- Email: [email protected]
The electrocardiogram during emotional and physical stress
International Elsevier
Journal of Psychophysiology
2 (1985) 213-219
273
PSP 00066
THE ELECTROCARDIOGRAM
THEO
H. HIJZEN
Deportment (Accepted
AND PHYSICAL
STRESS
and JEF. L. SLANGEN
of Psychophysiology. January
DURING EMOTIONAL
University of Utrecht, Sorbonnelmn
16, 3584 CA Utrecht (The Netherlands)
2nd. 1985)
Key words: emotional
stress - physical
stress - electrocardiogram
The electrocardiographic response pattern during exercise at low and high heart rate was compared with the response pattern during emotional stress. Qualitative differences between exercise and emotional stress were obtained, i.e. during exercise the ST segment was more depressed, T-wave amplitudes were larger and QT and PQ were significantly shorter than during emotional stress. The results do not support the suggestion that emotional stress evokes an exercise-like cardiovascular response pattern, which may lead to a metabolically maladaptive situation. The results are in accordance with the hypothesis that the ECG changes during emotional stress are similar to the ECG changes during right stellate stimulation, while the ECG changes during exercise are similar to the ECG changes obtained during left stellate stimulation.
INTRODUCTION It has been suggested that the control of somatic and autonomic functioning is mediated by common central nervous system structures. Such a central command theory is consistent with research demonstrating that stimulation of brain structures related to agonistic behavior and emotional display, results in an exercise-like cardiovascular response pattern (Abrahams et al., 1960; Smith et al., 1960). A central command theory also explains the intimate covariance between the vagally mediated heart rate changes and somatic activity during anticipation of an aversive event (Obrist, 1976) and the exercise-like cardiovascular response observed during emotional stress when strong sympathetic influences are evoked (Obrist et al., 1974). The theory implicates that the effects of psychological conditions on the cardiovascular Correspondence: T.H. Hijzen, ology. University of Utrecht. Utrecht. The Netherlands. 0167-X760/85/$03.30
Department of PsychophysiSorbonnelaan 16. 3584 CA
li 1985 Elsevier Science Publishers
B.V.
system (CVS) are restricted to the degree in which somatic activity is expressed or anticipated. Consequently, physical and emotional stress are supposed to have similar effects on cardiovascular, e.g. electrocardiographic (ECG), parameters. However, only a few studies directly compared the ECG obtained during physical stress with the ECG obtained during emotional stress. Stevenson et al. (1951) used a within subject design and concluded that the ECGs during physical and emotional stress were similar. No data were presented to corroborate this conclusion. Simonov et al. (1975) showed that emotional, but not physical, stress is associated with a decrease in the maximal amplitude of the ECG T-wave, while the R-T interval shortens significantly during physical stress only. The results of Simonov et al. suggest that the cardiovascular response pattern during physical stress is qualitatively different from the response pattern during emotional stress. Results, similar to those of Simonov et al. are obtained when the effects of stimulation of the left sympathetic nerves innervating the heart (LSS) are compared with the
274
effects of right stellate stimulation (RSS). LSS results in a more pronounced shortening of QT and RT than RSS (Randall. 1977; Sarnoff and Mitchell, 1962). RSS results in greater attenuation of the T-wave than LSS (Randall. 1977; Ucda et al.. 1964). Possibly, the left sympathetic cardiac nerves are dominant during physical stress, while the right cardiac nerves dominate during emotional stress. As far as LSS improves cardiac performance more effectively than RSS (Linden, 1972; Levy and Blattberg, 1977) this difference in autonomic balance seems quite functional. Based on these considerations it is hypothesized that the ECG during emotional stress is similar to the ECG found during RSS, while the ECG during physical stress is similar to the ECG pattern found during LSS. An exercise-like response during emotional stress is generally considered one of the prime factors in the etiology of cardiovascular disease, because the organism shows little somatic activity while at the same time the cardiovascular system is tuned for intense muscular effort, thus creating a metabolically maladaptive situation. (Obrist et al. 1974; Gilmore, 1974). In case qualitative differences between the ECG during physical and emotional stress are found, the relationship between emotional stress and cardiovascular disease must be reconsidered. Since ECG changes covary with heart rate, the exercise ECG and the emotional stress ECG must be compared at equal heart rates. In the normal frequency range (70-80 beats/min) parasympathetic influences on the heart are dominant. Sympathetic influences become increasingly important at higher heart rates. Therefore, in the present research the ECGs are analyzed at both low and high heart rates. Different degrees of emotional stress were induced in a field experimental situation, in which subjects were watching soccer games. The reports of Huelleman et al. (1971) and Rose and Dunn (1964) have shown that emotional stress does occur in such a situation. Some of these subjects participated in a physical exercise test. Electrocardiograms in both situations were compared at respectively high and low heart rates.
PROCEDURE Setting
und Subjects
The electrocardiograms of 58 male subjects were registered during the last 5 rounds of a European soccer competition. Three Dutch clubs participated in this competition. During the last rounds not only Dutch subjects were invited, but also subjects with the nationality of the opposing team. Altogether forty Dutch subjects participated and 18 non-Dutch subjects. During the final match subjects were seated in the stadium; the other matches were seen in the laboratory during a live television broadcasting. Each game was seen by 8 persons seated together. Spectators with different nationalities were always in different rooms. Their age ranged from 18 to 32 years, with an average of 24 years. For each match different subjects were recruited. The Dutch subjects were taken from the active supporters legion of the playing team and were therefore supposed to be highly involved in the performance of this team. Twenty-four of these subjects visited the laboratory about 3 months later and participated in a physical exercise test on a bicycle ergometer with an increasing work load from 5, 30, 60. 80 W, up to the maximum possible. Due to the limited amount of time that the exercise laboratory was at our disposal 24 persons could participate in the exercise test. These subjects were selected at random from the total sample of Dutch subjects. Three subjects initially selected were unable to come and were replaced by others. Data from one subject could not be used due to technical failure. The results of these 23 subjects are reported. All subjects were routinely examined by a qualified medical doctor and judged to be free of cardiovascular ailments or abnormal ECGs. Because subjects had to be highly involved in the performance of a team and only winning teams remained in the competition. subject selection depended upon the results of the soccer matches. Consequently, systematic variations of order of prcscntation of the soccer and excrcisc sessions was not possible. However, order effects were assumed to be absent if the exercise ECGs were a reliable replication of the normal exercise ECG reported by others.
