THE DEVELOPMENT OF PERIODICITY OF HEART RATE AND RESPIRATION DURING SLEEP IN NEWBORN BABIES W. Baust and /. Gagel University of Diisseldorf, Neurological Clinic and Pediatric Clinic, Diisseldorf/Germany

Sleep

development of periodicity

newborn infants

Introduction Since the fundamental work of Aserinsky and Kleitman (1955) it is well established that cyclic phenomea occur during sleep of infants. Kleitman (1963) described a cycle of between 50 and 60 minutes duration which he called the basic rest activity cycle. This has been confirmed by subsequent investigations {Monod et Pajot 1965, Stern et al. 1969). A cycle duration of 60 to 90 minutes was found by Goldie and van Velzer (1965). Although all investigators looked for a cycle duration similar to the one defined by Aserinsky and Kleitman, their results had in common the finding of a large interindividual variability. Sterman and Hoppenbrouwers (1971) made the interesting observation that fetal mobility, recorded through the abdominal Received: April 12, 1977

heart rate

respiratory rate

surface of pregnant women, shows two markedly different cycles of mobility. The authors used power spectral analysis for the evaluation of their data, a technique, previously introduced by Globus (1970) in order to determine sleep cycles in adults. The purpose of the present study is to investigate periodic phenomena of heart rate and respiration during sleep in newborn infants. Bearing in mind the large variability reported in previous investigations, a homogeneous sample was used under homogeneous conditions, assuming that these factors might be one of the reasons for the large interindividual variability. Respiration and heart rate were selected because these two parameters permit the most precise recording and therefore allow a very precise study of possible

Accepted: July 20, 1977

Address: W. B., University of Diisseldorf, Neurological Clinic, Moorenstr. 5, D-4000 Diisseldorf

387

Downloaded by: NYU. Copyrighted material.

Baust, W. and Gagel, ].: The development of periodicity of heart rate and respiration during sleep in newborn babies. Neuropadiatrie 8: 387—396 (1977). Poly graphic records ofEEG, ocular movements, respiration, and heart rate were made continuously in 10 full term babies between 2 and 7 days old. Power spectral analysis was then computed for heart and respiratory rate. The period durations obtained by this method ranged from 20 to 70 minutes. Most of the power spectra showed two maxima. In many cases cycles of respiration and heart rate were not identical. The experiments show that newborns have two separate cycles with different period durations. Most probably the shorter cycle originates from the fetal life, the longer one represents the development of the periodicity of the adult.

cyclic phenomena by means of mathematical methods such as autocorrelation analysis. Methods Experiments were carried out on 10 normal full-term newborn infants aged between 2 and 7 days, who had been bottle fed from the first day on. During the experimental sessions which lasted from 2 pm to 5 pm the infants were kept in an airconditioned environment (30" C, 60°/o humidity) with a low and constant noise level. The following physiological data were recorded: 1. Respiratory rate by means of a thermistor adjusted in front of the nostril, 2. heart rate by electrocardiography, 3. two bipolar leads of the EEG, 4. ocular movements, 5. electromyogram. The data were amplified through an ordinary polygraph and stored on FMtape for further data processing. After the electrodes were fixed and the recording equipment was adjusted the infant was fed. As soon as the onset of sleep occurred, whichwas monitored by the EEG and by direct observation of the infant, the recording was started. Subsequently heart and respiratory rates were computed continuously during the entire observation period of 3 hours by sequential interval histograms using a BIOMAC 1000 analyser. For this procedure the R-R interval in the ECG was selected for calculation of heart rate, respiratory intervals were measured from the ascending part of

Vol. 8, No. 4, 1977

1

Neumpidiatrie

inspiration to the corresponding point of the following respiratory cycle. Further data processing was carried out on a digital computer. Mean heart period duration and mean respiratory period duration were computed for each of consecutive 30 seconds intervals over the entire recording period, which lasted approximately 3 hours for each infant. As a measure of variability the variation coefficient was calculated again for each 30 second interval. These data were fed to the UCLA computer program BMD 02 T for autocorrelation and power spectral analysis. Results Visual examination of the original data does not allow to discover a possible periodicity. However, the plot of the calculated mean period durations for the 30 second intervals shows a periodic change in heart and respiratory rate, which is more or less pronounced due to interindividual differences or to random fluctuations. The lower part of Fig. 1 shows the course of the RR intervals of baby I1 over a period of 3 hours. Each cross represents the mean heart ~ e r i o dduration of one 30 second interval. Sleep stages are indicated by the horizontal bars at the top. I t becomes evident that periodicity is roughly correlated to sleep stages: RR intervals during active REM - sleep tend to be shorter. I t should be pointed out that periodicity of the heart rate is not as pronounced in the rest of the infants as it is in the case shown in the lower part of Fig. 1. The same observation was made concerning respiration; figu-

Downloaded by: NYU. Copyrighted material.

