative and possibly even stressful, regardless of whether the response involves involuntary attention or 'fascination' (Ulrich, 1983).
How does arousal theory fare in accounting for the findings regarding both intake/ attention and recovery? Whereas arousal theory predicts the observed surge in positive feelings that accompanied sharp declines in physiological arousal indicators during recovery, the main findings are at odds with arousal-based perspectives. Central to the
arousal perspective in environmental psychology is the notion that various arousalincreasing properties (e.g. movement, quantities of stimuli) strongly influence attention, affective responses, and activity levels in physiological systems (e.g. Berlyne, 1971; Mehrabian & Russell, 1974). According to arousal theory, the urban settings with their movement, quantities of visual stimuli and comparatively variable, intense sounds, should have elicited far more intake/attention than the natural environments, yet they elicited much less as indicated by heart rate responses. The findings did not support our hypothesis based on arousal theory that attention/intake responses to the high stimulation urban conditions (heavy traffic, many pedestrians) would be greater than to the low stimulation urban conditions (light traffic, few pedestrians). Most fundamentally, contrary to the prediction of arousal theory, there was no general relationship between quantities of stimulation and recovery influences. Findings from most of the recovery measures suggested that similarly low levels of restoration occurred during exposures to the high vs low stimulation urban settings. An exception to this pattern was the SCR finding suggesting somewhat more recovery during the high rather than low stimulation urban exposures, a result that runs directly counter to the prediction of arousal theory.
In general, the findings strongly suggest that content differences in terms of natural vs human-made properties, rather than variations in stimulation levels, were decisive in accounting for the differences in recovery and intake/attention. To explain this outcome, it might be argued from a culture-based point of view that American society conditions positive response dispositions to nature and negative dispositions to urban environments. However, a cultural perspective could also lead to a very different
interpretation. For instance, it is certainly the case that each of the subjects in this study had previously been exposed by television and other media to thousands of advertizements touting the positive characteristics of automobiles; accordingly, the individuals arguably should have evidenced strong intake and positively toned emotional responses to the urban settings with traffic.
On the basis of the heart rate findings, it does not appear that the nature settings produced restoration because they elicited an elaborated, active process of cognition
entailing positive associations or memories. Such mental activity is often associated with an acceleratory component to the cardiac response; the overall deceleration that actually occurred during the natural exposures is perhaps associated with parasympathetic influences involving reduced baroreceptor feedback, and increases in perceptual activity and sensitivity to external stimuli (Lacey & Lacey, 1974). To the extent that cognition may have played a role in the restorative effects of nature, conceivably the forest and water settings might have elicited effortless, well-formed affective/cognitive structures.
Regarding the less positive influences of the urban environments, a learning-based speculation is that the settings with traffic and pedestrians might have elicited cognitive appraisals that included slight risk or tension. Such appraisals would tend to be associated with mild sympathetic mobilization that could contribute an acceleratory component to heart rate and somewhat inhibit recovery. In physiological terms, perhaps a slight defense reaction (i.e. heart acceleration, higher SCR) can be a component
of responsiveness to many urban settings. But this perspective also implies the possibility, for instance, that the natural setting with fast-moving water could elicit mild defense reactions that would hinder recovery.
Alternatively, it seems reasonable to propose that evolutionary influences, along with learning, may underlie the differential intake and recovery responses to the natural vs urban environments. In an earlier theoretical section it was proposed that the rewards associated with natural settings during human evolution have been sufficiently critical to favor individuals who quickly and easily learned, and persistently retained, two related types of adaptive positive responding to nature: restoration following stress or taxing activities; and, in the absence of stress, positive emotional/attentional/ approach responses to nature contents and configurations, especially those that favored well-being or survival because of such advantages as high food potential and low risk. From these arguments it follows that biologically prepared learning may be evident in positive responses to many unthreatening natural settings, but that biological preparedness should not be manifested in responses to urban or built stimuli. An evolutionary perspective implies that adaptive responses to unthreatening natural settings should include quick-onset positive affects and sustained intake and perceptual sensitivity; in physiological terms this would often be a parasympathetically dominated response that included heart deceleration--similar to the observed response to the natural settings in the present study. Arguably, it would have been exceedingly maladaptive for early humans to easily and quickly learn, and over time not forget, to respond with perceptual rejection and sympathetic mobilization to unthreatening ambient environmental content. Sympathetic and other physiological mobilization would have been fatiguing, and over a prolonged period would be linked with chronic cardiovascular and endocrine responses that adversely affected health. Further, perceptual rejection or low attention would have been maladaptive from the standpoint of the need to exploit food, water or other survival related advantages of a natural setting. By contrast, an adaptive, parasympathetically dominated response would be associated with sustained yet non-taking intake and perceptual sensitivity, and maintenance or restoration of energy. Perhaps this biologically prepared constellation of responses would not only foster psychophysiological restoration following, for instance, encounters with threatening objects or situations, but would also function to maintain an appropriate 'base-line' state of environmental intake/ sensitivity which would leave the individual primed to very quickly mobilize and respond effectively to a sudden threat. Following a prolonged episode of fatiguing mobilization, it is not too great a speculation to suggest that some unthreatening natural settings, such as the water and forest settings in the study, might elicit a parasympathetically dominated response similar to a mild, eyes-open form of 'relaxation response' or wakeful, meditation-like state (e.g. Benson et al., 1974). Likewise, Katcher et al. (1984) have suggested that meditation-like relaxation can be induced by visual contemplation of a different configuration of natural content, an aquarium with fish.
The roles of culture, learning and possibly biological preparedness in differential response to natural vs urban settings represent fundamental issues that should be addressed in future research. Further studies are also needed to investigate psychophysiological influences of natural and urban content across varied situations including, for instance, contexts where individuals initially are unstressed, or are bored or chronically under stimulated. Also, research should examine person-based variables (e.g. perceived control) that may influence the extent of recovery during encounters with natural and urban settings. The results of the present study clearly imply the need for a broader agenda regarding environmental stress research that goes beyond a focus on stressors to include concern for settings that tend to facilitate recovery. The findings strongly suggest that environments of importance to well-being and stress are not confined to settings having extreme or unusual properties, such as loud noise or extreme temperatures, but also include very common environments that most urbanites in developed nations encounter daily.