215
Dutu collection and reduction From each subject the 3 orthogonal ECGs (X, Y, Z), according to a modified Frank lead system were recorded on analog tape. Each game was registered on video tape. From these tapes, between 50 and 60 offensive periods (ball remained in the possession of the Dutch team) were selected. The last 15 s of each offensive period was used for analysis. During each period the X, Y, Z lead was sampled at 2 ms for 15 s by a PDP 8E computer. A detailed description of the computerized ECG analysis can be found elsewhere (Simoons et al., 1975a). The absolute values of the differences between consecutive samples in the 3 leads were computed as well as the sum (SV) of these differences. This SV time function was used as an approximation of the spatial velocity of the heart vector. Results obtained via the SV function are fully comparable with results obtained by cross correlation of the spatial velocity time functions (Simoons et al., 1975a). The present approach is to be preferred because it is faster. Single representative complexes were obtained by averaging beats selected with the aid of the SV time function of the QRS complex and the ST segment. The onset and end of the QRS complex were determined using templates derived from the first reliably recorded averaged beat of a particular subject. This algorithm resulted in low measurement error and an accurate detection of the onset and end of the QRS complex, essential for reliable waveform analysis. Thus, for each selected 15 s period a single representative beat with a low noise level was obtained. In order to see whether a 15 s period could be represented by a single averaged beat, part of the data were split up into 3 periods of 5 s. On the average no significant differences appeared between these 5 s periods, but analysis at high work loads became less reliable due to noise and drift. Thus it was decided to represent each period by a single averaged beat. By the same procedure 15 s during each work load on the bicycle ergometer were represented by a single beat. From the remaining ECG complexes the following parameters were calculated: (1) the duration of: PPK-Q (Ppeak-Q), Q-TPK (Q-Tpeak), Q-TE (Q-Tend), S, STNEG (the negative part of the ST-segment); and (2) 8 time normalized amplitudes in the PQ-
interval, the QRS-complex and the ST-interval in lead X, Y, Z; i.e. l/8, 2/8, . . , 8/S PQ, QRS, ST. Each point represents the amplitude at 8 equal parts of these intervals. The 3/8 point in the PQ interval corresponds to the maximum spatial magnitude of the P-wave (PPK). The 8/8 amplitude in the ST-peak interval corresponds to the maximum spatial magnitude of the T-wave (TPK). All subjects watching soccer did reach a rate of a least 120 beats/min during exciting periods. The lowest heart rate common to all subjects was 80 beats/min. The data from each subject were used to calculate the within subjects linear regression equations with the ECG parameters (PPK-Q, QTPK, etc.) as the dependent variables and heart rate as the independent variable. From these equations, the parameter values at 80 beats/min (relatively low heart rate) and 120 beats/min (relatively high heart rate) were obtained. This procedure reduces the standard errors of the ECG parameters (Simoons et al., 1975b). Because the samples of several of the parameters were not normally distributed the non-parametric Wilcoxon matched-pairs signed-ranks test (Siegel, 1956) was used to test the significance of the differences in the ECG parameters at 80 and 120 beats/min during and between exercise and emotional excitement. Significance levels greater than or equal to 0.05 are reported. RESULTS Electrocardiographic chunges during physicul stress By comparing the exercise data at a heart rate of 120 beats/min with the data at a rate of 80 beats/min it can be shown that: the duration of PPK-Q and QT decreases while the duration of the S-wave increases significantly (Table 1); the amplitudes of the P-wave become enlarged in all leads (Fig. 1); QRS amplitudes, i.e. maximal Rwave values, do change significantly (Fig. 1); the S-wave becomes deeper with increasing exercise as shown by the amplitude-changes in X and Y (Fig. 1). The same trend is found in SX: the negative part of the ST-segment becomes more negative in X and Y because amplitudes change significantly (Fig. 1); and T-peak amplitudes in the Z-lead become more negative (Fig. 1).
216 TABLE
I
Time relutiom wlrhin the uwruged HR. heart rate: X, X-lead;
ECG.
Y. Y-lead;
“umber
of subjects
Ewruse
PPK Q Q-TPK Q-TE Duration Duration Negative Negative
S (X) S (Y) part ST (X) part ST(Y)
* Significance
SOKtT
1 (HR = X0)
2 (HR = 120)
3 (HR = X0)
4 (HR = 120)
106 256 341 25 15 11 4
95 221 29X 2x 20 22 19
106 265 351 23 13 41 20
104 250 323 23 16 42 39
lcveis according
to the Wilcoxon
teat for two “latched
---: IX
II -= -=
80BPM 120 ,,
Y
---=120
---z
-2
00 BPM 120
----:
---=
k
80BPM 120 ,,
I I II ----I
IZ
80BPt.I ,,
II
III
1111, 2
= 23.
---=120
80BPM ,,
sarnplcs.