Baust and Gagel

Periodicity of heart rate in newborns

Downloaded by: NYU. Copyrighted material.

POWER

"

0

.

"

30

'

~

"

60

'

l

'

"

90

'

~

l

120

"

r

~

~

l

~

~

~

~

~

l

~

~

150 180 MINUTES

Fig. I Heart rate of baby 11. Upper part: Power spectrum with one peak only. From the two diaect adjacent maxima the highest one (0.340 X lopS) corresponds t o a cycle duration of 49 minutes, the slightly smaller one to 65 minutes. Lower part: Plot of the original data. Each point represents the mean heart period duration of one 30 second interval. The horizontal bars indicate the stages of active sleep. Besides the cyclic changes in heart rate the intervals become longer at the end of the 3 hour recording period

~

a

~

Baust and Gagel

Vol. 8, No. 4, 1977

1

NeuropZdiatrie

CPS .I04

0

30

60

90

120

150

180 MINUTES

Fig. 2 Heart rate of baby VIII. Upper part: Power spectrum with less pronounced periodicity, although a maximum can be estimated at 0.170 X 10- cps = 98 minutes. Lower part: Plot of the original data, same record as in Figure I. The cycle duration of 98 minutes calculated by means of power spectral analysis can clearly be detected also in the original plot

Downloaded by: NYU. Copyrighted material.

POWER

Periodicity of heart rate in newborns

Downloaded by: NYU. Copyrighted material.

POWER

Fig. 3 Respiratory rate of baby VIII. Upper part: Power spectrum of respiratory rate with two adjacent peaks at 49 and 65 minutes cycle duration. Note that this maximum in the power spectrum is not identical with that for heart rate shown in Figure 2. Lower part: Plot of the original data of respiratory intervals. Compared to the original data of heart rate in the same baby (Figure 2) it can clearly be seen that respiratory periodicity is faster

Baust and Gage1

Vol. 8, No. 4, 1917

1 Ncuropidiatrie

.lo" 29.6

5.6

F

Downloaded by: NYU. Copyrighted material.

POWER

SEC

~ . ' . ~ " " . ~ ~ ' ~ ~ . " ' " ' ~ " ~ ' . ' " ' " ' . '

0

30

60

90

120

150 180 NINUTES

Fig. 4 Respiratory rate of baby IX. U p p e r part: Power spectrum with two separate peaks, the first one ranges from 65 to 98 minutes, the second one from 25 to 28 minutes. Lower part: Plot of the original data. The slow cycle of about 90 minutes can clearly be seen in the first half of the record, whereas a t the end of the 3 hours record the superimposed shorter cycle of about 25 minutes becomes dominant

Periodicity of heart rate in newborns

Table I Cycle durations (minutes) of heart and respiratory rate as estimated by the power spectra. If the peak in the spectrum was not markedly larger than its neighbors, the range is given with the highest peak underlined. Most of the power spectra showed two maxima. The largest one was defined as 1st maximum Respiratory rate

1st Maximum

2nd Maximum

1 I1 I11 IV V VI VII VIII IX

39 49 - 65 33 - 2 20-2-28 65 -98, 49 no values, data error 98 o r longer 49 28 -13

33 -65-98

X

--

1st Maximum

65 - 9 2 33 -2 -40 - 65

98 or longer

98 o r longer 22 65 39 98 o r longer

39-65-98 16

49-65

res 2 and 3 show breath and heart beat durations of baby VIII, demonstrating a case with less marked periodicity, although a cyclic variation can be detected by visual inspection of the plots. From the lots of the original data a trend of decreasing heart and respiratory rates is seen, which was similarly found in 9 out of the 10 infants investigated in this study. A more precise evaluation of periodicity, particularly in cases where it is obscured by random events, can be performed with mathematical methods such as power spectral analysis. After detrending of the time series two power spectra were computed for each of the infants: heart period duration and respiratory period duration. Only for baby VII power spectra for heart rate could not be calculated due to faulty recording. The results of the spectral analysis indicate that heart rate as well as respiratory rate and their variability show

