The Effects Of Aerobic Exercise On Cardiovascular, Facial Emg

The Effects of Aerobic Exercise on Cardiovascular, Facial EMG, and Self-Report Responses to Emotional Imagery ROGER B. FILLINGIM, PHD, DAVID L. ROTH, PHD, AND EDWIN W. COOK III, PHD This study examined the effects of exercise on subsequent psychophysiological and self-report responses to emotional imagery, using excitation-transfer theory as a guiding conceptual model. Twenty-four female undergraduates engaged in aerobic exercise (stationary cycling) for 15 minutes, and an equal number of subjects rested quietly for the same time period. All subjects then engaged in anger and sadness imagery trials. Cardiovascular, facial electromyographic, and self-report responses to the imagery were assessed. The results indicated that the subjects in the exercise group showed less peripheral vasoconstriction in response to the imagery than did the quiet rest subjects. Subjects in both groups displayed greater electromyographic activity in the depressor and zygomatic muscle regions during anger than sadness imagery, and subjects in the exercise group tended to show greater corrugator tension during sadness than during anger imagery. Few differences between the groups were found on self-report measures. These findings are discussed with reference to previous research, theoretical implications, and future directions.

INTRODUCTION

Many investigators have reported that exercise training leads to improved mood (1-6); however, the mechanisms whereby this change occurs are unclear. One possibility is that chronic exercise promotes psychologic health via the cumulative benefits of individual bouts of exercise (7). Consistent with this notion are findings that acute exercise decreases selfreports of tension and depression while increasing feelings of vigor (8-13). A related line of research has investigated the physiological effects of acute

From the University of Alabama at Birmingham, Birmingham, Alabama. Address reprint requests to: Roger B. Fillingim, Ph.D., The Pain Management Center, 3343 University Blvd. S., Jacksonville, Fl 32216. Received for publication January 28, 1991; revision received July 15, 1991

Psychosomatic Medicine 54:109-120 (1992) 0033-3174/92/5401-0109$O3 00/0 Copyright © 1992 by the American Psychosomatic Society

exercise. Decreased neuromuscular activity has been reported following acute exercise (14, 15). Also, several authors have reported decreased blood pressure after one bout of exercise (11, 16, 17). Other investigators have examined physiologic responses to stressful stimuli following an exercise session. Some studies have found that psychophysiological responses to mental tasks are not altered by single bouts of exercise (12, 18, 19). However, Peronnet and coworkers (20) found decreased adrenergic responses to physical and mental tasks following 2 hours of cycle ergometry. The discrepancies in these studies of physiologic reactivity could result from differences in the dependent variables, the exercise stimulus, the timing of post-exercise tasks, or the nature of the post-exercise tasks. These two lines of research suggest that acute exercise improves subjective mood and may reduce physiologic responses to psychological stress. 109

R. B. FILLINGIM et al.

Few studies have examined the subjective and psychophysiologic changes associated with acute exercise using an integrated model of emotional functioning. One relevant model is excitation-transfer theory (21). Excitation-transfer theory predicts that arousal from previous, potentially irrelevant stimuli can be misattributed to subsequent emotional stimuli, thereby enhancing emotional responses. Zillmann (21) has posited that such transfer of excitation will occur only to emotions characterized by autonomic activation (e.g., anger, fear, excitement) and not to less arousing emotional states (e.g., sadness). Additionally, misattribution is possible only if individuals remain unaware of their residual arousal. Several experiments have been performed in the context of excitation-transfer theory during which subjects first exercised and then their emotional responses to an unrelated stimulus were assessed. Findings have been consistent with predictions that exercise enhances aggressive behavior subsequent to provocation (22, 23) and in the absence of prior provocation (24). Additionally, exerciseinduced residual arousal has been found to enhance sexual excitement (25). These findings have also supported the proposition that individuals must be unaware of their residual arousal in order for excitation transfer to occur. One study found that transfer did not occur immediately after exercise, but only after a 6-minute delay (23). In the investigation of sexual excitement mentioned above, excitation transfer was found to occur only during the "residual phase" when subjects remained aroused physiologically but were unaware of this arousal (25). The present study contrasted predictions derived from excitation-transfer theory to alternative predictions based on 110

previous findings concerning the emotional benefits of exercise. In this experiment, subjects either engaged in a single bout of aerobic exercise or rested quietly for an equivalent time period. After this manipulation, all subjects engaged in anger and sadness imagery. Cardiovascular, facial electromyographic (EMG), and selfreport responses to imagery were assessed. Based on the assumption that residual arousal is misattributed only to emotions normally characterized by high subjective arousal and assuming that anger is characterized by greater sympathetic activation than is sadness, excitation-transfer theory would predict greater enhancement of responses to anger than sadness imagery (21). Alternatively, if exercise generally improves mood state through reduced psychophysiological reactivity to aversive stressors, then exercise ought to attenuate responses to both types of imagery in this paradigm. This study differs from previous work on excitation-transfer theory in that prior investigators have emphasized aggressive behavior as the dependent variable, which may differ from the subjective and physiologic concomitants of emotion. Additionally, earlier studies typically examined the emotion of anger, induced by shocking the subject. In the present study emotional imagery was utilized to induce both anger and sadness. METHOD

