Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 2001 Mar 1;531(Pt 2):535–543. doi: 10.1111/j.1469-7793.2001.0535i.x

Non-linear changes of electrocortical activity after antenatal betamethasone treatment in fetal sheep

Matthias Schwab *, Karin Schmidt *, Marcus Roedel *, Thomas Mueller †, Harald Schubert †, M Akthar Anwar ‡, Peter W Nathanielsz ‡
PMCID: PMC2278465  PMID: 11230525

Abstract

  1. We determined the effects of betamethasone on the fetal sheep electrocorticogram (ECoG) using linear (power spectral) and non-linear analysis. For non-linear analysis we used an algorithm based on the Wolf algorithm for the estimation of the leading Lyapunov exponent which calculates a prediction error based on the course of the time series in the phase space. A high prediction error stands for low predictibility or low regularity and vice versa.

  2. After 48 h of baseline recordings, vehicle (n = 6) or betamethasone (n = 7) at 10 μg h−1 was infused over 48 h to the sheep fetus at 128 days gestational age (0.87 of gestation).

  3. ECoG spectral analysis revealed no difference in power spectrum between vehicle- and betamethasone-treated fetuses. The prediction error of the ECoG during REM sleep was higher than during non-REM or quiet sleep in both groups (P < 0.0001) revealing lower causality of brain activity during REM sleep. During REM sleep, prediction error significantly decreased 18-24 h after onset of betamethasone treatment (P < 0.05) and returned to baseline values within the following 24 h of continued betamethasone treatment. No ECoG changes were found during quiet sleep. Non-linear ECoG changes during metabolically active REM sleep accompanied the previously described decrease in cerebral blood flow.

  4. These results suggest that betamethasone in doses used in perinatal medicine acutely alters complex neuronal activity.

Synthetic glucocorticoids are widely used in perinatal medicine to accelerate fetal lung maturation in babies at risk of premature delivery (reviews in Crowley et al. 1990; Ballard & Ballard, 1995). Acute reduction in fetal body movements, fetal breathing movements and heart rate has been reported following maternal betamethasone treatment in human fetuses (Mulder et al. 1997; Senat et al. 1998). Such behaviour changes indicate an altered functional state of the brain that we hypothesised would be revealed by changes in the electrocortical activity. The aim of the present study was to investigate the changes in brain function that accompany betamethasone administration to the sheep fetus in doses that produce plasma concentrations similar to those in human umbilical cord blood of babies whose mothers received antenatal glucocorticoids (Derks et al. 1997).

We analysed the fetal electrocorticogram (ECoG) by linear (power spectral analysis) and non-linear methods. It has been shown in humans that the EEG is a highly complex signal that can probably not be described entirely by linear methods (Pritchard et al. 1995; Palus, 1996; Theiler & Rapp, 1996; Micheloyannis et al. 1998). We have previously determined that the fetal sheep ECoG contains deterministic and non-linear portions (Schwab et al. 2000b). For non-linear ECoG analysis, we used an algorithm based on the Wolf algorithm (calculation of the leading Lyapunov exponent) which calculates a so-called point prediction error based on the course of the time series in the phase space. Thus, it is a measure of local exponential divergence and quantifies the theoretical predictability or causality of such a time series (Schmidt et al. 1997; Schwab et al. 2000b). A high prediction error reveals low predictability or causality. The reverse also holds, namely a low prediction error indicates a high degree of regularity. This approach has proved to be effective in detecting distinct changes of the functional state of the fetal brain (Schmidt et al. 1997; Schwab et al. 2000b). We used the combination of these methods to detect an effect of betamethasone treatment on complex neuronal interactions.

METHODS

Surgical procedure

All procedures were approved by the Cornell University Animal Use and Care Committee and were performed in facilities approved by the American Association for the Accreditation of Laboratory Animal Care. Thirteen Rambouillet-Colombia ewes bred on a single occasion and of known gestational age were acclimated to the animal facilities for at least 5 days before surgery and kept in rooms with controlled light-dark cycles (14 h light:10 h dark; lights off at 21.00 h and lights on at 07.00 h). Alfalfa cubes and water were provided ad libitum. Food and water were withdrawn for 24 h before surgery, which was performed at 120 ± 1 days gestational age (dGA). Ewes were pretreated with 1 g of ketamine (Ketaset, Fort Doge Animal Health, IA, USA) i.m. and 0.8 mg of glycopyrrolate (American Regent Laboratories, Inc., subsidiary of Luitpold Pharmaceuticals, Inc., Shirley, NY, USA) i.m. Halothane (Halocarbon Laboratories, River Edge, NJ, USA) was administered by face mask to permit tracheal intubation. Anaesthesia was maintained with 1.5 % halothane-98.5 % oxygen. Surgery was performed using techniques previously described in detail (Nathanielsz et al. 1980). Briefly, ewes were instrumented with catheters inserted into the carotid artery for blood sampling and into the jugular vein for postoperative administration of antibiotics. The fetuses were instrumented with polyvinyl catheters (Tygon, Norton Performance Plastics; 0.1 mm i.d., 0.18 mm o.d.) inserted into the axillary artery, jugular and pedal vein as well as in the amniotic cavity. Stainless-steel screw electrodes soldered to multistranded stainless-steel wire (Cooner, Chatsworth, CA, USA; no. AS 632) were implanted bilaterally on the fetal parietal skull to record fetal electrocorticogram (ECoG) and above the orbit on each side to monitor the electro-oculogram (EOG). Stainless-steel wire electrodes were implanted into the muscles of the left and right shoulder and above the sternum to record the electrocardiogram (ECG), on the fetal diaphragm to monitor fetal breathing movements (FBM), in the neck muscles to record tonic nuchal muscle activity and on the surface of the uterine body to record myometrial activity.