Electrocardiogruphic
chunges during emotionul
stress
Again the data at a rate of 120 beats/min are compared with the data at a rate of 80 beats/min: Q-TPK and Q-TE duration is less at 120 beats/min (Table 1); QRS amplitudes in the X-lead diminish (Fig. 1); the S-wave deepens in Z and Y as shown by amplitude changes (Fig. 1); and T-peak values diminish significantly (Fig. 1). The exercise
ECG
compared
to the emotion
ECG
There is a significant difference in PPK-Q-duration between exercise and soccer at high heart rates (Table 1); the duration of PQ is greater during soccer; the duration of QT is greater during the soccer match at all heart rates (Table 1); S duration is never different while the negative part of ST is larger during soccer at 80 and 120 beats/min; during soccer P-wave amplitudes in the X-lead are consistently smaller (Fig. 2); the amplitudes at 3/8, 4/8, 5/8 QRS in Z are significantly smaller during soccer (Fig. 2); S and SX are less pronounced during soccer as shown by the Fig. 1. Time normalized aznplitudea in the EC6 durmg cxcrcisc and soccer at 80 and 120 beats/min. EC& were drawn from the 24 amplitudes at time normalized intervals of PPK-Q, QRS. and S-TPK. TPK. the last amplitude shown. corrcaponds to the maximum spatial magnitude of the T-wave. Levels of significnncc are shown below the corrraponding amplitude. -, P < 0.05: --, P -c 0.02; ---, P x 0.01. Exercise ECG’s on the left side (uninterrupted line). boxer ECG’a on the right aide (dashed line).
277
III1 I Ill,,
I
T};
80 BPM
Y
I >=
--_>=
I
III111
II
120 BPM
I
80 BPM
Fig. 2. Time normalized amplitudes in the ECG; exercise vs soccer at 80 and 120 beats/mm ECGs were drawn from the 24 amplitudes at time normalized intervals of PPK-Q, QRS, and S-TPK. TPK, the last amplitude shown, corresponds to the maximum spatial magnitude of the T-wave. Levels of significance are shown below the corresponding amplitude. -, P < 0.05; --, P -C 0.02: ---, P i 0.01. Exercise = uninterrupted line; soccer = dashed line.
differences in amplitudes (Fig. 2); and T-peak values are consistently lower during the soccer match. (Fig. 2).
DISCUSSION From the evidence presented in Table 1 and Fig. 2 it can be concluded that within the same
subject and at comparable heart rates some of the electrocardiographic changes during emotional stress differ significantly from the electrocardiographic changes during exercise. With increasing workload during exercise PQ and QT intervals become shorter, the ST segment is depressed, the amplitude of the P-wave increases and S becomes more negative. These results are similar to the results obtained by Simoons et al. (1975b). Because Simoons lead and computer system were used in this experiment, the similarity in results indicates that order effects of the soccer-exercise sessions were absent. During stimulation of the left sympathetic nerves to the heart (LSS), QT duration is less than during stimulation of the right sympathetic nerves (RSS) (Randall, 1977; Sarnoff and Mitchel, 1962). The present results show that QT is significantly shorter during exercise than during emotional stress. The significantly shorter PQ-interval during exercise is in agreement with the greater dromotropic effect of LSS as opposed to RSS (Randall, 1977; Wallace and Sarnoff, 1974). During LSS, the S-wave deepens (Rothberger and Winterberg, 1910), the T-wave increases and ST becomes depressed below the isoelectric line. During RSS ST elevation occurs and T decreases while a decrease in S is not reported (Randall, 1977; Ueda et al., 1964). In accordance with these observations it has been found in the present study that during exercise the S-wave is deeper and the ST segment is more depressed than during emotional stress, as shown by the significant differences in amplitudes especially during high heart rates. Finally, T-peak amplitudes are larger during exercise than during emotional stress. Although the reported ECG changes during stimulation of the cardiac nerves were obtained in animal research, body surface potential patterns between dogs and man have enough substantial similarities (Abildskov, 1972a) to conclude that, as predicted, the electrocardiographic differences between emotional stress and exercise are qualitatively similar to the electrocardiographic differences between RSS and LSS. The results imply that the central nervous system control of the cardiovascular system is less rigidly programmed for exercise responses than is often assumed. As
278
far as alternations in autonomic balance are correlated with cardiovascular pathology (Abildskov, 1972b) the present results suggest that the effect of emotional stress on autonomic balance has to be considered as a possible etiological factor in cardiovascular disease apart from the exercise-like response pattern. The exercise and the emotion ECG were analyzed at a heart rate of 80 beats/min and at a heart rate of 120 beats/min, because sympathetic and parasympathetic influences on the heart differ at these rates. Ueda et al. (1964) showed that vagal stimulation of the heart has no influence on the waveform of the ECG. The results presented in Fig. 2 suggest therefore that P, S and T wave amplitudes can be used as markers of sympathetic influences even if parasympathetic influences dominate the heart. Besides nervous system influences, the ECG is also affected by other factors like respiration, heart position, and cardiac blood volume. It is, however, unlikely that the ECG changes reported are primarily caused by differences in heart position, because Simoons et al. (1975b) showed that during an identical exercise test the spatial orientation of the maximum P, QRS and T vectors did not change significantly. Biberman et al. (1971) showed that breathing rapidly and deeply resulted in low or inverted T waves. The T wave change found during emotional stress is therefore hard to explain by respiratory differences alone. During submaximal exercise the left ventricular end-diastolic volume is reduced. This should cause a decrease in T wave because the repolarization vectors are radially oriented (Abildskov et al., 1971). Thus, the lower T waves found during emotional stress cannot be explained by changes in cardiac blood volume. One final point should be elaborated: LSS and RSS have different cardiac dynamic consequences. LSS increases cardiac output and contractile force, the development of tension in the ventricles is more rapid and the duration of ejection shorter (Sarnoff and Mitchell, 1962). The improvement in myocardial performance during sympathetic stimulation and during exercise is also achieved by a more synchronous contraction of the myocardial fibers (Randall and Priola, 1965). RSS increases heart rate relatively more, but the inotropic conse-
quences are less than during LSS (Linden, 1972). Apparently, the predominantly inotropic effects of left sympathetic dominance are quite functional during exercise, and the comparatively smaller inotropic effects of right sympathetic dominance in situations where heavy muscular effort is not performed might be functional while watching soccer. In conclusion: both at 80 beats/min and at 120 beats/min the ECG during exercise differed significantly from the ECG during emotional stress. The differences that have been observed are consistent with the idea that left-sympathetic influences on the heart are dominant during exercise, and right-sympathetic influences on the heart are dominant during emotional stress. The results obtained are not consistent with theories, in which it is assumed that exercise and emotional stress have similar effects on the heart. It is concluded that any theory about the mechanism by which psychological factors may contribute to cardiovascular disease must take into account that emotional stress is not necessarily accompanied by an exercise-like cardiovascular response pattern.