Subjects Subjects were 48 female undergraduate students who participated in order to fulfill course requirements. Females were used because they have been found to show greater subjective and facial EMG responses to emotional imagery (26). Equal numbers of subjects were randomly assigned to exercise and quiet rest groups. One of the 48 subjects failed to

Psychosomatic Medicine 54:109-120 (1992)

EXERCISE AND EMOTIONAL IMAGERY report for the second session due to facial irritation from the EMG skin preparation and was replaced with another subject. The mean age of subjects in the exercise and quiet rest groups was 19.7 (SD = 4.6) and 21.2 (SD = 5.0) years, respectively.

Procedure Subjects participated in two experimental sessions, one involving anger imagery and one involving sadness imagery. Sessions were separated by 7 to 14 days. The structure was the same for both of these sessions. Minor procedural differences are noted below. Upon arrival for each session, participants completed the Profile of Mood States (POMS) (27), which assessed overall mood. Then, after skin preparation, surface electrodes were placed over the right subclavicular space and a left intercostal space for measurement of electrocardiogram (EKG). A photoplethysmograph was attached to the last digit of the middle finger on the nondominant hand for pulse volume (PV) recording. Facial EMG electrodes were placed over the corrugator, zygomatic, and depressor muscle regions using placements described in Fridlund and Cacioppo (28). These sites were chosen because corrugator activity is generally elevated in negative affective states whereas zygomatic activity is reduced. Depressor activity pulls the mouth into a frown (29), which was hypothesized to be characteristic of sadness. EMG electrodes were placed on the left side of the face based on previously reported advantages in left-sided EMG intensity (30). A ground electrode was placed at the midline above the inner brows (28). All EMG electrode impedances were below 10 Kohms. To reduce reactivity to the facial EMG electrodes, subjects were informed that these sensors measured their brain waves. Subjects were then seated in a 1 x 1 meter acoustic isolation chamber for a 10-minute baseline followed by a practice neutral imagery trial to acquaint subjects with the imagery procedure. Cardiovascular responses were monitored throughout the baseline period. Next, subjects were moved to an adjacent room and either performed exercise or rested quietly for 15 minutes. The exercise stimulus was stationary cycling with the initial resistance set to 300 kilopond meters per minute. The resistance was increased gradually throughout the first 6 minutes of exercise as needed to maintain a heart rate (HR) between 140

Psychosomatic Medicine 54:109-120 (1992)

and 160 bpm. Quiet rest subjects were instructed to sit and relax until the experimenter returned. Following the 15-minute experimental period, subjects returned to the recording chamber. After an 8-minute rest period, the two emotional imagery trials were conducted. In accordance with excitation-transfer theory, imagery was conducted at this time based on pilot data, which indicated that the residual phase, during which subjects remained physiologically activated but were unaware of this arousal, lasted from 8 to 14 minutes after exercise. These data were collected by having an independent group of 15 females perform stationary cycle ergometry at the same relative intensity as subjects in the main experiment. Afterwards their heart rate, pulse transit time (PTT), and subjective level of arousal were monitored. The residual phase was the time period during which subjects rated their physiological arousal less than 10% higher than baseline, while their pulse transit time remained significantly elevated from baseline. Pulse transit time was utilized to determine residual arousal as it has been supported as a valid index of sympathetic activation (31). and it correlates highly with systolic blood pressure (32), which has been frequently used in previous excitation transfer research (22, 25, 33-35). Subjects in both exercise and quiet rest groups engaged in both anger and sadness imagery. Two imagery scripts of each emotion were developed using questionnaire responses and ratings of scripts by two independent samples. Subjects were presented with two parallel imagery scripts of each emotion to increase generalizability; the orders of the content of imagery (anger vs. sadness) and of the two parallel imagery scripts were counterbalanced. The texts of one anger and one sadness script are presented below. Anger: After work one day you come out to your car and see a huge dent in the left front fender. Somebody has hit your car and left without leaving a note. As you look at the damage you become furious. Your heart pounds and your face gets hot. You think of how much money it will cost to get it repaired and you become even angrier. You can feel your muscles tighten. You get in your car to leave and realize that your car is too damaged to drive. You grip the steering wheel tightly and shake it in frustration. You breathe more rapidly and your anger grows as you look around for a phone.