After surgery the ewes were returned to a metabolism cage and provided with free access to food and water. Animals were allowed at least 5 days of postoperative recovery before the start of any observations. During this time ewes received a daily dose of 1 g ampicillin (AMP-Equine; SmithKline Beecham, West Chester, PA, USA) i.v. and 1 g ampicillin into the amniotic cavity. In addition, ewes were treated with phenylbutazone orally (Equiphene paste, Luitpold Pharmaceuticals, Shirley, NY, USA), 0.5 g twice daily for 3 days, to provide postoperative analgesia. All catheters were maintained patent via a continuous infusion of heparin at 12.5 i.u. ml−1 in 0.9 % NaCl solution delivered at 0.5 ml h−1.

At the end of the experiment the ewes were anaesthetised with 4 % halothane and the fetuses delivered by Caesarean section. Ewes were killed by i.v. injection of pentobarbital sodium solution (Fatal-Plus, Vortech Pharmaceuticals, Dearborn, MI, USA).

Experimental procedure

Baseline recordings of fetal arterial blood pressure and amniotic pressure, ECoG, EOG, nuchal electromyogram (EMG) and FBM were started at 126 dGA. All variables were monitored continuously throughout the experiment. Daily fetal and maternal arterial blood samples were taken at 09.00 h throughout the study for measurement of pH and blood gases using a blood gas analyser (ABL600, Radiometer, Copenhagen, Denmark; measurements corrected to 39°C). Haemoglobin concentration and oxygen saturation were measured photometrically (Hemoximeter OSM2, Radiometer). At 128 dGA, after 48 h of baseline recordings, vehicle (n = 7) or betamethasone at a rate of 10 μg h−1(n = 6, Celestone Soluspan, Schering, Kenilworth, NJ, USA) infusion was started into the fetal jugular vein at 11.00 h and maintained over the next 48 h.

Data acquisition and analysis

ECoG, EOG, nuchal EMG and FBM signals were obtained as previously described (Nathanielsz et al. 1980). Fetal ECoG, EOG, diaphragm and nuchal EMG signals were filtered (band-pass ranges: ECoG and EOG 0.5-100 Hz, EMG 10-1000 Hz), amplified and digitised using a 16-channel A/D board (DT 2801F, Data Translation, Marlborough, MA, USA). All biophysical variables were sampled at 128 Hz, stored continuously on a PC hard disk and, additionally, recorded on a multichannel chart recorder (TA11, Gould, Valley View, OH, USA). For linear and non-linear ECoG analysis, we chose five artefact-free 10 min ECoG epochs of the respective sleep states from each animal recorded in the early morning hours of each experimental day at a time when the animals were not disturbed by laboratory personnel or experimental procedures. Non-REM or quiet sleep was characterised by a high-voltage, slow-frequency ECoG pattern, absence of sustained rapid eye movements, and occasional nuchal tone. REM sleep was distinguished by a low-voltage, high-frequency ECoG, occurrence of rapid eye and breathing movements, and general absence of nuchal tone. ECoG recordings during myometrial contractures were not used for data analysis.

For power spectral analysis, ECoG was quantified continuously in 4 s summaries over 10 min. Fast Fourier transformation was used to evaluate absolute and relative spectral band power of the total band (1.5-30 Hz), delta band (1.5-4 Hz), theta band (4-8 Hz), alpha band (8-13 Hz) and beta band (13-30 Hz) as well as the spectral edge frequency of the total band (1.5-30 Hz). The spectral edge frequency is defined as the frequency below which 95 % of the power resides. The software used was WATISA (Institute of Medical Statistics, Computer Sciences and Documentation, University of Jena, Germany).

For non-linear analysis, we developed an algorithm based on the Wolf algorithm for the estimation of the leading Lyapunov exponent, which quantifies the local exponential divergence of neighbouring trajectories in the phase space similar to Gao & Zheng (1994), who did not use this approach for single time points but for defined short time windows. The local exponential divergence is a measure of the theoretical predictability and causality of the investigated time series.

The spectra of the investigated fetal ECoG are confined to frequencies between 1.5 and 30 Hz. To reduce the data length for non-linear analysis, the data sets were sampled down to 64 Hz after a digital low pass filtering with an upper cut-off frequency of 32 Hz. A test procedure using data sets with a sampling frequency of 128 and 64 Hz showed no differences in the resulting predictability. For each time point we evaluated a single prediction error based on the course of the time series in the phase space. We averaged the single prediction error for each data point to produce a prediction error of the respective data set.

The theoretical starting point of our algorithm is a one-dimensional time series {x(ti)}i=1,…n as a projection of a measured biomedical process. According to Takens (1981) it is possible to transform such a time series in a multidimensional phase space by means of a time delay τ.