ACKNOWLEDGEMENTS This research was supported by the Netherlands Organization for the Advancement of Pure Research and conducted at the Physiological Laboratory of the University of Utrecht. Thanks are due to the members of this laboratory for their kind assistance. Especially the persistent and time consuming effort of Ed Smallenburg is gratefully acknowledged.
REFERENCES Abildskov. J.A. (1972a) Adrenergic effects on the QT interval of the electrocardiogram. Amer. Hrarl J., 92: 210-216. Abildakov. J.A. (1972b) Central nervous system influence upon electro-cardiographic waveforms. In R.C. Schlant and J.W. Hurst (Ed%), Adtmtces tn ElectrocurdtoRruph~, Prune and Stratton, New York, pp. 423-429. Abildskov, J.A.. Millar. K., Burgess, M.J. and Green. L. (1971) Characteristics of ventricular recovery as defined by the vector-cardiographic T-loop. Amer J. Car&l. 28: 670-677.
279 Abrahams, V.C.. Hilton, S.M. and Zbrozyna. A. (1960) Active muscle vasodilatation produced by stimulation of the brain stem: its significance for the defence reaction. J. Phvsiol. (Lo&), 154: 491-513. Biberman. L., Sarma. R.N. and Surawicz. B. (1971) T-wave abnormalities during hyperventilation and isoproterenol infusion. Amer. Heart J., 81: 166-174. Gilmore. J.P. (1974) Physiology of stress. In R.S. Eliot (Ed.). Stress and the Heart, Futura Publishing Company. New York, pp. 69-90. Huelleman, K.D., Mayer, H. and Stahlheber, R. (1971) Fernsehen und Herz-Kreislauf-Regulation: Kreislaufuntersuchungen bei Herzinfarkt-patienten und Normal-personen waehrend der Fernsehuebertragung van Fussballweltmeisterschaftspielen. Muench. med. Wschr., 113: 14Oll1406. Levy, M.N. and Blattberg, B. (1977) Correlation of the mechanical responses of the heart with the norepinephrine overflow during cardiac sympathetic neural stimulation in the dog. Cardiouasc. Res., 11: 481-488. Linden, R.J. (1972) Function of nerves of the heart. Curdtooasc. Res., 6: 605-626. Obrist, P.A. (1976) The cardiovascular-behavioral interaction - as it appears today. Psychophysiology. 13: 95-107. Obrist, P.A.. Lawler, J.E. and Gaebelin, C.J. (1974) A psychobiological perspective on the cardiovascular system. In L.V. DiCara (Ed.). Limhic and Autonomic Nervous Systems Research. Plenum Press, New York, pp. 311-334. Randall. W.C. (1977) Sympathetic control of the heart. In W.C. Randall (Ed.), Neural Regularton of the Hem-t. Oxford Univ. Press, New York. pp. 43-95. Randall, W.C. and Priola, D.V. (1965) Sympathetic influences on synchrony of myocardial contraction. In W.C. Randall (Ed.), Nervous Control o/the Heart, Williams and Wilkins, Baltimore. MD, pp. 214-245. Rose. K.D. and Dunn, F.L. (1964) The heart of a spectator sportsman. Med. Tms., 92: 940-951.
Rothberger, J. and Winterberg, H. (1910) Ueber die Beziehung der Herznerven zur Form des Elektrokardiogramms. Pliigers. Arch. ges. Physiol., 135: 506-557. Sarnoff. S.J. and Mitchell, J.H. (1962) The control of the function of the heart. In W.F. Hamilton and P. Dow (Eds.), Handbook of Physrology, Sect. 2: Cwcuirtron, Vol. I, Americ. Physiol. Sot. Washington, DC, pp. 489-532. Siegel, S. (1956) Nonpurametrrc Stottstrcs for the Behaororal Scrences, McGraw-Hill, New York, 75. pp. Simonov, P.V., Frolov. M.V. and Sviridov, E.P. (1975) Characteristics of the electrocardiogram under physical and emotional stress in man. Aoiut. Space environ. Med., 46: 141-144. Simoons, M.L., Boom, H.B.K. and Smallenburg, E. (1975a). On-line processing of orthogonal exercise electrocardiograms. Camp. biomed. Rex, 8: 105-117. Simoons, M.L. and Hugenholtz, P.G. (1975b) Gradual changes of ECG waveform during and after exercise in normal subjects, Circulation, 52: 570-576. Smith O.A., Jr., Rushmer. R.F. and Lasher. E.P. (1960) Similarity of cardiovascular responses to exercise and to diencephalon stimulation. Amer. J. Physiol., 198: 1130-l 142. Stevenson, I., Duncan, C.H., Ripley, H.S. (1951) Variations in the electrocardiogram changes in emotional state. Genatrits. 6: 164-178. Ueda. H., Yanai, Y., Murao. S., Harumi, K.. Mashima S.. Kuroiwa, A., Sugimoto, T. and Shiromura, K. (1964) Electrocardiographic and vectorcardiographic changes produced by electrical stimulation of the cardiac nerves. Jup. Heurt J., 5: 359-372. Wallace. A.G. and Sarnoff, S.J. (1964) Effects of cardiac sympathetic nerve stimulation on conductions in the heart. Circular. Rex. 14: 86-92.