111

R. B. FILLINGIM et al. Sadness: You are at home one day when a friend calls to tell you that someone close to you has died. You are stunned by this sad news. Your heart feels very heavy and tears start to well up in your eyes as you begin to fully realize what this means. You will never see her again and this makes you feel very blue. Your body feels numb and you slump down in your chair. Your eyes drift around the room as you think about how much you will miss her. It's hard to believe that she is gone. As you hang up the phone your breathing is slow and heavy. You feel very alone. Imagery trials were conducted as in previous work (36-39). Each imagery trial consisted of a 30minute rest period, a 50-minute read period, a 30minute image period, and a 30-minute recovery period. Physiological measures were obtained throughout the rest, image, and recovery periods and during the last 30 minutes of the read period. Following the recovery period, subjects rated how angry, sad, and aroused they felt while imagining the scene, and subjects also rated how pleasant, vivid, and easy to imagine the scene was. Ratings were made using 75-mm visual analog scales. After the imagery trials, each subject completed another POMS. Finally, subjects were either reminded of their next session or thoroughly debriefed and excused.

Apparatus and Physiological a Data Collection Silver-silver chloride electrodes were used to obtain the EKG and facial EMG. A GRASS impedance meter was used to measure impedances at facial EMG sites. GRASS P511 pre-amps were used for facial EMG data collection, and GRASS 7P4 and 7P1 pre-amps were used for EKG and peripheral pulse signals, respectively. These psychophysiological measures were amplified with a GRASS Model 7D polygraph. The three channels of EMG activity (depressor, corrugator, zygomatic) were passed through analog filters (30-300 Hz), sampled at a rate of 500 Hz, and then digitally rectified and smoothed with a low-pass, single pole Butterworth filter with a time constant of 100 milliseconds. Although this sampling rate leads to aliasing of activity in the 250- to 300Hz range (40), most of the power in facial EMG activity is below 200 Hz (28). The EKG signal was

112

filtered (5-35 Hz) and then input to a Schmitt trigger device, which signaled the computer to record the interbeat interval with a resolution of 2 milliseconds. A GRASS photoplethysmograph was utilized to obtain pulse volume. Between each Schmitt trigger the computer recorded the minimum and maximum peripheral photoplethysmograph level, and subtracting the minimum from the maximum yielded pulse volume. Additionally, pulse transit time was recorded on-line as the time interval between detection of the R-wave and the occurrence of the peak of the peripheral pulse that followed it (41). A ZENITH 248 computer equipped with a METRABYTE Dash-16 laboratory interface and a physiologic data acquisition program (VPM) (42) sampled the physiological data, performed the digital filtering, and computed the peripheral pulse values described above. A Bodyguard 990 bicycle ergometer was used for the exercise manipulation, and a REALISTIC cassette tape player and two REALISTIC Minimus 7 speakers were used to present the imagery scripts. Imagery trials were conducted in an acoustically and electrically isolated chamber, and the physiological signals were relayed to the GRASS polygraph in an adjacent control room.

Data Reduction After on-line correction, the EMG, EKG, and peripheral pulse data were further reduced prior to data analysis. The band-pass filtered, rectified, smoothed, and sampled EMG data were linearly transformed from A/D units to microvolts for data analysis. Mean heart rate was calculated for each data collection period by a computer program that converted interbeat intervals to rate per minute units (42). The mean rate is a weighted average of these individual rates such that each individual rate is assigned a weight proportional to its temporal contribution to the time period being described. Because pulse volume is a relative measure not suitable for interindividual comparisons, percent pulse volume change from baseline was used for data analyses: %PV (for period) period PV - baseline PV baseline PV This formula calculated a %PV for each period of interest relative to the individual's original baseline

Psychosomatic Medicine 54:109-120 (1992)

EXERCISE AND EMOTIONAL IMAGERY PV. Positive numbers represent an increase in volume while negative numbers reflect a decrease. Analyses of variance (ANOVAs) were conducted to determine the reliability of mean differences, and the Greenhouse-Geisser correction was imposed on repeated measures tests with 2 or more degrees of freedom.

