Exp. Pathol. 1991; 42: 251-256 Gustav Fischer Verlag lena
Friedrich Schiller University lena, 1) Institute of Pathological Physiology, 2) Children's Hospital, lena, F.R.G.
Parameters for sensitive cerebral monitoring during the neonatal period in intensive care medicine By M. EISELT1), M. ROTHER 1), R. BOLWIN2) and U. ZWIENER1)
Address for correspondence: Dr. med. M. EISELT, Friedrich Schiller University, Institute of Pathological Physiology, LobderstraBe 3, D-O-6900 lena, F.R.G. Key words: EEG-monitoring; neonatal EEG; amplitude; EEG-patterns; righUleft hemisphere
Summary In order to bring about further reduction of neonatal mortality and morbidity, immediate and highly sensitive detection of brain-threatening situations is necessary. We investigated 52 newborns by means of 8-channel EEG recorded with normal (15 mm/ s) and compressed write-out (1 cm/min). A correct estimation of amplitude of background activity was possible by both methods. For differentiation between healthy newborns and newborns at risk, the right hemisphere amplitude of the interburst period was the most appropriate. The amplitude of the continuous EEG and the interburst period was reduced in the group of newborns at risk. In the group of the most severely disturbed newborns there was an amplitude difference between both hemispheres with greater reduction over the right hemisphere. These results emphasize the importance of the amplitude of background activity for prognostication and the necessity of using 2 interhemispheric derivations for an independent estimation of changes in both hemispheres.
Introduction During the last few years neonatal EEG has been widely accepted as a reliable parameter for describing the actual cerebral function. Until now, the visual analysis of the neonatal EEG has been of great importance. The occurrence of special patterns and the pattern-related amplitude of the EEG plays the major role (2, 11). Recent results emphasize the importance of the amplitude of background activity, especially of the interburst period for the estimation of the severity of disturbances of cortical functions (2). The prognosis further correlates well with the duration of the occurrence of these disturbances as well as the aetiology. Because EEG patterns and their amplitudes have a great prognostic importance, a method for longterm monitoring of cerebral functions recording both parameters is necessary. Thus, the aim of this study was, firstly, to compare 8 channel EEG recorded with normal write-out (15 mm/s) with that recorded with compressed write-out (I cm/min) concerning the measured amplitude of background activity, especially of the interburst period, and, secondly, to find EEG parameters which made it possible to distinguish between healthy newborns and newborns at risk. 17*
Exp. Pathol. 42 (1991) 4
251
Method We recorded the 8-channel EEG of 52 newborns between the 1st and 7th day of life (12 newborns without risk factors and 40 newborns at risk). Follow-up examinations were made until the age of 1Y2 years (table 1). EEG was recorded in a bipolar manner: Fpl-C3, C3-0 1, Fpl-T3, T3-
Table 1. Characterization of the investigated newborns (CA = conceptional age at the time of EEG monitoring examination) newborns not at risk: 12 newborns, CA = 34.9 ± 4.1 weeks normal cardiotocogram postnatal pH > = 7.2 Apgar-Score at 5 min> 6 normotrophy no artificial ventilation during the neonatal period results of follow-up examination at 1 and 112 year: - without any abnormality = 12 newborns (healthy) newborns at risk: 40 newborns, CA = 32.9±4.6 weeks pathological cardiotocogram postnatal pH ~ 7.2 Apgar-Score at 5 min> 7 hypotrophy artificial ventilation during the neonatal period results of follow-up examination at 1 and 11/2 year: - without any abnormality = 14 newborns (healthy) - with transient or slide neurological abnormalities = 6 newborns - infantile cerebral palsy = 7 newborns - died during the first 18 month of life: without pathological finding of the brain by autopsy = 6 newborns with pathological finding of the brain by autopsy = 7 newborns
01, Fp2-C4, C4-02, Fp2-T4, T4-02 [in accordance with the International 10- 20 system modified by HELLSTROM (6)]. Impedance was between 5 to 10kQ. The EEG was recorded with normal (15 mm/s) and compressed write-out (1 cm/min). The comparison was made concerning the amplitude and special EEG patterns (table 2). For statistical analysis Spearman's rank correlation coefficient, paired t-test and multidimensional analysis of variance and discriminance analysis (4) were used.