Journal of Psychophysiology
2 (1985) 213-219
273
PSP 00066
THE ELECTROCARDIOGRAM
THEO
H. HIJZEN
Deportment (Accepted
AND PHYSICAL
STRESS
and JEF. L. SLANGEN
of Psychophysiology. January
DURING EMOTIONAL
University of Utrecht, Sorbonnelmn
16, 3584 CA Utrecht (The Netherlands)
2nd. 1985)
Key words: emotional
stress - physical
stress - electrocardiogram
The electrocardiographic response pattern during exercise at low and high heart rate was compared with the response pattern during emotional stress. Qualitative differences between exercise and emotional stress were obtained, i.e. during exercise the ST segment was more depressed, T-wave amplitudes were larger and QT and PQ were significantly shorter than during emotional stress. The results do not support the suggestion that emotional stress evokes an exercise-like cardiovascular response pattern, which may lead to a metabolically maladaptive situation. The results are in accordance with the hypothesis that the ECG changes during emotional stress are similar to the ECG changes during right stellate stimulation, while the ECG changes during exercise are similar to the ECG changes obtained during left stellate stimulation.
INTRODUCTION It has been suggested that the control of somatic and autonomic functioning is mediated by common central nervous system structures. Such a central command theory is consistent with research demonstrating that stimulation of brain structures related to agonistic behavior and emotional display, results in an exercise-like cardiovascular response pattern (Abrahams et al., 1960; Smith et al., 1960). A central command theory also explains the intimate covariance between the vagally mediated heart rate changes and somatic activity during anticipation of an aversive event (Obrist, 1976) and the exercise-like cardiovascular response observed during emotional stress when strong sympathetic influences are evoked (Obrist et al., 1974). The theory implicates that the effects of psychological conditions on the cardiovascular Correspondence: T.H. Hijzen, ology. University of Utrecht. Utrecht. The Netherlands. 0167-X760/85/$03.30
Department of PsychophysiSorbonnelaan 16. 3584 CA
li 1985 Elsevier Science Publishers
B.V.
system (CVS) are restricted to the degree in which somatic activity is expressed or anticipated. Consequently, physical and emotional stress are supposed to have similar effects on cardiovascular, e.g. electrocardiographic (ECG), parameters. However, only a few studies directly compared the ECG obtained during physical stress with the ECG obtained during emotional stress. Stevenson et al. (1951) used a within subject design and concluded that the ECGs during physical and emotional stress were similar. No data were presented to corroborate this conclusion. Simonov et al. (1975) showed that emotional, but not physical, stress is associated with a decrease in the maximal amplitude of the ECG T-wave, while the R-T interval shortens significantly during physical stress only. The results of Simonov et al. suggest that the cardiovascular response pattern during physical stress is qualitatively different from the response pattern during emotional stress. Results, similar to those of Simonov et al. are obtained when the effects of stimulation of the left sympathetic nerves innervating the heart (LSS) are compared with the
274
effects of right stellate stimulation (RSS). LSS results in a more pronounced shortening of QT and RT than RSS (Randall. 1977; Sarnoff and Mitchell, 1962). RSS results in greater attenuation of the T-wave than LSS (Randall. 1977; Ucda et al.. 1964). Possibly, the left sympathetic cardiac nerves are dominant during physical stress, while the right cardiac nerves dominate during emotional stress. As far as LSS improves cardiac performance more effectively than RSS (Linden, 1972; Levy and Blattberg, 1977) this difference in autonomic balance seems quite functional. Based on these considerations it is hypothesized that the ECG during emotional stress is similar to the ECG found during RSS, while the ECG during physical stress is similar to the ECG pattern found during LSS. An exercise-like response during emotional stress is generally considered one of the prime factors in the etiology of cardiovascular disease, because the organism shows little somatic activity while at the same time the cardiovascular system is tuned for intense muscular effort, thus creating a metabolically maladaptive situation. (Obrist et al. 1974; Gilmore, 1974). In case qualitative differences between the ECG during physical and emotional stress are found, the relationship between emotional stress and cardiovascular disease must be reconsidered. Since ECG changes covary with heart rate, the exercise ECG and the emotional stress ECG must be compared at equal heart rates. In the normal frequency range (70-80 beats/min) parasympathetic influences on the heart are dominant. Sympathetic influences become increasingly important at higher heart rates. Therefore, in the present research the ECGs are analyzed at both low and high heart rates. Different degrees of emotional stress were induced in a field experimental situation, in which subjects were watching soccer games. The reports of Huelleman et al. (1971) and Rose and Dunn (1964) have shown that emotional stress does occur in such a situation. Some of these subjects participated in a physical exercise test. Electrocardiograms in both situations were compared at respectively high and low heart rates.
PROCEDURE Setting
und Subjects
The electrocardiograms of 58 male subjects were registered during the last 5 rounds of a European soccer competition. Three Dutch clubs participated in this competition. During the last rounds not only Dutch subjects were invited, but also subjects with the nationality of the opposing team. Altogether forty Dutch subjects participated and 18 non-Dutch subjects. During the final match subjects were seated in the stadium; the other matches were seen in the laboratory during a live television broadcasting. Each game was seen by 8 persons seated together. Spectators with different nationalities were always in different rooms. Their age ranged from 18 to 32 years, with an average of 24 years. For each match different subjects were recruited. The Dutch subjects were taken from the active supporters legion of the playing team and were therefore supposed to be highly involved in the performance of this team. Twenty-four of these subjects visited the laboratory about 3 months later and participated in a physical exercise test on a bicycle ergometer with an increasing work load from 5, 30, 60. 80 W, up to the maximum possible. Due to the limited amount of time that the exercise laboratory was at our disposal 24 persons could participate in the exercise test. These subjects were selected at random from the total sample of Dutch subjects. Three subjects initially selected were unable to come and were replaced by others. Data from one subject could not be used due to technical failure. The results of these 23 subjects are reported. All subjects were routinely examined by a qualified medical doctor and judged to be free of cardiovascular ailments or abnormal ECGs. Because subjects had to be highly involved in the performance of a team and only winning teams remained in the competition. subject selection depended upon the results of the soccer matches. Consequently, systematic variations of order of prcscntation of the soccer and excrcisc sessions was not possible. However, order effects were assumed to be absent if the exercise ECGs were a reliable replication of the normal exercise ECG reported by others.