RESULTS

Cardiovascular Effects of Exercise Means and standard deviations on cardiovascular measures during the initial baseline, the experimental manipulation period, and immediately prior to the first imagery trial (i.e., 8 months after exercise/rest) are presented in Table 1. These data are presented to demonstrate that following the experimental manipulation the exercise group was characterized by excessive arousal, which is necessary for excitation-transfer to occur. As expected, the exercise group showed much higher heart rates during exercise than the control group showed while resting. Subjects who exercised showed significantly higher heart rates and percent pulse volumes, and lower pulse transit times during the baseline period of the first imagery trial (i.e., 8 minutes after exercise) than at the initial baseline (HR: TABLE 1. Means and Standard Deviations for Exercise and No Exercise Groups before, during, and after the Experimental Period Croup Exercise

Pre-HRJ During HR Post-HR Pre-PTT Post-PTT Post-%PV

No Exercise

Mean

SD

Mean

SD

82.19 147.76 91.39 315.66 306.88 34.08

7 85 8.16 8.41 19.10 17.79 55.71

82.08 82.50 78.93 312.89 315.19 -6.79

9.06 10.27 9.20 17.91 18.10 33.67

0

HR, heart rate (in bpm); PTT, pulse transit time (in msecs), %PV, percent pulse volume.

Psychosomatic Medicine 54:109-120 (1992)

F(l,45) = 91.25, p < 0.0001; %PV: F(l,46) = 13.16, p < 0.001; PTT: F(l,45) = 12.10, p < 0.005). Conversely, the no exercise group evinced no differences in pulse transit times or percent pulse volumes, but showed lower heart rates during baseline of the first imagery trial than at the initial experimental baseline (PTT: F(l,45) = 0.72, p > 0.4; %PV: F(l,46) = 0.52, p > 0.4; HR: F(l,45) = 10.18, p < 0.005). In both experimental sessions the exercise group showed significantly higher heart rates (anger session: F(l,46) = 27.79, p < 0.0001; sadness session: F(l,46) = 13.55, p < 0.001) and percent pulse volume (anger session: F(l,46) = 6.44, p < 0.01; sadness session: F(l,46) = 6.26, p < 0.05) during baseline of the first imagery trial than the no exercise group. Thus, the exercise group remained cardiovascularly activated from exercise at the beginning of the first imagery trial relative to their own baseline values and compared with the no exercise group during the same period. Responses to Emotional Imagery Ratings. On the imagery visual analog scales, analysis of variance revealed that anger scripts were rated higher on both the anger and arousal scales, while sadness scripts were given higher ratings on the sadness scale (anger: F(l,44) = 89.73, p < 0.0001; arousal: F(l,44) = 24.39, p < 0.0001; sadness: F(l,44) = 46.01, p < 0.0001). Subjects rated the anger and sadness scripts as being equally unpleasant, vivid, and easy to imagine. No significant effects involving exercise condition, affect order, or trial emerged on the affective or imagery ratings. These data indicate that, not surprisingly, subjects perceived anger scripts as not only more anger-inducing but also more arousing than sadness scripts, and predictably, the latter were 113

R. B. FILLINGIM et al.

rated higher in sadness. However, contrary to theoretical predictions, exercise did not affect self-report response to imagery. Facial EMG Responses. The activity in the depressor, corrugator, and zygomatic muscle regions for the two experimental groups during imagery is depicted in Figure 1. Separate 2 (group) X 2 (affect order) between x 3 (period) X 2 (affect) x 2 (imagery trial) within mixed ANOVAs were performed on EMG activity in the three facial muscle regions (depressor, zygomatic, corrugator). The baseline of the imagery period was omitted from the EMG analyses because, unlike cardiovascular data, facial EMG activity is not generally subject to the Law of Initial Values and real physiological baseline is 0 for muscle tension (see Ref. 28); therefore, baseline comparisons are typically uninformative. The baseline values, however, are presented in Figure 1 for visual comparison. As shown in Figure 1, tension in zygomatic and depressor muscle regions was significantly higher during anger than sadness imagery (depressor: F(l,46) = 10.29, p < 0.005; zygomatic: F(l,46) = 8.61, p < 0.005). The affect X group interaction was significant for corrugator activity, F(l,46) = 4.09, p < 0.05. This finding, illustrated in the middle panel of Figure 1, reflects a tendency toward higher corrugator responses during sadness imagery than anger imagery for the exercise group and no differential corrugator response to anger and sadness imagery for nonexercisers. The only significant period effect was that for the depressor muscle region, F(l.5,66.7) = 5.39, p < 0.05. Follow-up analyses indicated that depressor muscle region activity was higher during the recovery period than during the read and image periods (read: F(l,46) = 8.02, p < 0.01; image: F(l,46) = 4.42, p < 0.05). In summary, tension in the depressor and zygomatic muscle regions was higher during anger than sadness imagery, and the 114