Starting with any point y(ti) on a trajectory in the phase space we searched for the nearest (Euclidean) neighbour in the phase space. The distance of this point to the starting point is D(ti). After a specific time step k the distance of the evolved points in the phase space is D‘(ti +k). This procedure is repeated for every point y(ti) in the phase space. We are able to estimate single prediction errors PPi at every time point i according to:

graphic file with name tjp0531-0535-m1.jpg (1)

where i = 1, …n– (De– 1)τ–k, ti is actual time, y(ti) is a point in the phase space, D(ti) is the smallest Euclidean distance at the time ti, D‘(ti+k) is the evolved distance at the time ti+k, De is the embedding dimension, τ is the time delay, k is the evolving steps, f is the sampling frequency and PPi is the single prediction error.

The logarithmic relation of eqn (1) is a measure of the local exponential divergence (or convergence) of neighbouring trajectories in the phase space. A high positive value of the prediction error means a very low predictability and causality; PPi < 0 represents a periodic/quasiperiodic process or convergence to a steady state.

As in all non-linear measures, the numerical values of the prediction error depend on the availability of a sufficient number of data points and several other parameters as the embedding dimension, the time delay and the evolving time. For a one-to-one transformation of the measured one-dimensional time series in a multi-dimensional phase space an embedding dimension De≥ 2 CD is necessary, where CD is the correlation dimension of the process. To define an appropriate time lag τ we chose the first point with ACF < 1/e, where ACF is the autocorrelation function. Thus, we used an embedding dimension of 16 and a time delay of 150 ms. The influence of the evolving time k was tested carefully taking into consideration the different frequency distribution of the ECoG during REM and quiet sleep. We have chosen the evolving time such that it reflects similar portions of relative spectral power during epochs of high and low ECoG frequency states (REM and quiet sleep). An evolving time of 75 ms seemed to be appropriate for both sleep states. ECoG epochs of 10 min containing 38 400 data points proved to be sufficient for calculation of the prediction error.

All results are given as means ±s.e.m. Results of power spectral analysis are normalised to the values at 126 dGA. Non-parametric tests were used for statistical analysis as spectral parameters of the ECoG do not follow a normal distribution (Gasser et al. 1981). As we did not test the distribution of the prediction error we had to assume a non-Gaussian distribution as well. Differences of the spectral values and the prediction error within the experimental groups were tested for significance by Wilcoxon's sign rank test. The Mann-Whitney rank sum test was used for comparisons between the experimental groups. P values < 0.05 were considered to be significant.

RESULTS

Fetal physiological parameters remained unchanged during vehicle or betamethasone infusion (Table 1). No differences were found between the vehicle- and betamethasone-treated fetuses either during the baseline or the infusion periods.

Table 1.

Physiological parameters before and during vehicle or betamethasone treatment

pH P∞2 (mmHg) PO2 (mmHg) O2% Hb (g dl−1)
Vehicle Beta Vehicle Beta Vehicle Beta Vehicle Beta Vehicle Beta
Baseline 7.37 ± 0.01 7.35 ± 0.01 47.4 ± 2.2 50.1 ± 0.9 23.3 ± 1.2 24.4 ± 1.0 57 ± 3 64 ± 3 11.3 ± 0.9 10.4 ± 0.4
24 h 7.32 ± 0.04 7.36 ± 0.01 49.5 ± 1.6 49.4 ± 0.5 21.4 ± 1.6 22.7 ± 0.9 58 ± 6 59 ± 3 10.6 ± 0.9 11.5 ± 0.5
48 h 7.35 ± 0.01 7.37 ± 0.01 47.5 ± 4.1 49.0 ± 2.0 23.2 ± 4.1 23.2 ± 1.3 54 ± 2 60 ± 3 10.8 ± 0.4 11.5 ± 0.7
Open in a new tab

Values are means ± S.E.M.; vehicle-treated fetuses, n = 7; betamethasone (Beta)-treated fetuses, n = 6.

Maturational changes in fetal ECoG

Power spectral analysis of the ECoG showed the typical frequency distribution of the fetal ECoG with the maximum relative spectral power in the delta band during quiet sleep (Fig. 1) and a higher portion of theta, alpha and beta activity during REM sleep than during quiet sleep revealing ECoG desynchronisation in REM sleep (Fig. 1). A significant increase of absolute ECoG power occurred over the experimental period in both animal groups (P < 0.05, Figs 2 and 3). During REM sleep, ECoG power of the beta band increased (P < 0.05, Fig. 2). During quiet sleep, ECoG power increased in all frequency bands (P < 0.05, Fig. 3). Betamethasone had no effect because the increase of spectral power was also present in the vehicle-treated animals as well as before onset of betamethasone exposure at 128 dGA. Analysis of spectral edge frequency as a sensitive measure of changes in the frequency distribution of the fetal ECoG (Szeto, 1990) clearly demonstrated a shift to higher frequencies in REM and to lower frequencies in quiet sleep indicating functional maturation of sleep states (P < 0.05, Fig. 4). Separation of spectral edge frequency between REM and quiet sleep occurred in both groups of fetuses and is hence independent of exposure to betamethasone. The frequency shift to higher frequencies in REM sleep is due to a decrease of relative spectral power in the delta band and an increase of relative spectral power in the beta band (P < 0.05, Fig. 1). The frequency shift to lower frequencies in quiet sleep is due to an increase of relative spectral power in the delta band and a decrease of relative spectral power in the other frequency bands (P < 0.05, Fig. 1).