215
Dutu collection and reduction From each subject the 3 orthogonal ECGs (X, Y, Z), according to a modified Frank lead system were recorded on analog tape. Each game was registered on video tape. From these tapes, between 50 and 60 offensive periods (ball remained in the possession of the Dutch team) were selected. The last 15 s of each offensive period was used for analysis. During each period the X, Y, Z lead was sampled at 2 ms for 15 s by a PDP 8E computer. A detailed description of the computerized ECG analysis can be found elsewhere (Simoons et al., 1975a). The absolute values of the differences between consecutive samples in the 3 leads were computed as well as the sum (SV) of these differences. This SV time function was used as an approximation of the spatial velocity of the heart vector. Results obtained via the SV function are fully comparable with results obtained by cross correlation of the spatial velocity time functions (Simoons et al., 1975a). The present approach is to be preferred because it is faster. Single representative complexes were obtained by averaging beats selected with the aid of the SV time function of the QRS complex and the ST segment. The onset and end of the QRS complex were determined using templates derived from the first reliably recorded averaged beat of a particular subject. This algorithm resulted in low measurement error and an accurate detection of the onset and end of the QRS complex, essential for reliable waveform analysis. Thus, for each selected 15 s period a single representative beat with a low noise level was obtained. In order to see whether a 15 s period could be represented by a single averaged beat, part of the data were split up into 3 periods of 5 s. On the average no significant differences appeared between these 5 s periods, but analysis at high work loads became less reliable due to noise and drift. Thus it was decided to represent each period by a single averaged beat. By the same procedure 15 s during each work load on the bicycle ergometer were represented by a single beat. From the remaining ECG complexes the following parameters were calculated: (1) the duration of: PPK-Q (Ppeak-Q), Q-TPK (Q-Tpeak), Q-TE (Q-Tend), S, STNEG (the negative part of the ST-segment); and (2) 8 time normalized amplitudes in the PQ-
interval, the QRS-complex and the ST-interval in lead X, Y, Z; i.e. l/8, 2/8, . . , 8/S PQ, QRS, ST. Each point represents the amplitude at 8 equal parts of these intervals. The 3/8 point in the PQ interval corresponds to the maximum spatial magnitude of the P-wave (PPK). The 8/8 amplitude in the ST-peak interval corresponds to the maximum spatial magnitude of the T-wave (TPK). All subjects watching soccer did reach a rate of a least 120 beats/min during exciting periods. The lowest heart rate common to all subjects was 80 beats/min. The data from each subject were used to calculate the within subjects linear regression equations with the ECG parameters (PPK-Q, QTPK, etc.) as the dependent variables and heart rate as the independent variable. From these equations, the parameter values at 80 beats/min (relatively low heart rate) and 120 beats/min (relatively high heart rate) were obtained. This procedure reduces the standard errors of the ECG parameters (Simoons et al., 1975b). Because the samples of several of the parameters were not normally distributed the non-parametric Wilcoxon matched-pairs signed-ranks test (Siegel, 1956) was used to test the significance of the differences in the ECG parameters at 80 and 120 beats/min during and between exercise and emotional excitement. Significance levels greater than or equal to 0.05 are reported. RESULTS Electrocardiographic chunges during physicul stress By comparing the exercise data at a heart rate of 120 beats/min with the data at a rate of 80 beats/min it can be shown that: the duration of PPK-Q and QT decreases while the duration of the S-wave increases significantly (Table 1); the amplitudes of the P-wave become enlarged in all leads (Fig. 1); QRS amplitudes, i.e. maximal Rwave values, do change significantly (Fig. 1); the S-wave becomes deeper with increasing exercise as shown by the amplitude-changes in X and Y (Fig. 1). The same trend is found in SX: the negative part of the ST-segment becomes more negative in X and Y because amplitudes change significantly (Fig. 1); and T-peak amplitudes in the Z-lead become more negative (Fig. 1).
216 TABLE
I
Time relutiom wlrhin the uwruged HR. heart rate: X, X-lead;
ECG.
Y. Y-lead;
“umber
of subjects
Ewruse
PPK Q Q-TPK Q-TE Duration Duration Negative Negative
S (X) S (Y) part ST (X) part ST(Y)
* Significance
SOKtT
1 (HR = X0)
2 (HR = 120)
3 (HR = X0)
4 (HR = 120)
106 256 341 25 15 11 4
95 221 29X 2x 20 22 19
106 265 351 23 13 41 20
104 250 323 23 16 42 39
lcveis according
to the Wilcoxon
teat for two “latched
---: IX
II -= -=
80BPM 120 ,,
Y
---=120
---z
-2
00 BPM 120
----:
---=
k
80BPM 120 ,,
I I II ----I
IZ
80BPt.I ,,
II
III
1111, 2
= 23.
---=120
80BPM ,,
sarnplcs.