exercise group tended to show higher corrugator activity during sadness than anger imagery; however, in general facial EMG activity did not change significantly across imagery periods. No other significant main effects or interactions emerged. Cardiovascular Responses. A series of 2 (group) x 2 (affect order) between x 4 (period) X 2 (affect) X 2 (trial) within mixed ANOVAs were conducted on the three cardiovascular measures (pulse transit time, percent pulse volume, heart rate). Heart rate and percent pulse volume were found to be significantly higher during imagery for the exercise than for the no exercise group (heart rate: F(l,45) = 25.87, p < 0.0001; percent pulse volume: F(l,45) = 14.26, p < 0.0005). The group X period interaction for percent pulse volume (but not heart rate or pulse transit time) was reliable, F(2.1,96.0) = 3.41, p < 0.04, and pulse volume responses to imagery for the two groups are presented in Figure 2. In order to explicate this finding, an analysis of change scores from baseline for each imagery period was conducted, which revealed a significant Group x Period interaction, F(l.9,89.4) = 3.76, p < 0.03. Follow-up comparisons on the change scores using an appropriate pooled error term (44) revealed that, compared with the exercise group, the quiet rest group showed a significantly greater decrease in percent pulse volume from the baseline to the image period and a trend toward a greater decrease during the read period, but there were no group differences in change scores at recovery (image: F(l,69) = 4.41, p < 0.05; read: F(l,69) = 3.62, p < 0.01; recovery: F(l,69) = 0.13, p > 0.5). Thus, as can be seen in Figure 2, the exercise group showed less vasoconstriction in response to emotional imagery than the quiet rest group.

Psychosomatic Medicine 54:109-120 (1992)

R. B. FILLINGIM et al.

• — • EXERCISE O - - O NO EXERCISE

•10--

BASE

READ

IMAGE

ANGER IMAGERY

RECVR

BASE

READ

IMAGE

RECVR

SADNESS IMAGERY

Fig. 1. Facial EMG activity from zygomatic, corrugator, and depressor muscles during anger and sadness imagery for the exercise and no exercise groups.

Psychosomatic Medicine 54:109-120 (1992)

115

R. B. FIILINGIM et al.

based on facial EMG, with anger imagery characterized by greater depressor and zygomatic activity. Also, there was a tendency for subjects who exercised to show ^.« j. • higher corrugator activity during sadness T 1 1 • 1 • 1 1 than during anger imagery, while no such . T • difference emerged for the quiet rest t o -o T group. The exercise group evinced less l J \T ovasoconstriction in response to imagery 1 9' than did the quiet rest group. Finally, both BASE READ IMAGE RECVR BASE READ IMAGE RECVR groups showed decreases in tension and ANGER IMAGERY S A D N E S S IMAGERY confusion across the experimental sesFig. 2. Pulse volume responses to emotional im- sions, but the quiet rest group also showed agery for the exercise and no exercise reductions in vigor, while the exercise groups. group increased slightly. These findings are not supportive of predictions from excitation-transfer theProfile of Mood States ory that exercise-induced arousal would Change scores were computed on each enhance the experience of high arousal of the six POMS scale scores by subtract- (anger) but not low arousal (sadness) emoing the pre-score from the post score. tional imagery. There are several possible Analyses of variance indicated that only explanations for this failure to support tension and confusion decreased signifi- excitation-transfer theory. One important cantly pre to post experimental period in consideration is the potency of the experboth groups (tension: F(l,44) = 24.35, p < imental manipulation of emotion and 0.0001; confusion: F(l,44) = 11.00, p < arousal. Although subjects described the 0.005). The group X pre-post interaction imagery scenes as provoking anger and for vigor was significant, F(l,44) = 9.12, p sadness, it is unlikely that their subjective < 0.005. Inspection of the means indicated emotional experience reached the level that vigor did not change (p > 0.2) for of an authentic sadness or anger response. subjects who exercised but decreased sig- Additionally, while sadness imagery was nificantly for the quiet rest group, F(l,46) perceived as less arousing than anger im= 10.71, p < 0.002. These data suggest that agery, the cardiovascular data do not supwhile both groups decreased in tension port the notion that sadness and anger and confusion, the exercise group in- imagery were differentially arousing. Becreased slightly in vigor and quiet rest cause affective arousal appears not to group decreased on this measure. have been strongly manipulated, differential effects of exercise on sadness and anger imagery were less likely. Other possible explanations include DISCUSSION that the methodology used in the current This experiment examined the effects investigation differs substantially from of exercise versus quiet rest on self-report previous work in this area. Earlier studies and psychophysiological responses to typically employed exercise of shorter duemotional imagery. Exercise was not ration, which likely produced different found to alter self-report responses to cardiovascular and psychological reeither sadness or anger imagery. Anger sponse patterns. Thus, the physiologic and sadness imagery were distinguishable and subjective emotional status of our • — • EXERCISE O - - O N 0 EXERCISE