Results Estimation of the amplitude of background activity The estimation of the amplitude of the EEG background activity recorded with compressed write-out (1 cm/min) was as good as using the 8-channel EEG recorded with normal write-out (15 mm/s) [interbrust period of the discontinuous EEG (r = 0.83)]. During the occurrence of continuous as well as the burst period of discontinuous EEG there was also a significant correlation 252
Exp. Pathol. 42 (1991) 4
Table 2. Estimated parameters in the EEG EEG recorded with compressed write-out amplitude was measured in ten I-min-intervals continuous EEG pattern: - 95 % of all waves were within the limits discontinuous EEG: - this EEG was divided into 2 pattern components Burst-period Interburst-period - 95 % of all waves were within the limits the frequency of the occurrence of Burst-Interburst cycles/lO minutes EEG recorded with normal write out the amplitude during the Interburst-period
o ==
normal amplitude of background activity 1 == amplitude is within the limits > 5 f! V < 20 f! V 2 == amplitude is < 5 f! V 3 == isoelectric EEG
between the amplitude measured in the EEG-monitoring (I cm/min) and the amplitude measured in the EEG recorded with normal write-out (r == 0.68-0.73, significant alpha < 0.05).
Pattern related changes The EEG amplitude of the interburst period measured over the right hemisphere was the best of all parameters concerning a differentiation of healthy newborns and newborns with neurological abnormalities in the follow-up at I Y2 years or newborns who died (the variable with the greatest contribution to the discriminant score, T == 0.152). The next efficient parameter was the amplitude ofthe burst period measured overthe right hemisphere (T == 0.099) and the third the age-correlated 1eo _ a~ m p_ llt_ u_ d. ~(p ~V_)__________________________~ 180
140 " 120 100
eo eo 40
20
o _
,
t hUt
.p"',. _
left h
l.pIt.,.
Fig.l. Mean amplitude of the continuous EEG measured over the right and the left hemisphere in the different follow-up groups (Abbreviations used in fig. I to 3: healthy == newborns without any evidence of psychoneurological abnormalities at the follow-up at I and 3 Yz years; rep == newborns who suffer from infantile cerebral palsy; died withouUwith == newborns who died within this follow-up period withouUwith evidence of brain damage observable by autopsy). burst-interburst-duration (T == 0.009). If we use only the amplitude of the interburst and the Exp. Pathol. 42 (1991) 4
253
400 am plllude - --'(tJ_V_, _ _ _ _ _ _ _ _ _ _ _ __
_
I
I
h.m phere
_
lefl
Fig. 2. Mean amplitude of the burst period of the discontinuous EEG measured over the right and the left hemisphere in the different follow-up groups. The amplitude measured over the right hemisphere is significantly reduced (alpha < 0.05) in newborns who died without and with evidence of brain damage observable by autopsy.
1·0
amplitud (V)
120 '
.pilel'
Ie I hem
.ph.,.
Fig. 3. Mean amplitude of the interburst period of the discontinuous EEG measured over the right and the left hemisphere in the different follow-up groups . The amplitude measured over the right hemisphere is significantly reduced (alpha < 0.01) in newborns who died with evidence of brain damage observable by autopsy.
amplitude of the burst period , the following discriminant function between both groups can be concluded: D = (-1.4
* amp!.
of interburst)
* (1.1 * amp!.
of burst)
EEG-amplitude over the right and left hemisphere
In healthy newborns at the age of I Y2 years (without any pathological findings in the psychomotoric examination) in all patterns of pattern components there are no statistically significant differences between the hemispheres (figs. 1, 2, 3).In newborns who developed infantile cerebral palsy or died within the first I V2 years (with or without evidence of brain damage after autopsy) there was a decreased amplitude measured in the continuous EEG as well as the interburst period of the discontinuous EEG. During the interqurst period the amplitude was lower over the right hemisphere (alpha < 0.0 I) in newborn who died with evidence of brain damage
254
Exp. Pathol. 42 (1991) 4
observed by autopsy. During the burst period the amplitude was lower over the right hemisphere in newborns who died with and without evidence of brain damage observed by autopsy. In contrast to the amplitude of the continuous EEG and the amplitude of the interburst period, the amplitude of bursts was increased in newborns who developed infantile cerebral palsy and was reduced in newborns who died with brain damage.