Electrocardiogruphic
chunges during emotionul
stress
Again the data at a rate of 120 beats/min are compared with the data at a rate of 80 beats/min: Q-TPK and Q-TE duration is less at 120 beats/min (Table 1); QRS amplitudes in the X-lead diminish (Fig. 1); the S-wave deepens in Z and Y as shown by amplitude changes (Fig. 1); and T-peak values diminish significantly (Fig. 1). The exercise
ECG
compared
to the emotion
ECG
There is a significant difference in PPK-Q-duration between exercise and soccer at high heart rates (Table 1); the duration of PQ is greater during soccer; the duration of QT is greater during the soccer match at all heart rates (Table 1); S duration is never different while the negative part of ST is larger during soccer at 80 and 120 beats/min; during soccer P-wave amplitudes in the X-lead are consistently smaller (Fig. 2); the amplitudes at 3/8, 4/8, 5/8 QRS in Z are significantly smaller during soccer (Fig. 2); S and SX are less pronounced during soccer as shown by the Fig. 1. Time normalized aznplitudea in the EC6 durmg cxcrcisc and soccer at 80 and 120 beats/min. EC& were drawn from the 24 amplitudes at time normalized intervals of PPK-Q, QRS. and S-TPK. TPK. the last amplitude shown. corrcaponds to the maximum spatial magnitude of the T-wave. Levels of significnncc are shown below the corrraponding amplitude. -, P < 0.05: --, P -c 0.02; ---, P x 0.01. Exercise ECG’s on the left side (uninterrupted line). boxer ECG’a on the right aide (dashed line).
277
III1 I Ill,,
I
T};
80 BPM
Y
I >=
--_>=
I
III111
II
120 BPM
I
80 BPM
Fig. 2. Time normalized amplitudes in the ECG; exercise vs soccer at 80 and 120 beats/mm ECGs were drawn from the 24 amplitudes at time normalized intervals of PPK-Q, QRS, and S-TPK. TPK, the last amplitude shown, corresponds to the maximum spatial magnitude of the T-wave. Levels of significance are shown below the corresponding amplitude. -, P < 0.05; --, P -C 0.02: ---, P i 0.01. Exercise = uninterrupted line; soccer = dashed line.
differences in amplitudes (Fig. 2); and T-peak values are consistently lower during the soccer match. (Fig. 2).
DISCUSSION From the evidence presented in Table 1 and Fig. 2 it can be concluded that within the same
subject and at comparable heart rates some of the electrocardiographic changes during emotional stress differ significantly from the electrocardiographic changes during exercise. With increasing workload during exercise PQ and QT intervals become shorter, the ST segment is depressed, the amplitude of the P-wave increases and S becomes more negative. These results are similar to the results obtained by Simoons et al. (1975b). Because Simoons lead and computer system were used in this experiment, the similarity in results indicates that order effects of the soccer-exercise sessions were absent. During stimulation of the left sympathetic nerves to the heart (LSS), QT duration is less than during stimulation of the right sympathetic nerves (RSS) (Randall, 1977; Sarnoff and Mitchel, 1962). The present results show that QT is significantly shorter during exercise than during emotional stress. The significantly shorter PQ-interval during exercise is in agreement with the greater dromotropic effect of LSS as opposed to RSS (Randall, 1977; Wallace and Sarnoff, 1974). During LSS, the S-wave deepens (Rothberger and Winterberg, 1910), the T-wave increases and ST becomes depressed below the isoelectric line. During RSS ST elevation occurs and T decreases while a decrease in S is not reported (Randall, 1977; Ueda et al., 1964). In accordance with these observations it has been found in the present study that during exercise the S-wave is deeper and the ST segment is more depressed than during emotional stress, as shown by the significant differences in amplitudes especially during high heart rates. Finally, T-peak amplitudes are larger during exercise than during emotional stress. Although the reported ECG changes during stimulation of the cardiac nerves were obtained in animal research, body surface potential patterns between dogs and man have enough substantial similarities (Abildskov, 1972a) to conclude that, as predicted, the electrocardiographic differences between emotional stress and exercise are qualitatively similar to the electrocardiographic differences between RSS and LSS. The results imply that the central nervous system control of the cardiovascular system is less rigidly programmed for exercise responses than is often assumed. As
278
far as alternations in autonomic balance are correlated with cardiovascular pathology (Abildskov, 1972b) the present results suggest that the effect of emotional stress on autonomic balance has to be considered as a possible etiological factor in cardiovascular disease apart from the exercise-like response pattern. The exercise and the emotion ECG were analyzed at a heart rate of 80 beats/min and at a heart rate of 120 beats/min, because sympathetic and parasympathetic influences on the heart differ at these rates. Ueda et al. (1964) showed that vagal stimulation of the heart has no influence on the waveform of the ECG. The results presented in Fig. 2 suggest therefore that P, S and T wave amplitudes can be used as markers of sympathetic influences even if parasympathetic influences dominate the heart. Besides nervous system influences, the ECG is also affected by other factors like respiration, heart position, and cardiac blood volume. It is, however, unlikely that the ECG changes reported are primarily caused by differences in heart position, because Simoons et al. (1975b) showed that during an identical exercise test the spatial orientation of the maximum P, QRS and T vectors did not change significantly. Biberman et al. (1971) showed that breathing rapidly and deeply resulted in low or inverted T waves. The T wave change found during emotional stress is therefore hard to explain by respiratory differences alone. During submaximal exercise the left ventricular end-diastolic volume is reduced. This should cause a decrease in T wave because the repolarization vectors are radially oriented (Abildskov et al., 1971). Thus, the lower T waves found during emotional stress cannot be explained by changes in cardiac blood volume. One final point should be elaborated: LSS and RSS have different cardiac dynamic consequences. LSS increases cardiac output and contractile force, the development of tension in the ventricles is more rapid and the duration of ejection shorter (Sarnoff and Mitchell, 1962). The improvement in myocardial performance during sympathetic stimulation and during exercise is also achieved by a more synchronous contraction of the myocardial fibers (Randall and Priola, 1965). RSS increases heart rate relatively more, but the inotropic conse-
quences are less than during LSS (Linden, 1972). Apparently, the predominantly inotropic effects of left sympathetic dominance are quite functional during exercise, and the comparatively smaller inotropic effects of right sympathetic dominance in situations where heavy muscular effort is not performed might be functional while watching soccer. In conclusion: both at 80 beats/min and at 120 beats/min the ECG during exercise differed significantly from the ECG during emotional stress. The differences that have been observed are consistent with the idea that left-sympathetic influences on the heart are dominant during exercise, and right-sympathetic influences on the heart are dominant during emotional stress. The results obtained are not consistent with theories, in which it is assumed that exercise and emotional stress have similar effects on the heart. It is concluded that any theory about the mechanism by which psychological factors may contribute to cardiovascular disease must take into account that emotional stress is not necessarily accompanied by an exercise-like cardiovascular response pattern.