2nd Maximum

a periodic change over the three hours of recording time. As shown in Table I the cycle duration varies considerably. Aside from a few cases without detectable peaks of power, almost all data show two separate maxima in their powerspectrum. The first maximum was defined as the point which indicates considerably more power than the neighbors. If the maximum consisted of two or three directly adjacent points which did not markedly differ in their amplitude, Table I shows this range where the highest point is underlined. The definition for the second maximum was the same except that this had to be lower than the first one. The upper part of Fig. 1 illustrates an example for the absence of a second maximum (baby I1 heart rate). In this case the first maximum consists of two points with almost identical power. The slightly higher value corresponds to a cycle of 49 minutes, the lower one to a cycle of 65 minutes. Thus the under-

Downloaded by: NYU. Copyrighted material.

Heart rate Baby

lined value of 49 minutes in Table I does not imply that this is the true cycle for heart rate in baby 11. It simply indicates that this is the highest peak; the cycle duration could also be somewhere between 49 and 65 minutes. Fig. 4 gives an example for two separate peaks (baby IX, respiratory rate). As seen in Table I the cycle durations in this particular case range from 65 to 98 minutes and from 25 to 28 minutes. A general statement can be made from Table I: with a few exceptions two peaks of power were found for each of the physiological parameters. In the case of heart rate, cycle duration ranged from 16 to 98 minutes and longer, for respiration from 20 to 98 and longer. Cycles of longer than 98 minutes duration could not be verified due to the method of analysis employed. Since the recording time was 180 minutes, the estimation of cycle durations longer than 98 minutes was not possible. I t should be noted that cycle durations for heart rate and respiration were not identical in all cases. This finding can be particularly well demonstrated in baby VIII (Figs. 2 and 3). The maximum of power for heart rate was found to be 98 minutes, whereas the corresponding values for respiration ranged from 49 to 65 minutes. Bearing these results in mind the plots of the original data were visually scored under this particular aspect (see Figs. 2 and 3). On the basis of periodicity found in the power spectra it was possible to discover identical cyclic

Voi. 8, N o . 1, 1977

1

Neuropiidiatric

changes in the two time series, i. e. cyclic variations of heart and respiratory rates do not go parallel in the case of baby VIII.

Discussion The conclusions to be drawn from the results reported in this study are limited due to the methodological restrictions when using autocorrelation. The strongest objection which could be raised is the relatively short observation period. This limitation is given by the spontaneous awakening of the baby due to the 3-4 hour schedule. Thus heart and respiratory rates during uninterrupted sleep could only be recorded between two periods of wakefulness. The recording ~ e r i o dof three hours was selected for reasons of comparability only. A time series of this duration of course does not permit selection of low frequency rhythms with sufficient precision. A second problem concerns the definition of a maximum in the power spectrum. There is no statistical procedure available to estimate the significance of differences of adjacent maxima. According to Mercer (1960) there is evidence of periodicity if one of the points indicates considerably more power than its neighbors. This automatically includes a subjective error introduced by the person who does the scoring. The method of power spectral analysis seems nevertheless to be a useful design to distinguish periodic phenomena from random fluctuations in the data obtained by the present study. The general conclusions to be drawn from our results are:

Downloaded by: NYU. Copyrighted material.

Baust and Gagel

Periodicity of heart rate in newborns

If we had done so with the values shown in Table I, the mean value of our data would well fit into the range of 50-60 minutes of Kleitman's basic rest activity cycle. 3. In most of the power spectra two maxima could be found. This points to the existance of two different cycles of different length during the sleep of human newborns. In a study of Sterman and Hoppenbrouwers (1971) the fetal motility was recorded through the abdominal surface of pregnant women. Using power spectral analysis, two separate cycles of motility were found, one with a mean of 39.6 11.8 minutes and a second one with a mean of 96.4rf: 13.2 minutes. The authors discuss the possibility that the slow cycle corresponds to the adult rest-activity cycle, whereas the fast cycle would be identical to the cycle at term reported by Stern et al. (1969). Since our results generally agree with the findings of Sterman and Hoppenbrouwers one might conclude that the two fetal cycles continue during the newborn period. 4. Although visual inspection of the data in Table I suggests that periodicity of heart rate parallels that of respiratory rate, this is obviously not true in all cases. An explanation of this phenomenon must remain speculative. Although various physiological parameters show their own cyclic pattern, the latter two are obviously not yet under the control of a central rhythmic pacemaker. Those babies in which the periodicity of heart rate and

*

Downloaded by: NYU. Copyrighted material.

I. The power maxima for heart and respiratory rate are not within the range of 50 to 60 minutes. 2. The cycle durations show a rather large variability. A large variabili~t~ of cycle durations has been described by all previous investigators. Aserinsky and Kleitman (1955) found a motility cycle somewhere in the range of 37-96 minutes. A rhythm of activity during the sleep of newborns with durations between 60 and 90 minutes was described by Goldie and van Velzer (1965). Stern et al. (1969) found a mean length of sleep cycles in newborn infants of 47 minutes with a standard deviation of 10.4, which means that about 96O/o of their values must be within the range from 27 to 67 minutes. The largest range of variability was seen byMonod and Pajot (1965). Although their cycle duration varied between 23 and 136 minutes, the authors finally reach the conclusion that the usual duration of the sleep cycle is 50 to 60 minutes. Harper et al. (1974) found that the periodicity is very variable at one week ranging from 0.5 to 4 cycles per hour. In a recent review Parmelee (1974) has reported average cycle durations of 47 minutes at term with no data about variability. The previous findings together with our own results clearly demonstrate that cycle durations during sleep of newborn infants show a very large interindividual variability which, for our oppinion does not allow calculation of arithmetic means.

respiration is dissociated are perhaps less developed than others, although clinical examination did not show signs of unmaturity. This explanation is supported by the results of Monod et al. (1967) who found that babies with de~elopm~ental disturbances showed abnormal respiratory and cardiac rhythms not correlated with the sleep cycle. References Aserinsky, E. and Kleitman, N.: A motility cycle in sleeping infants as manifested by ocular and gross bodily activity. J. Appl. Physiol. 8: 11-18 (1955). Globus, G.: Quantification of the REM sleep cycle as a rhythm. Psychophysiology 7: 248-253 (1970). Goldie, L. and van Velzer, C.: Innate sleep rhythms. Brain 88: 1043-1056 (1965). Harper, R. M., Sclabassi, R. J. and Estrin, T.: Time series and sleep research. IEEE Trans. Automatic Control 19: 932-943 (1974). Kleitman, N.: Sleep and Wakefulness. (University of Chicago Press, Chicago and London 1963).

Vol. 8, No. 1, 1977

1

Newopidiatrie

6. Mercer, D.M.A.: Analytical methods for the stddy of periodic dhenomena obscured by random fluctuations. Cold Spring Harbor Symp. Quant. Biol. 25: 73-86 (1960). Monod, N. et Pajot, N.: Le sommeil du nouveau-nC du prCmaturC. I. Analyse des Ctudes polygraphiques (mouvements oculaires, respiration et EEG) chez le nouveaunC a terme. Biol. Neonat. 8: 281-307 (1965). Monod, N., Eliet-Flescher, J. et DreyfusBrisac, C.: Le sommeil du nouveau-n6 et du prkmatur& 111. Les troubles de l'organization du sommeil chlez le nouveau-nC pathologiques. Analyse des Ctudes polygraphiques. Biol. Neonat. 11: 216-247 (1967). Parmelee, A. H. Jr.: Ontogeny of sleep patterns and associated periodicities in infants. Pre- and postnatal development of the human brain. Mod. Probl. Paediatr. 13: 298-311 (1974). . . 10. Sterman, M. B. and Hoppenbrouwers, T.: The development of sleep - waking and rest - activity, patterns from fsetus to adult in man. In: Brain Development and Behavior. Sterman, M. B., McGinty, D. J. and Adinolfi, A. M. (Eds) pp 203-227. Academic Press, New York 1971. Stern, E., Parmelee, A. H., Akiyama, Y., Schultz, M. A. and Wenner, W. H.: Sleep cycles characteristics in infants. Pediatrics 43: 65-70 (1969). L

Downloaded by: NYU. Copyrighted material.

Baust and Gagel

The development of periodicity of heart rate and respiration during sleep in newborn babies.

THE DEVELOPMENT OF PERIODICITY OF HEART RATE AND RESPIRATION DURING SLEEP IN NEWBORN BABIES W. Baust and /. Gagel University of Diisseldorf, Neurologi...
341KB Sizes 0 Downloads 0 Views