Discussion The EEG amplitude representation in EEG-monitoring provided greater amounts in comparison to the EEG recorded with normal write-out (15 mm/s). This may result from the fact that very slow waves « 0.5 Hz), not regarded in normal write-out, are more visible in compressed write-out, and thus may contribute to the increase in the amplitude . Nevertheless, changes of the amplitude of background activity measured in the normal write-out (15 mm/s) are also measurable in the compressed write-out. Thus, this EEG-monitoring method provides the same information concerning the development of the amplitude of background activity as the normal EEG-recording and, in addition, makes it possible to estimate amplitude changes of low frequency range. As shown by our results, a pattern-related estimation of the background activity is necessary. The discriminant function can be interpreted as follows: With a decreasing amplitude of the interburst period the probability of a bad prognosis increases, but, in contrast, an increasing amplitude ofthe burst period increases the probability of a bad prognosis, for instance cerebral palsy. Thus, the amplitude of burst and interburst has an inverse direction of changes with increasing severity of brain damage. This increase of burst amplitude seems to disappear if the severity of brain damage further increase. Until now, the brain structure in human neonates generating the burst of discontinuous EEG has been unknown (7). In animal experiments it can be shown that burst originates in subcortical structures and causes, by means of subcortico-cortical pathways, the discontinuous EEG (5). Mild to moderate impairment of the function of the brain leads to a more synchronized discharge of neuronal populations (greater vulnerability of inhibitory neurons (10)). Their expression in the EEG is an increase in amplitude. Further increase of functional impairment leads to a reduction of the magnitude and number of postsynaptic potentials and correlates with a decrease of amplitude (6). This correlates with a very bad prognosis (death of the newborn) in our investigation. In healthy newborns there is no significant difference of the amplitude of background activity between the 2 hemispheres in all EEG patterns and pattern-related components. This result agrees with the findings of PETERS et al. (8). In newborns who died, a reduced amplitude, especially over the right hemisphere, was observed during the discontinuous EEG. This result was confirmed by data measured by pattern-related automatic EEG-anal ysis (9). It reflects a greater degree of functional disturbance of the right hemisphere . The reduction of the amplitude over the left hemisphere in dependence on the severity of disturbances is not quite obvious. This can be explained by a different degree of amplitude reduction in different frequenllY ranges, e.g. high frequencies disappear earlier than los frequencies. With the EEG-monitoring method it is impossible to distinguish between the different contributions of each frequency range to the whole amplitude. Thus, it may be possible that low frequency persists over the left hemisphere with only a small reduction in amplitude, whereas higher frequencies have already disappeared. As shown by investigations of Aso et al. (1) the sensitivity of pathological EEG asymmetry is low (40%) for the focality of morphological damage , although the specifity is relatively high (85 %). This underlines the necessity of using 2 interhemispheric derivations for an independent estimation of changes in both hemispheres . For prognostication a pattern-related and localization-related estimation of the amplitude of EEG background activity are of great importance.