Figure 1. Relative spectral power of the ECoG in REM and in quiet sleep in the vehicle- and betamethasone-treated fetuses 48 h before and during 48 h of exposure.

Figure 1

Open in a new tab

Total spectral power equals 1.0. □, vehicle-treated, n = 7; ▪, betamethasone-treated fetuses, n = 6; means +s.e.m.; *P < 0.05 in comparison to 126 dGA.

Figure 2. Spectral power of the ECoG normalised to 126 dGA in REM sleep in the vehicle- and betamethasone-treated fetuses 48 h before and during 48 h of exposure.

Figure 2

Open in a new tab

□, vehicle-treated, n = 7; ▪, betamethasone-treated fetuses, n = 6; means +s.e.m.; *P < 0.05 in comparison to 126 dGA.

Figure 3. Spectral power of the ECoG normalised to 126 dGA in quiet sleep in the vehicle- and betamethasone-treated fetuses 48 h before and during 48 h of exposure.

Figure 3

Open in a new tab

□, vehicle-treated, n = 7; ▪, betamethasone-treated fetuses, n = 6; means +s.e.m.; *P < 0.05 in comparison to 126 dGA.

Figure 4. Spectral edge frequency of the total band (1.5-30 Hz) of the ECoG in REM and in quiet sleep in the vehicle- and betamethasone-treated fetuses 48 h before and during 48 h of exposure.

Figure 4

Open in a new tab

□, vehicle-treated, n = 7; ▪, betamethasone-treated fetuses, n = 6; means +s.e.m.; *P < 0.05 in comparison to 126 dGA.

The prediction error of the ECoG during REM sleep was significantly higher than during quiet sleep in both animal groups revealing the lower causality of brain activity during REM sleep (P < 0.0001, Fig. 5). The higher standard deviation of the prediction error during REM than during quiet sleep is due to fluctuations of the predictability of the ECoG within the evaluated 10 min epochs and reveals the lower predictability of the ECoG in REM sleep as well. The prediction error increased during REM sleep and decreased during quiet sleep with gestational age (P < 0.05, Fig. 5). Changes in the frequency distribution and predictability of the ECoG show a continuous functional development of the brain after emergence of organised sleep states.

Figure 5. Changes of the prediction error of the ECoG during betamethasone exposure.

Figure 5

Open in a new tab

□, vehicle-treated, n = 7; ▪, betamethasone-treated fetuses, n = 6; means +s.e.m.*P < 0.05 in comparison to 126 dGA, †P < 0.05 in comparison to 128 dGA, ‡P < 0.05 in comparison to vehicle-treated fetuses.

Effects of betamethasone

Betamethasone exposure did not have a significant effect on spectral power of the ECoG either in REM or in quiet sleep (Figs 1–4).

After 18 h of betamethasone exposure, the prediction error of the metabolically active REM sleep was significantly decreased (P > 0.05, Fig. 5). However, the prediction error of REM sleep was still significantly higher than that of quiet sleep (P > 0.0001). The prediction error of REM sleep returned to baseline levels on the second day of continued betamethasone treatment. Betamethasone exposure did not alter the prediction error of quiet sleep. The prediction error in REM sleep of the vehicle-infused fetuses did not show any decrease over the experimental period.

DISCUSSION

Using non-linear ECoG analysis, we were able to demonstrate that betamethasone at doses clinically used in perinatal medicine alters the electrical brain activity in fetal sheep during REM sleep. These changes could not be demonstrated by power spectral analysis suggesting an effect of betamethasone on complex neuronal interactions that cannot be described by this linear approach. In humans, changes of non-linear measures of the EEG such as correlation dimension, Kolmogorov entropy, leading Lyapunov exponent or mutual dimension have been demonstrated under physiological conditions such as evoked emotions (Jeong et al. 1997; Aftanas et al. 1998; Micheloyannis et al. 1998) or in such neurological disorders as schizophrenia (Röschke et al. 1994), epilepsy (Stam et al. 1998), Alzheimer's disease (Jeong et al. 1998) and Creutzfeldt-Jacob disease (Stam et al. 1997). Nevertheless, the better performance of the non-linear approach in describing the effect of betamethasone in our study is no proof of non-linearity by itself. We have previously determined by surrogate data testing that the ECoG of fetal sheep at a similar gestational age contains deterministic and non-linear portions during unaffected REM and quiet sleep (Schwab et al. 2000b). A complete consideration of the usefulness of non-linear and linear methods in quantifying the fetal ECoG would require a comparison of changes in the prediction error based on non-linear analysis with changes of the prediction error based on a linear autoregressive (AR) model.