ACKNOWLEDGEMENTS This research was supported by the Netherlands Organization for the Advancement of Pure Research and conducted at the Physiological Laboratory of the University of Utrecht. Thanks are due to the members of this laboratory for their kind assistance. Especially the persistent and time consuming effort of Ed Smallenburg is gratefully acknowledged.
REFERENCES Abildskov. J.A. (1972a) Adrenergic effects on the QT interval of the electrocardiogram. Amer. Hrarl J., 92: 210-216. Abildakov. J.A. (1972b) Central nervous system influence upon electro-cardiographic waveforms. In R.C. Schlant and J.W. Hurst (Ed%), Adtmtces tn ElectrocurdtoRruph~, Prune and Stratton, New York, pp. 423-429. Abildskov, J.A.. Millar. K., Burgess, M.J. and Green. L. (1971) Characteristics of ventricular recovery as defined by the vector-cardiographic T-loop. Amer J. Car&l. 28: 670-677.
279 Abrahams, V.C.. Hilton, S.M. and Zbrozyna. A. (1960) Active muscle vasodilatation produced by stimulation of the brain stem: its significance for the defence reaction. J. Phvsiol. (Lo&), 154: 491-513. Biberman. L., Sarma. R.N. and Surawicz. B. (1971) T-wave abnormalities during hyperventilation and isoproterenol infusion. Amer. Heart J., 81: 166-174. Gilmore. J.P. (1974) Physiology of stress. In R.S. Eliot (Ed.). Stress and the Heart, Futura Publishing Company. New York, pp. 69-90. Huelleman, K.D., Mayer, H. and Stahlheber, R. (1971) Fernsehen und Herz-Kreislauf-Regulation: Kreislaufuntersuchungen bei Herzinfarkt-patienten und Normal-personen waehrend der Fernsehuebertragung van Fussballweltmeisterschaftspielen. Muench. med. Wschr., 113: 14Oll1406. Levy, M.N. and Blattberg, B. (1977) Correlation of the mechanical responses of the heart with the norepinephrine overflow during cardiac sympathetic neural stimulation in the dog. Cardiouasc. Res., 11: 481-488. Linden, R.J. (1972) Function of nerves of the heart. Curdtooasc. Res., 6: 605-626. Obrist, P.A. (1976) The cardiovascular-behavioral interaction - as it appears today. Psychophysiology. 13: 95-107. Obrist, P.A.. Lawler, J.E. and Gaebelin, C.J. (1974) A psychobiological perspective on the cardiovascular system. In L.V. DiCara (Ed.). Limhic and Autonomic Nervous Systems Research. Plenum Press, New York, pp. 311-334. Randall. W.C. (1977) Sympathetic control of the heart. In W.C. Randall (Ed.), Neural Regularton of the Hem-t. Oxford Univ. Press, New York. pp. 43-95. Randall, W.C. and Priola, D.V. (1965) Sympathetic influences on synchrony of myocardial contraction. In W.C. Randall (Ed.), Nervous Control o/the Heart, Williams and Wilkins, Baltimore. MD, pp. 214-245. Rose. K.D. and Dunn, F.L. (1964) The heart of a spectator sportsman. Med. Tms., 92: 940-951.
Rothberger, J. and Winterberg, H. (1910) Ueber die Beziehung der Herznerven zur Form des Elektrokardiogramms. Pliigers. Arch. ges. Physiol., 135: 506-557. Sarnoff. S.J. and Mitchell, J.H. (1962) The control of the function of the heart. In W.F. Hamilton and P. Dow (Eds.), Handbook of Physrology, Sect. 2: Cwcuirtron, Vol. I, Americ. Physiol. Sot. Washington, DC, pp. 489-532. Siegel, S. (1956) Nonpurametrrc Stottstrcs for the Behaororal Scrences, McGraw-Hill, New York, 75. pp. Simonov, P.V., Frolov. M.V. and Sviridov, E.P. (1975) Characteristics of the electrocardiogram under physical and emotional stress in man. Aoiut. Space environ. Med., 46: 141-144. Simoons, M.L., Boom, H.B.K. and Smallenburg, E. (1975a). On-line processing of orthogonal exercise electrocardiograms. Camp. biomed. Rex, 8: 105-117. Simoons, M.L. and Hugenholtz, P.G. (1975b) Gradual changes of ECG waveform during and after exercise in normal subjects, Circulation, 52: 570-576. Smith O.A., Jr., Rushmer. R.F. and Lasher. E.P. (1960) Similarity of cardiovascular responses to exercise and to diencephalon stimulation. Amer. J. Physiol., 198: 1130-l 142. Stevenson, I., Duncan, C.H., Ripley, H.S. (1951) Variations in the electrocardiogram changes in emotional state. Genatrits. 6: 164-178. Ueda. H., Yanai, Y., Murao. S., Harumi, K.. Mashima S.. Kuroiwa, A., Sugimoto, T. and Shiromura, K. (1964) Electrocardiographic and vectorcardiographic changes produced by electrical stimulation of the cardiac nerves. Jup. Heurt J., 5: 359-372. Wallace. A.G. and Sarnoff, S.J. (1964) Effects of cardiac sympathetic nerve stimulation on conductions in the heart. Circular. Rex. 14: 86-92.