Exp. Pathol. 42 (199\) 4
255
References I. Aso, K., SCHER, M. S., BARMADA, M. A.: Neonatal electroencephalography and neuropathology. J. Clin. Neurophysiol. 1989; 6: 103-123. 2. EISELT, M., ZWIENER, U., ROTHER, M., FRENZEL, J.: Zur Notwendigkeit der Erfassung neonataler Krampfaktivitiit mit der polygrafischen erweiterten Elektroenzephalografie. Piidiatrie und Grenzgebiete 1988; 27: 245-258. 3. HELLSTROM, B., KARLSSON, B., MDsSBICHLER, H.: Electrode placement in EEG of infants and its anatomical relationship studied radiographically. Electroencephalogr. Clin. Neurophysiol. 1963; 15: 115-117. 4. HENRION, G., HENRION, A., HENRION, R.: Beispiele zur Datenanalyse mit BASIC-Programmen. I. Auflage, Deutscher Verlag der Wissenschaften, Berlin 1988; 363. 5. HUTTENLOCHER, P. R.: Development of cortical neuronal activity in the neonatal cat. Exp. Neurol. 1967; 17: 247-262. 6. OKADA, Y., YONEDA, K., TANIMOTO, M.: Effect of deprivation of oxygen and/or glucose on the neurotransmission and on the level of ATP and creatine-P in the guinea pig hippocampal slices in vitro. Neuroscience 1987; 22: 397. 7. PARMELEE, A. H., AKIYAMA , Y., STERN, E.: A periodic cerebral rhythm in newborn infants. Exp. Neurol. 1969; 25: 575-584. 8. PETERS, J. F., VARNER, J. L., ELLINGSON, R. J.: Interhemispheric amplitude symmetry in the EEGsofnormal fullterm, low risk premature, and trisomy-21 infants. Electroencephalog. Clin. Neurophysiol. 1981; 51: 165-169. 9. ROTHER, M., EISELT, M., WITTE, H., ZWIENER, U.: Rechnergestiitzte EEG-Analyse im Neugeborenenalter. Wiss. Zeitschr. der WPU Rostock, 1990; 65-67. 10. SCHWARTZKROIN, P. A.: Epileptogenesis in the immature central nervous system. In: Electrophysiology of epilepsy (eds. SCHWARTZKROIN, P. A., WHEAL, H.). Academic Press, London 1984.pp. 390-412. I!. WATANABE, K., MIYAZAKI, S., HARA, K.: Behavioral state cycles, background EEG and prognosis of newborns with perinatal hypoxia. Electroencephalogr. Clin. Neurophysiol. 1980; 49: 618-625.
256
Exp. Pathol. 42 (1991) 4
Friedrich Schiller University lena, 1) Institute of Pathological Physiology, 2) Children's Hospital, lena, F.R.G.
Parameters for sensitive cerebral monitoring during the neonatal period in intensive care medicine By M. EISELT1), M. ROTHER 1), R. BOLWIN2) and U. ZWIENER1)
Address for correspondence: Dr. med. M. EISELT, Friedrich Schiller University, Institute of Pathological Physiology, LobderstraBe 3, D-O-6900 lena, F.R.G. Key words: EEG-monitoring; neonatal EEG; amplitude; EEG-patterns; righUleft hemisphere
Summary In order to bring about further reduction of neonatal mortality and morbidity, immediate and highly sensitive detection of brain-threatening situations is necessary. We investigated 52 newborns by means of 8-channel EEG recorded with normal (15 mm/ s) and compressed write-out (1 cm/min). A correct estimation of amplitude of background activity was possible by both methods. For differentiation between healthy newborns and newborns at risk, the right hemisphere amplitude of the interburst period was the most appropriate. The amplitude of the continuous EEG and the interburst period was reduced in the group of newborns at risk. In the group of the most severely disturbed newborns there was an amplitude difference between both hemispheres with greater reduction over the right hemisphere. These results emphasize the importance of the amplitude of background activity for prognostication and the necessity of using 2 interhemispheric derivations for an independent estimation of changes in both hemispheres.