We have previously demonstrated that 24 h of betamethasone exposure results in significant reductions of cerebral blood flow (CBF) in all brain regions but the hippocampus (Schwab et al. 2000a). Blood flow decreased in the cerebral cortex and in the basal ganglia by about 35-40 % and in the thalamus and hindbrain by about 45-50 % in comparison to baseline levels. Forty-eight hours after onset of betamethasone infusion, CBF reduction was diminished to about 25-30 % by which time CBF was not significantly different from baseline values but still significantly different from that of vehicle-treated fetuses except in the parietotemporal, parietooccipital and occipital cortex. The CBF decrease at 24 h of betamethasone exposure would tend to alter the cerebral oxidative metabolism leading to insuffiency of neuronal metabolism since about half of cerebral glucose and oxygen metabolism is used to meet the energy requirements of the ionic pumps that maintain the ionic gradients across cellular membranes (Erecinska & Silver, 1989). The first occurrence of altered EEG pattern in the adult brain of humans (Trojaborg & Boysen, 1973) or cats (Hossmann & Schuier, 1980) was found at a flow reduction of about 50-60 %. Although we demonstrated betamethasone-induced flow reductions of up to 50 % in the cortex and thalamus, regions essential to ECoG generation, we were unable to demonstrate betamethasone-induced ECoG changes using power spectral analysis. Using non-linear ECoG analysis, however, we were able to detect a distinct reduction of the prediction error of the ECoG in REM sleep during the period of decreased CBF induced by betamethasone infusion. The hindbrain as the region with the most pronounced CBF reduction (Schwab et al. 2000a) plays a substantial role in induction and maintenance of REM sleep (Steriade & McCarley, 1990). Moreover, cortical activation during REM sleep requires an increase in cerebral oxidative metabolism (Richardson et al. 1985; Abrams et al. 1988) and is associated with an increase of CBF (Jensen et al. 1986). The impaired CBF that occurs during betamethasone infusion is likely to prevent the neurons from completely meeting the metabolic demand of cortical activation during REM sleep.

The discussion so far has been based on the notion that the increase in the predictability and causality of the ECoG is secondary to the CBF decrease. This inference is supported by the observation that cerebral metabolic rate and cerebral perfusion are closely coupled in all mammalian species studied to date (Iadecola, 1993). However, the CBF reduction and the increase in the predictability of the ECoG might also result from a catabolic effect of betamethasone on the CNS. It has been shown that glucocorticoids decrease cerebral glucose uptake (Landgraf et al. 1978) and, conversely, adrenalectomy in rats increases local cerebral glucose utilisation, which can be reversed by glucocorticoids (Kadekaro et al. 1988).

Direct actions of glucocorticoids on neuronal excitability and transmitter-activated ionic conductance may explain the betamethasone-induced increased causality of the ECoG during REM sleep (reviews in Joels & de Kloet, 1992, 1994). Glucocorticoid effects on intrinsic membrane properties are dependent on the receptor type and condition of the membrane potential (hyperpolarisation/depolarisation) (reviews in Joels & de Kloet, 1992, 1994). Occupancy of mineralocorticoid or type I receptors that are located predominantly in the hippocampus and hypothalamus (Ahima et al. 1991) suppresses 5HT hyperpolarisations, enhances excitatory amino acid-mediated synaptic transmission and decreases Ca2+-dependent conductance. In contrast, glucocorticoid or type II receptors particularly interact with the inhibitory noradrenergic system and activation of type II receptors suppresses excitability in the hippocampal CA1 area (Joels & de Kloet, 1992, 1994). Betamethasone has a specific affinity for type II receptors that are present in most brain regions including the cerebral cortex (Ahima et al. 1990; Cintra et al. 1994). Type II receptors are present in the brain of fetal sheep at the gestational age at which we conducted our study (Rose et al. 1985; Yang et al. 1990). However, there is only very limited knowledge of glucocorticoid effects on membrane properties of neurons outside the hippocampus (Hua & Chen, 1989). The inhibition of the excitability of larger cortical neuronal populations could contribute to the increased causality of the ECoG observed in the present study. In rabbits, inhibition of neuronal excitability by dexamethasone preventing epileptiform activity has been reported as well (Pieretti et al. 1992).

The increase of the predictability and, thus, causality of the ECoG in REM sleep during betamethasone exposure was not accompanied by a change of the incidence of REM or quiet sleep (Derks et al. 1997). In adult humans, however, glucocorticoids caused a decrease of the incidence of REM sleep and an increase of the incidence of quiet sleep (Born et al. 1991; Bohlhalter et al. 1997; Steiger et al. 1998). It appears that activation of type II receptors mediates decreased incidence of REM sleep (Born et al. 1991). Spectral analysis revealed significant increases in delta and theta power (Bohlhalter et al. 1997). In adult rabbits, cortisone treatment reduced the amount of time spent in REM sleep as well but did not alter the amount of time spent in quiet sleep (Toth et al. 1992). In adult rats, prolonged cortisone treatment reduced the amount of quiet sleep (Bradbury et al. 1998).

Changes of electrical membrane properties may be due to both genomic and non-genomic mechanisms. Glucocorticoids rapidly depress the neuronal firing activity (Hua & Chen, 1989). This action may be due to glucocorticoid effects on membrane-associated receptor-gated ion-channels or activation of second messenger systems that modulate electrical membrane characteristics (for review see Moore et al. 1995). Delayed actions of glucocorticoids on neuronal firing pattern are likely to be due to a genomic action of glucocorticoids via intracellular receptors (Joels & de Kloet, 1992; McEwen, 1994). Genomic actions of glucocorticoids on lymphocytes have been demonstrated with onset latencies of only 20 min (Pfaff et al. 1971).