116

Psychosomatic Medicine 54:109-120 (1992)

EXERCISE AND EMOTIONAL IMAGERY

subjects may have differed from that of subjects in prior research. Additionally, past experiments have generally used aggressive behavior as the primary dependent measure, which may differ from physiological and subjective emotional responses. The facial EMG findings must be interpreted cautiously as facial muscle activity was generally not found to change significantly across imagery periods; therefore, facial EMG responses may have been affected by factors other than the imagery. It would be expected based on previous findings (36) that EMG activity would be higher during the image period than during read and recovery; however, this was not the case, and, in fact depressor activity was highest during recovery. This discrepancy is likely due to the fact that previous studies using this imagery procedure instructed subjects in progressive muscle relaxation prior to imagery, and subjects were instructed to use this relaxation during the recovery period (35-37). Subjects in the current study were instructed to relax but received no relaxation training; therefore, they may have been less able to reduce muscle tension than subjects in previous studies. The most interesting finding from the current study is the greater pulse volume reduction (i.e., greater vasoconstriction) in response to imagery in the quiet rest versus the exercise group. Previous research has reported that emotional imagery influenced cardiovascular responses during exercise (45). A recent study from our laboratory (13) also found associations among acute exercise, psychological stress, and peripheral vasoregulation, in that mental stress during exercise prevented the thermoregulatory vasodilatation response. Hence, mental activity may influence exercise-induced vasoregulatory responses, and peripheral vascular activity may be especially sensitive to interactions between exercise Psychosomatic Medicine 54:109-120 (1992)

and stress responses. The current finding is even more noteworthy when one considers that exercise subjects' pulse volume values were elevated with respect to their original experimental baseline and the quiet rest group showed lower values than their original baseline. Thus, considerations such as the Law of Initial Values, regression to the mean, and the natural process of recovery from exercise predict that the exercise group would show greater decrements in pulse volume. In fact the reverse occurred, suggesting that exercise protected subjects from the vasoconstrictive effects of aversive emotional stimuli. This finding differs from previous research indicating that acute exercise does not alter cardiovascular reactivity to mental stressors (12, 18): however, there are several notable differences between this study and earlier studies. We used unpleasant emotional imagery while previous studies have employed mental arithmetic problems; therefore, it is possible that exercise protects individuals from the vasoconstrictive effects of emotional stimuli but not mental tasks. Earlier studies allowed 15 to 20 minutes of recovery after exercise before introducing the stressor, while we intentionally had subjects perform imagery earlier in the post-exercise period. Thus, the cardiovascular status of exercisers during imagery in the current study differed from that of the exercisers during stress in the previous experiment. Future research examining peripheral vasoregulatory responses to cognitive and affective stressors at various stages of recovery from exercise may help clarify these issues. The mechanism whereby exercise prevented vasoconstriction is unclear. One possibility is that exercise somehow altered the subjective experience of the imagery which mediated the physiological response. However, this seems unlikely as exercise did not alter self-report ratings of imagery. Another possibility is that fol117

R. B. FILLINGIM et al.

lowing exercise there is a refractory period during which individuals are less physiologically responsive to emotional or stressful stimuli. Consistent with this hypothesis are findings of decreased adrenergic responses to stressful laboratory tasks following one bout of exercise (28) and desensitization of beta-adrenergic receptors following a single exercise session (46). Thus, decreased post-exercise adrenergic responsivity may have prevented vasoconstriction. It is tempting to hypothesize that these peripheral vascular effects of acute exercise may be clinically meaningful. For example, reduced vasoconstriction may be related to reductions in elevated blood pressure which have been associated with prolonged physical activity (47, 48) and acute exercise (11, 49). While such interpretations are premature, they merit further investigation. While the present research is consistent with previous research indicating that exercise is no more effective than quiet rest at reducing tension (8, 11), we did not replicate other findings that acute exercise reduces anxiety and depression compared to a control condition (9, 10, 12). It

is possible that engaging in unpleasant emotional imagery blocked the anxiolytic and anti-depressant effects of exercise. Also, our quiet rest group showed a significant decrease in anxiety, while control subjects in some earlier studies have not (10, 12). Thus, the difference between the current work and prior research may not rest in the effects of exercise, but in the effects of the control conditions. Somewhat consistent with a previous report that exercise enhances vigor (9), our exercise group maintained their pre-experimental level of vigor while the quiet rest group decreased. Hence, exercise may have protected subjects from the enervation experienced by the quiet rest group. In summary, the present investigation did not support predictions derived from excitation-transfer theory regarding the effects of exercise on response to sadness and anger imagery. For both affective scripts exercise appears to have blocked the vasoconstriction response. However, ratings of tension, anger, and depression were unaffected by the exercise manipulation.