Introduction During the last few years neonatal EEG has been widely accepted as a reliable parameter for describing the actual cerebral function. Until now, the visual analysis of the neonatal EEG has been of great importance. The occurrence of special patterns and the pattern-related amplitude of the EEG plays the major role (2, 11). Recent results emphasize the importance of the amplitude of background activity, especially of the interburst period for the estimation of the severity of disturbances of cortical functions (2). The prognosis further correlates well with the duration of the occurrence of these disturbances as well as the aetiology. Because EEG patterns and their amplitudes have a great prognostic importance, a method for longterm monitoring of cerebral functions recording both parameters is necessary. Thus, the aim of this study was, firstly, to compare 8 channel EEG recorded with normal write-out (15 mm/s) with that recorded with compressed write-out (I cm/min) concerning the measured amplitude of background activity, especially of the interburst period, and, secondly, to find EEG parameters which made it possible to distinguish between healthy newborns and newborns at risk. 17*
Exp. Pathol. 42 (1991) 4
251
Method We recorded the 8-channel EEG of 52 newborns between the 1st and 7th day of life (12 newborns without risk factors and 40 newborns at risk). Follow-up examinations were made until the age of 1Y2 years (table 1). EEG was recorded in a bipolar manner: Fpl-C3, C3-0 1, Fpl-T3, T3-
Table 1. Characterization of the investigated newborns (CA = conceptional age at the time of EEG monitoring examination) newborns not at risk: 12 newborns, CA = 34.9 ± 4.1 weeks normal cardiotocogram postnatal pH > = 7.2 Apgar-Score at 5 min> 6 normotrophy no artificial ventilation during the neonatal period results of follow-up examination at 1 and 112 year: - without any abnormality = 12 newborns (healthy) newborns at risk: 40 newborns, CA = 32.9±4.6 weeks pathological cardiotocogram postnatal pH ~ 7.2 Apgar-Score at 5 min> 7 hypotrophy artificial ventilation during the neonatal period results of follow-up examination at 1 and 11/2 year: - without any abnormality = 14 newborns (healthy) - with transient or slide neurological abnormalities = 6 newborns - infantile cerebral palsy = 7 newborns - died during the first 18 month of life: without pathological finding of the brain by autopsy = 6 newborns with pathological finding of the brain by autopsy = 7 newborns
01, Fp2-C4, C4-02, Fp2-T4, T4-02 [in accordance with the International 10- 20 system modified by HELLSTROM (6)]. Impedance was between 5 to 10kQ. The EEG was recorded with normal (15 mm/s) and compressed write-out (1 cm/min). The comparison was made concerning the amplitude and special EEG patterns (table 2). For statistical analysis Spearman's rank correlation coefficient, paired t-test and multidimensional analysis of variance and discriminance analysis (4) were used.
Results Estimation of the amplitude of background activity The estimation of the amplitude of the EEG background activity recorded with compressed write-out (1 cm/min) was as good as using the 8-channel EEG recorded with normal write-out (15 mm/s) [interbrust period of the discontinuous EEG (r = 0.83)]. During the occurrence of continuous as well as the burst period of discontinuous EEG there was also a significant correlation 252
Exp. Pathol. 42 (1991) 4
Table 2. Estimated parameters in the EEG EEG recorded with compressed write-out amplitude was measured in ten I-min-intervals continuous EEG pattern: - 95 % of all waves were within the limits discontinuous EEG: - this EEG was divided into 2 pattern components Burst-period Interburst-period - 95 % of all waves were within the limits the frequency of the occurrence of Burst-Interburst cycles/lO minutes EEG recorded with normal write out the amplitude during the Interburst-period
o ==
normal amplitude of background activity 1 == amplitude is within the limits > 5 f! V < 20 f! V 2 == amplitude is < 5 f! V 3 == isoelectric EEG
between the amplitude measured in the EEG-monitoring (I cm/min) and the amplitude measured in the EEG recorded with normal write-out (r == 0.68-0.73, significant alpha < 0.05).
Pattern related changes The EEG amplitude of the interburst period measured over the right hemisphere was the best of all parameters concerning a differentiation of healthy newborns and newborns with neurological abnormalities in the follow-up at I Y2 years or newborns who died (the variable with the greatest contribution to the discriminant score, T == 0.152). The next efficient parameter was the amplitude ofthe burst period measured overthe right hemisphere (T == 0.099) and the third the age-correlated 1eo _ a~ m p_ llt_ u_ d. ~(p ~V_)__________________________~ 180
140 " 120 100
eo eo 40
20
o _
,
t hUt
.p"',. _
left h
l.pIt.,.