Further studies are required to determine which of these three major mechanisms accounts for the non-linear ECoG changes observed. The various mechanisms are not mutually exclusive and may operate at different concentrations and following different exposure periods. Pre-existing fetal stress or other altered baseline situations are also likely to be of considerable importance.

In conclusion, non-linear ECoG analysis provides information that enhances findings obtained by power spectral analysis. Our results suggest that betamethasone at doses resembling those to which the human fetus is exposed during antenatal glucocorticoid therapy acutely alters complex neuronal activity and these changes may contribute to the altered behavioural function of the human and sheep fetuses during glucocorticoid treatment.

Acknowledgments

We thank Dr Xiu-Ying Ding for the excellent surgical assistance, Dr R. A. Wentworth and Doug Kliever for the help with the data acquisition system, and Karen Moore for the help with the manuscript. This work was supported by the Deutsche Akademie der Naturforscher Leopoldina, Boehringer Ingelheim Fonds and NIH grant HD 21840.

References

  1. Abrams RA, Hutchison AA, Jay TM, Sokoloff L, Kennedy C. Local cerebral glucose utilisation non-selectively elevated in rapid eye movement sleep of the fetus. Brain Research. 1988;468:65–70. doi: 10.1016/0165-3806(88)90008-9. [DOI] [PubMed] [Google Scholar]
  2. Aftanas LI, Lotova NV, Koshkarov VI, Makhnev VP, Mordvintsev YN, Popov SA. Non-linear dynamic complexity of the human EEG during evoked emotions. International Journal of Psychophysiology. 1998;28:63–76. doi: 10.1016/s0167-8760(97)00067-6. [DOI] [PubMed] [Google Scholar]
  3. Ahima RS, Harlan RE. Charting of type II glucocorticoid receptor-like immunoreactivity in the rat central nervous system. Neuroscience. 1990;39:579–604. doi: 10.1016/0306-4522(90)90244-x. [DOI] [PubMed] [Google Scholar]
  4. Ahima RS, Krozowski Z, Harlan RE. Type I corticosteroid receptor-like immunoreactivity in the rat CNS: distribution and regulation by corticosteroids. Journal of Comparative Neurology. 1991;313:522–538. doi: 10.1002/cne.903130312. [DOI] [PubMed] [Google Scholar]
  5. Ballard PL, Ballard RA. Scientific basis and therapeutic regimens for use of antenatal glucocorticoids. American Journal of Obstetrics and Gynecology. 1995;173:254–262. doi: 10.1016/0002-9378(95)90210-4. [DOI] [PubMed] [Google Scholar]
  6. Bohlhalter S, Murck H, Holsboer F, Steiger A. Cortisol enhances non-REM sleep and growth hormone secretion in elderly subjects. Neurobiology and Aging. 1997;18:423–429. doi: 10.1016/s0197-4580(97)00036-5. [DOI] [PubMed] [Google Scholar]
  7. Born J, Dekloet ER, WenZ H, Kern W, Fehm HL. Gluco- and antimineralocorticoid effects on human sleep: a role of central corticosteroid receptors. American Journal of Physiology. 1991;260:E183–188. doi: 10.1152/ajpendo.1991.260.2.E183. [DOI] [PubMed] [Google Scholar]
  8. Bradbury MJ, Dement WC, Edgar DM. Effects of adrenalectomy and subsequent corticosterone replacement on rat sleep state and EEG power spectra. American Journal of Physiology. 1998;275:R555–565. doi: 10.1152/ajpregu.1998.275.2.R555. [DOI] [PubMed] [Google Scholar]
  9. Cintra A, Zoli M, Rosen L, Agnati LF, Okret S, Wikstrom AC, Gustaffsson AJ, Fuxe K. Mapping and computer assisted morphometry and microdensitometry of glucucorticoid receptor immunoreactive neurons and glial cells in the rat central nervous system. Neuroscience. 1994;62:843–897. doi: 10.1016/0306-4522(94)90481-2. [DOI] [PubMed] [Google Scholar]
  10. Crowley P, Chalmers I, Keirse MJ. The effects of corticosteroid administration before preterm delivery: an overview of the evidence from controlled trials. British Journal of Obstetrics and Gynaecology. 1990;97:11–25. doi: 10.1111/j.1471-0528.1990.tb01711.x. [DOI] [PubMed] [Google Scholar]
  11. Derks JB, Giussani DA, Jenkins SL, Wentworth RA, Visser GH, Padbury JF, NathanielsZ PW. A comparative study of cardiovascular, endocrine and behavioural effects of betamethasone and dexamethasone administration to fetal sheep. Journal of Physiology. 1997;499:217–226. doi: 10.1113/jphysiol.1997.sp021922. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Erecinska M, Silver IA. ATP and brain function. Journal of Cerebral Blood Flow and Metabolism. 1989;9:2–19. doi: 10.1038/jcbfm.1989.2. [DOI] [PubMed] [Google Scholar]
  13. Gao J, Zheng Z. Direct dynamical test for deterministic chaos. Europhysics Letters. 1994;25:485–490. [Google Scholar]