REFERENCES 1. Blumenthal JA, Williams RS, Needels TL, Wallace AG: Psychological changes accompany aerobic exercise in healthy middle-aged adults. Psychosom Med 44:29-36, 1982 2. Emery CF, Blumenthal JA: The effects of exercise training on psychological functioning in healthy Type A men. Psychol Health 2:367-379, 1988 3. Greist JH: Exercise intervention with depressed outpatients. In Morgan WP, Goldston SE, (eds), Exercise and Mental Health. Washington DC, Hemisphere Publishing, 1987, 99-104 4. Moses J, Steptoe A, Matthews A, Edwards S: The effects of exercise training on mental well-being in the normal population: A controlled trial. J Psychosom Res 33:47-61,1989 5. Roth DL, Holmes DS: Influence of aerobic exercise training and relaxation training on physical and psychologic health following stressful life events. Psychosom Med 49:355-365, 1987 6. Steptoe A, Edwards S, Moses ], Matthews A: The effects of exercise training on mood and perceived coping ability in anxious adults from the general population. J Psychosom Res 33:537-547, 1989 7. Haskell WL: Physical activity and health: Need to define the required stimulus. Am J Cardiol 55:4D9D, 1985

118

Psychosomatic Medicine 54:109-120 (1992)

EXERCISE AND EMOTIONAL IMAGERY 8. Bahrke MS, Morgan WP: Anxiety reduction following exercise and meditation. Cognit Ther Res 2:323333, 1978 9. Berger BG, Owen DR: Mood alteration with swimming—Swimmers really do "feel better." Psychoshom Med 45:425-433, 1983 10. Boutcher SH, Landers SM: The effects of vigorous exercise on anxiety, heart rate, and alpha activity of runners and nonrunners. Psychophysiology 25:696-702, 1988 11. Raglin JS, Morgan WP: Influence of exercise and quiet rest on state anxiety and blood pressure. Med Sci Sports Exerc 19:456-463, 1987 12. Roth DL: Acute emotional and psychophysiological effects of aerobic exercise. Psychophysiology 29:593602, 1989 13. Roth DL, Bachtler SD, Fillingim RB: Acute emotional and cardiovascular effects of stressful mental work during aerobic exercise. Psychophysiology 27:694-701, 1990 14. DeVries HA: Tension reduction with exercise. In Morgan WP, Goldston SE (eds), Exercise and Mental Health. Washington DC, Hemisphere Publishing, 1987, 99-104 15. DeVries HA, Wiswell RA, Bulbulian R, Moritani T: Tranquilizer effect of exercise. Am J Phys Med 60:57-66, 1981 16. Overton JM, Joyner M, Tipton CM: Reductions in blood pressure after acute exercise by hypertensive rats. J Appl Physiol 64:748-752, 1988 17. Paulev PE, Jordal R, Kristenen O, Ladefoged J: Therapeutic effect of exercise on hypertension. Eur J Appl Physiol 53:180-185, 1984 18. McGowan CR, Robertson RJ, Epstein LH: The effect of bicycle ergometer exercise at varying intensities on the heart rate, EMG and mood state responses to a mental arithmetic stressor. Res Q Exercise Sport 56:131-137, 1985 19. Russell PO, Epstein LH, Erickson KT: Effects of acute exercise and cigarette smoking on autonomic and neuromuscular responses to a cognitive stressor. Psychol Rep 53:199-206.1983 20. Peronnet F, Massicotte D, Paquee JE, Brisson G, de Champlain J: Blood pressure and catecholamine responses to various challenges during exercise-recovery in man. Eur J Appl Physiol 58:551-555, 1989 21. Zillmann D: Transfer of excitation of emotional behavior. In Cacioppo JT, Petty RE, (eds), Social Psychophysiology: A Sourcebook. New York, Guilford Press, 1983, 215-246 22. Zillmann D, Katcher AH, Milavsky B: Excitation transfer physical exercise to subsequent aggressive behavior. J Exp Soc Psychol 8:247-259, 1972 23. Zillmann D, Johnson RC, Day KD: Attribution of apparent arousal and proficiency of recovery from sympathetic activation affecting excitation transfer to aggressive behavior. J Exp Soc Psychol 10:503515, 1974 24. Caprara GV, Renzi P, D'Augello D, D'Imperio G, Rielli I, Travaglia G: Interpolating physical exercise between instigation to aggress and aggression: The role of irritability and emotional susceptibility. Aggressive Behav 12:83-91, 1986 25. Cantor JR, Zillmann D, Bryant J: Enhancement of experienced sexual arousal to erotic stimuli through misattribution of unrelated residual excitation. J Pers Soc Psychol 32:69-75, 1975 26. Schwartz GE, Brown SL, Ahern AL: Facial muscle patterning and subjective experience during affective imagery: Sex differences. Psychophysiology 17:75-82, 1980 27. McNair DM, Lorr M, Droppelmann LF: Profile of Mood States. San Diego, Educational and Industrial Testing Service, 1971 28. Fridlund AJ, Cacioppo JT: Guidelines for human electromyographic research. Psychophysiology 23:567589, 1986 29. Fridlund AJ, Izard CR: Electromyographic studies of facial expressions of emotions and patterns of emotions. Social Psychophysiology: A Sourcebook. New York, Guilford Press, 243-286 30. Schwartz GE, Ahern AL, Brown SL: Lateralized facial muscle response to positive and negative emotional stimuli. Psychophysiology 16:561-571, 1979 31. Weiss T, Del Bo A, Reichek N, Engelman K: Pulse transit time in the analysis of autonomic nervous system effects on the cardiovascular system. Psychophysiology 17:202-207, 1980 32. Marie GV, Lo CR, Jones JV, Johnston DW: The relationship between arterial blood pressure and pulse transit time during dynamic and static exercise. Psychophysiology 21:521-527,1984

Psychosomatic Medicine 54:109-120 (1992)

119

R. B. FILLINGIM et al. 33. Cantor JR, Zillmann D, Day KD: Relationship between cardiorespiratory fitness and physiological responses to films. Percept Mot Skills 46:1123-1130,1978 34. Wright RA, Weeks JL, Burch D, Hernandez H: Effect of residual excitation upon appraisals of a potential aversive outcome. Manuscript submitted for publication, 1990 35. Zillmann D. Excitation transfer in communication-mediated aggressive behavior. J Exp Soc Psychol 7:491-434, 1971 36. Cook EW III, Melamed BG, Cuthbert BN, McNeil DW, Lang PJ: Emotional imagery and the differential diagnosis of anxiety. J Consult Clin Psychol 56:734-740, 1988 37. Lang PJ, Kozak MJ, Miller GA, Levin DN, McLean A Jr: Emotional imagery: Conceptual structure and pattern of somato-visceral response. Psychophysiology 17:179-192, 1980 38. Lang PJ, Levin DN, Miller GA, Kozak MJ: Fear behavior, fear imagery, and the psychophysiology of emotion: The problem of affective response integration. J Abnorm Psychol 92:276-306, 1983 39. Miller GA, Levin DN, Kozak MJ, Cook EW III, McLean AJ, Lang PJ: Individual differences in imagery and the psychophysiology of emotion. Cognit Emotion 1:367-390, 1987 40. Cook EW III, Miller GA: Digital filtering: Background and tutorial for psychophysiologists. Psychophysiology, in press 41. Obrist PA, Light KC, McCubbin JA, Hutcheson JS, Hoffer JL: Pulse transit time: Relationship to blood pressure and myocardial performance. Psychophysiology 16:292-301, 1979 42. Cook EW III, Atkinson L, Lang PJ: Stimulus control and data acquisition for IBM PC's and compatibles [Abstract]. Psychophysiology 24:726-727, 1987 43. Graham FK: Constraints on measuring heart rate and period sequentially through real and cardiac time. Psychophysiology 15:492-495, 1978 44. Winer BJ: StalisticaJ Principles in Experimental Design, 2nd Edition. New York: McGraw-Hill, 1971 45. Schwartz GE, Weinberger DA, Singer JA: Cardiovascular differentiation of happiness, sadness, anger and fear following imagery and exercise. Psychosom Med 43:343-364, 1981 46. Friedman DB, Ordway GA, Williams RS: Exercise-induced functional desensitization of canine cardiac beta-adrenergic receptors. J Appl Physiol 62:1721-1723, 1987 47. Hagberg JM, Goldring D, Ehsani AA, Heath GW, Hernandez A, Schectman K, Holloszy JO: Effects of exercise training on the blood pressure and hemodynamic features of hypertensive adolescents. Am J Cardiol 52:763-768, 1983 48. Kenney WK, Zambraski EJ: Physical activity in human hypertension: A mechanisms approach. Sports Med 1:459-473, 1984 49. Hannum SM, Kasch FW: Acute post-exercise blood pressure response of hypertensive and normotensive men. Scand J Sports Sci 3:11-15.

120

Psychosomatic Medicine 54:109-120 (1992)