Fig.l. Mean amplitude of the continuous EEG measured over the right and the left hemisphere in the different follow-up groups (Abbreviations used in fig. I to 3: healthy == newborns without any evidence of psychoneurological abnormalities at the follow-up at I and 3 Yz years; rep == newborns who suffer from infantile cerebral palsy; died withouUwith == newborns who died within this follow-up period withouUwith evidence of brain damage observable by autopsy). burst-interburst-duration (T == 0.009). If we use only the amplitude of the interburst and the Exp. Pathol. 42 (1991) 4
253
400 am plllude - --'(tJ_V_, _ _ _ _ _ _ _ _ _ _ _ __
_
I
I
h.m phere
_
lefl
Fig. 2. Mean amplitude of the burst period of the discontinuous EEG measured over the right and the left hemisphere in the different follow-up groups. The amplitude measured over the right hemisphere is significantly reduced (alpha < 0.05) in newborns who died without and with evidence of brain damage observable by autopsy.
1·0
amplitud (V)
120 '
.pilel'
Ie I hem
.ph.,.
Fig. 3. Mean amplitude of the interburst period of the discontinuous EEG measured over the right and the left hemisphere in the different follow-up groups . The amplitude measured over the right hemisphere is significantly reduced (alpha < 0.01) in newborns who died with evidence of brain damage observable by autopsy.
amplitude of the burst period , the following discriminant function between both groups can be concluded: D = (-1.4
* amp!.
of interburst)
* (1.1 * amp!.
of burst)
EEG-amplitude over the right and left hemisphere
In healthy newborns at the age of I Y2 years (without any pathological findings in the psychomotoric examination) in all patterns of pattern components there are no statistically significant differences between the hemispheres (figs. 1, 2, 3).In newborns who developed infantile cerebral palsy or died within the first I V2 years (with or without evidence of brain damage after autopsy) there was a decreased amplitude measured in the continuous EEG as well as the interburst period of the discontinuous EEG. During the interqurst period the amplitude was lower over the right hemisphere (alpha < 0.0 I) in newborn who died with evidence of brain damage
254
Exp. Pathol. 42 (1991) 4
observed by autopsy. During the burst period the amplitude was lower over the right hemisphere in newborns who died with and without evidence of brain damage observed by autopsy. In contrast to the amplitude of the continuous EEG and the amplitude of the interburst period, the amplitude of bursts was increased in newborns who developed infantile cerebral palsy and was reduced in newborns who died with brain damage.
Discussion The EEG amplitude representation in EEG-monitoring provided greater amounts in comparison to the EEG recorded with normal write-out (15 mm/s). This may result from the fact that very slow waves « 0.5 Hz), not regarded in normal write-out, are more visible in compressed write-out, and thus may contribute to the increase in the amplitude . Nevertheless, changes of the amplitude of background activity measured in the normal write-out (15 mm/s) are also measurable in the compressed write-out. Thus, this EEG-monitoring method provides the same information concerning the development of the amplitude of background activity as the normal EEG-recording and, in addition, makes it possible to estimate amplitude changes of low frequency range. As shown by our results, a pattern-related estimation of the background activity is necessary. The discriminant function can be interpreted as follows: With a decreasing amplitude of the interburst period the probability of a bad prognosis increases, but, in contrast, an increasing amplitude ofthe burst period increases the probability of a bad prognosis, for instance cerebral palsy. Thus, the amplitude of burst and interburst has an inverse direction of changes with increasing severity of brain damage. This increase of burst amplitude seems to disappear if the severity of brain damage further increase. Until now, the brain structure in human neonates generating the burst of discontinuous EEG has been unknown (7). In animal experiments it can be shown that burst originates in subcortical structures and causes, by means of subcortico-cortical pathways, the discontinuous EEG (5). Mild to moderate impairment of the function of the brain leads to a more synchronized discharge of neuronal populations (greater vulnerability of inhibitory neurons (10)). Their expression in the EEG is an increase in amplitude. Further increase of functional impairment leads to a reduction of the magnitude and number of postsynaptic potentials and correlates with a decrease of amplitude (6). This correlates with a very bad prognosis (death of the newborn) in our investigation. In healthy newborns there is no significant difference of the amplitude of background activity between the 2 hemispheres in all EEG patterns and pattern-related components. This result agrees with the findings of PETERS et al. (8). In newborns who died, a reduced amplitude, especially over the right hemisphere, was observed during the discontinuous EEG. This result was confirmed by data measured by pattern-related automatic EEG-anal ysis (9). It reflects a greater degree of functional disturbance of the right hemisphere . The reduction of the amplitude over the left hemisphere in dependence on the severity of disturbances is not quite obvious. This can be explained by a different degree of amplitude reduction in different frequenllY ranges, e.g. high frequencies disappear earlier than los frequencies. With the EEG-monitoring method it is impossible to distinguish between the different contributions of each frequency range to the whole amplitude. Thus, it may be possible that low frequency persists over the left hemisphere with only a small reduction in amplitude, whereas higher frequencies have already disappeared. As shown by investigations of Aso et al. (1) the sensitivity of pathological EEG asymmetry is low (40%) for the focality of morphological damage , although the specifity is relatively high (85 %). This underlines the necessity of using 2 interhemispheric derivations for an independent estimation of changes in both hemispheres . For prognostication a pattern-related and localization-related estimation of the amplitude of EEG background activity are of great importance.
Exp. Pathol. 42 (199\) 4
255
References I. Aso, K., SCHER, M. S., BARMADA, M. A.: Neonatal electroencephalography and neuropathology. J. Clin. Neurophysiol. 1989; 6: 103-123. 2. EISELT, M., ZWIENER, U., ROTHER, M., FRENZEL, J.: Zur Notwendigkeit der Erfassung neonataler Krampfaktivitiit mit der polygrafischen erweiterten Elektroenzephalografie. Piidiatrie und Grenzgebiete 1988; 27: 245-258. 3. HELLSTROM, B., KARLSSON, B., MDsSBICHLER, H.: Electrode placement in EEG of infants and its anatomical relationship studied radiographically. Electroencephalogr. Clin. Neurophysiol. 1963; 15: 115-117. 4. HENRION, G., HENRION, A., HENRION, R.: Beispiele zur Datenanalyse mit BASIC-Programmen. I. Auflage, Deutscher Verlag der Wissenschaften, Berlin 1988; 363. 5. HUTTENLOCHER, P. R.: Development of cortical neuronal activity in the neonatal cat. Exp. Neurol. 1967; 17: 247-262. 6. OKADA, Y., YONEDA, K., TANIMOTO, M.: Effect of deprivation of oxygen and/or glucose on the neurotransmission and on the level of ATP and creatine-P in the guinea pig hippocampal slices in vitro. Neuroscience 1987; 22: 397. 7. PARMELEE, A. H., AKIYAMA , Y., STERN, E.: A periodic cerebral rhythm in newborn infants. Exp. Neurol. 1969; 25: 575-584. 8. PETERS, J. F., VARNER, J. L., ELLINGSON, R. J.: Interhemispheric amplitude symmetry in the EEGsofnormal fullterm, low risk premature, and trisomy-21 infants. Electroencephalog. Clin. Neurophysiol. 1981; 51: 165-169. 9. ROTHER, M., EISELT, M., WITTE, H., ZWIENER, U.: Rechnergestiitzte EEG-Analyse im Neugeborenenalter. Wiss. Zeitschr. der WPU Rostock, 1990; 65-67. 10. SCHWARTZKROIN, P. A.: Epileptogenesis in the immature central nervous system. In: Electrophysiology of epilepsy (eds. SCHWARTZKROIN, P. A., WHEAL, H.). Academic Press, London 1984.pp. 390-412. I!. WATANABE, K., MIYAZAKI, S., HARA, K.: Behavioral state cycles, background EEG and prognosis of newborns with perinatal hypoxia. Electroencephalogr. Clin. Neurophysiol. 1980; 49: 618-625.
256
Exp. Pathol. 42 (1991) 4