  14. Gasser T, Bächer P, Möcks J. Transformation towards the normal distribution of broad band spectral parameters of the EEG. Electroencephalography and Clinical Neurophysiology. 1981;53:119–124. doi: 10.1016/0013-4694(82)90112-2. [DOI] [PubMed] [Google Scholar]
  15. Hossmann K-A, Schuier F. Experimental brain infarcts in cats. I. Pathophysiological observations. Stroke. 1980;11:583–592. doi: 10.1161/01.str.11.6.583. [DOI] [PubMed] [Google Scholar]
  16. Hua SY, Chen YZ. Membrane receptor-mediated electrophysiological effects of glucocorticoid on mammalian neurons. Endocrinology. 1989;124:687–691. doi: 10.1210/endo-124-2-687. [DOI] [PubMed] [Google Scholar]
  17. Iadecola C. Regulation of the cerebral microcirculation during neural activity: is nitric oxide the missing link? Trends in Neuroscience. 1993;16:206–214. doi: 10.1016/0166-2236(93)90156-g. [DOI] [PubMed] [Google Scholar]
  18. Jensen A, Bamford OS, Dawes GS, Hofmeyr G, Parkes MJ. Changes in organ blood flow between high and low voltage electrocortical activity in fetal sheep. Journal of Developmental Physiology. 1986;8:187–194. [PubMed] [Google Scholar]
  19. Jeong J, Joung MK, Kim SY. Quantification of emotion by nonlinear analysis of the chaotic dynamics of electroencephalogram during perception of 1/f music. Biological Cybernetics. 1997;78:217–225. doi: 10.1007/s004220050428. [DOI] [PubMed] [Google Scholar]
  20. Jeong J, Kim SY, Han SH. Non-linear dynamical analysis of the EEG in Alzheimer's disease with optimal embedding dimension. Electroencephalography and Clinical Neurophysiology. 1998;106:220–228. doi: 10.1016/s0013-4694(97)00079-5. [DOI] [PubMed] [Google Scholar]
  21. Joels M, De Kloet ER. Control of neuronal excitability by corticosteroid hormones. Trends in Neuroscience. 1992;15:25–30. doi: 10.1016/0166-2236(92)90345-9. [DOI] [PubMed] [Google Scholar]
  22. Joels M, De Kloet ER. Mineralocorticoid and glucocorticoid receptors in the brain. Implications for ion permeability and transmitter systems. Progress in Neurobiology. 1994;43:1–36. doi: 10.1016/0301-0082(94)90014-0. [DOI] [PubMed] [Google Scholar]
  23. Kadekaro M, Ito M, Gross PM. Local cerebral glucose utilization is increased in acutely adrenalectomized rats. Neuroendocrinology. 1988;47:329–334. doi: 10.1159/000124933. [DOI] [PubMed] [Google Scholar]
  24. Landgraf R, Mitro A, Hess J. Regional net uptake of 14C-glucose by rat brain under the influence of corticosterone. Endocrinologia Experimentalis. 1978;12:119–129. [PubMed] [Google Scholar]
  25. McEwen BS. Steroid hormone actions on the brain: when is the genome involved? Hormones and Behaviour. 1994;28:396–405. doi: 10.1006/hbeh.1994.1036. [DOI] [PubMed] [Google Scholar]
  26. Micheloyannis S, Flitzanis N, Papanikolaou E, Bourkas M, Terzakis S, Arvanitis S, Stam JC. Usefulness of non-linear EEG analysis. Acta Neurologica Scandinavica. 1998;97:13–19. doi: 10.1111/j.1600-0404.1998.tb00603.x. [DOI] [PubMed] [Google Scholar]
  27. Moore FL, Orchinik M, Lowry C. Functional studies of corticosterone receptors in neuronal membranes. Receptor. 1995;5:21–28. [PubMed] [Google Scholar]
  28. Mulder EJ, Derks JB, Visser GH. Antenatal corticosteroid therapy and fetal behaviour: a randomised study of the effects of betamethasone and dexamethasone. British Journal of Obstetrics and Gynaecology. 1997;104:1239–1247. doi: 10.1111/j.1471-0528.1997.tb10969.x. [DOI] [PubMed] [Google Scholar]
  29. Nathanielsz PW, Bailey A, Poore ER, Thorburn GD, Harding R. The relationship between myometrial activity and sleep state and breathing in fetal sheep throughout the last third of gestation. American Journal of Obstetrics and Gynecology. 1980;138:653–659. doi: 10.1016/0002-9378(80)90083-6. [DOI] [PubMed] [Google Scholar]
  30. Palus M. Nonlinearity in normal human EEG: cycles, temporal asymmetry, nonstationarity and randomness, not chaos. Biological Cybernetics. 1996;75:389–396. doi: 10.1007/s004220050304. [DOI] [PubMed] [Google Scholar]
  31. Pfaff DW, Silva MT, Weiss JM. Telemetered recording of hormone effects on hippocampal neurons. Science. 1971;172:394–395. doi: 10.1126/science.172.3981.394. [DOI] [PubMed] [Google Scholar]
  32. Pieretti S, Di Giannuario A, Loiozz A, Sagratella S, Scotti De Carolis A, Capasso A, Sorrentino L. Dexamethasone prevents epileptiform activity induced by morphine in in vivo and in vitro experiments. Journal of Pharmacology and Experimental Therapeutics. 1992;263:830–839. [PubMed] [Google Scholar]
  33. Pritchard WS, Duke DW, Krieble KK. Dimensional analysis of resting human EEG II: surrogate-data testing indicates nonlinearity but not low-dimensional chaos. Psychophysiology. 1995;32:486–491. doi: 10.1111/j.1469-8986.1995.tb02100.x. [DOI] [PubMed] [Google Scholar]
  34. Richardson BS, Patrick JE, Abduljabbar H. Cerebral oxidative metabolism in the fetal lamb: relationship to electrocortical state. American Journal of Obstetrics and Gynecology. 1985;153:426–431. doi: 10.1016/0002-9378(85)90081-x. [DOI] [PubMed] [Google Scholar]
  35. Röschke J, Mann K, Fell J. Nonlinear EEG dynamics during sleep in depression and schizophrenia. International Journal of Neuroscience. 1994;75:271–284. doi: 10.3109/00207459408986309. [DOI] [PubMed] [Google Scholar]
  36. Rose JC, Kute TE, Winkler L. Glucocorticoid receptors in sheep brain tissues during development. American Journal of Physiology. 1985;249:E345–349. doi: 10.1152/ajpendo.1985.249.4.E345. [DOI] [PubMed] [Google Scholar]
  37. Schmidt K, Eiselt M, Schwab M, Bauer R, Zwiener U, Hoyer D. The quality of the point prediction error to quantify and classify state changes. In: Witte H, Zwiener U, Schack B, editors. Quantitative and Topological EEG and MEG Analysis. Erlangen, Germany: Druckhaus Mayer; 1997. pp. 263–265. [Google Scholar]
  38. Schwab M, Roedel M, Anwar MA, Buchwalder LF, Müller T, Schubert H, Walter B, NathanielsZ PW. Effects of betamethasone administration to the fetal sheep in late gestation on fetal cerebral blood flow. Journal of Physiology. 2000a;528:619–632. doi: 10.1111/j.1469-7793.2000.00619.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Schwab M, Schmidt K, Witte H, Abrams RM. Investigation of nonlinear ECoG changes during spontaneous sleep state changes and cortical arousal in fetal sheep. Cerebral Cortex. 2000b;10:142–48. doi: 10.1093/cercor/10.2.142. [DOI] [PubMed] [Google Scholar]
  40. Senat MV, Minoui S, Multon O, FernandeZ H, Frydman R, Ville Y. Effects of dexamethasone and betamethasone on fetal heart rate variability in preterm labour: a randomised study. British Journal of Obstetrics and Gynaecology. 1998;105:749–755. doi: 10.1111/j.1471-0528.1998.tb10206.x. [DOI] [PubMed] [Google Scholar]
  41. Stam CJ, Nicolai J, Keunen RWM. Nonlinear dynamical analysis of periodic lateralized epileptiform discharges. Clinical Electroenephalography. 1998;29:101–105. doi: 10.1177/155005949802900209. [DOI] [PubMed] [Google Scholar]
  42. Stam CJ, Van Woerkom TCAM, Keunen RWM. Non-linear analysis of the electroencephalogram in Creutzfeldt-Jakob disease. Biological Cybernetics. 1997;77:247–256. doi: 10.1007/s004220050385. [DOI] [PubMed] [Google Scholar]
  43. Steiger A, Antonijevic IA, Bohlhalter S, Frieboes RM, Friess E, Murck H. Effects of hormones on sleep. Hormone Research. 1998;49:125–130. doi: 10.1159/000023158. [DOI] [PubMed] [Google Scholar]
  44. Steriade M, McCarley RW. Brainstem Control of Sleep and Wakefulness. New York: Plenum; 1990. [Google Scholar]
  45. Szeto HH. Spectral edge frequency as a simple quantitative measure of the maturation of electrocortical activity. Pediatric Research. 1990;27:289–292. doi: 10.1203/00006450-199003000-00018. [DOI] [PubMed] [Google Scholar]
  46. Takens F. Detecting strange attractors in turbulence. In: Rand D, Young LS, editors. Dynamical Systems in Turbulence. New York: Springer; 1981. pp. 366–381. [Google Scholar]
  47. Theiler J, Rapp PE. Re-examination of the evidence for low-dimensional, nonlinear structure in the human electroencephalogram. Electroencephalography and Clinical Neurophysiology. 1996;98:213–222. doi: 10.1016/0013-4694(95)00240-5. [DOI] [PubMed] [Google Scholar]
  48. Toth LA, Gardiner TW, Krueger JM. Modulation of sleep by cortisone in normal and bacterially infected rabbits. American Journal of Physiology. 1992;263:R1339–1346. doi: 10.1152/ajpregu.1992.263.6.R1339. [DOI] [PubMed] [Google Scholar]
  49. Trojaborg W, Boysen G. Relation between EEG, rCBF and internal carotid artery pressure during carotid endarteryectomy. Electroencephalography and Clinical Neurophysiology. 1973;34:61–69. doi: 10.1016/0013-4694(73)90151-x. [DOI] [PubMed] [Google Scholar]
  50. Yang K, Jones SA, Challis JR. Changes in glucocorticoid receptor number in the hypothalamus and pituitary of the sheep fetus with gestational age and after adrenocorticotropin treatment. Endocrinology. 1990;126:11–17. doi: 10.1210/endo-126-1-11. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES