e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 7 8 0 e7 8 9
Official Journal of the European Paediatric Neurology Society
Original article
Functional connectivity in preterm infants derived from EEG coherence analysis E.J. Meijer a, K.H.M. Hermans a,b, A. Zwanenburg c, W. Jennekens d, H.J. Niemarkt e, P.J.M. Cluitmans f, C. van Pul a, P.F.F. Wijn a,b, P. Andriessen e,g,* xima Medical Center, Veldhoven, The Netherlands Clinical Physics, Ma Applied Physics, Eindhoven University of Technology, Eindhoven, The Netherlands c Biomedical Engineering, Maastricht University, Maastricht, The Netherlands d Medical Physics and Technology, Maasstad Hospital, Rotterdam, The Netherlands e xima Medical Center, Veldhoven, The Netherlands Neonatal Intensive Care Unit, Ma f Medical Electrical Engineering, Eindhoven University of Technology, The Netherlands g Faculty of Health, Medicine and Life Science, University of Maastricht, Maastricht, The Netherlands a
b
article info
abstract
Article history:
Objective: To quantify the neuronal connectivity in preterm infants between homologous
Received 26 May 2014
channels of both hemispheres.
Received in revised form
Methods: EEG coherence analysis was performed on serial EEG recordings collected from
12 August 2014
preterm infants with normal neurological follow-up. The coherence spectrum was divided
Accepted 16 August 2014
in frequency bands: dnewborn(0e2 Hz), qnewborn(2e6 Hz), anewborn(6e13 Hz), bnewborn(13 e30 Hz). Coherence values were evaluated as a function of gestational age (GA) and post-
Keywords:
natal maturation.
Preterm infant
Results: All spectra show two clear peaks in the dnewborn and qnewborn-band, corresponding to
EEG
the delta and theta EEG waves observed in preterm infants. In the dnewborn-band the peak
Coherence
magnitude coherence decreases with GA and postnatal maturation for all channels. In the
Maturation
qnewborn-band, the peak magnitude coherence decreases with GA for all channels, but increases
Computer-assisted signal
with postnatal maturation for the frontal polar channels. In the bnewborn-band a modest
processing
magnitude coherence peak was observed in the occipital channels, which decreases with GA. Conclusions: Interhemispherical connectivity develops analogously with electrocortical maturation: signal intensities at low frequencies decrease with GA and postnatal maturation, but increase at high frequencies with postnatal maturation. In addition, peak magnitude coherence is a clear trend indicator for brain maturation. Significance: Coherence analysis can aid in the clinical assessment of the functional connectivity of the infant brain with maturation. © 2014 European Paediatric Neurology Society. Published by Elsevier Ltd. All rights reserved.
xima Medical Center, Neonatal Intensive Care Unit, P.O. Box 7777, 5500 MB Veldhoven, The Netherlands. Tel.: * Corresponding author. Ma þ31408889350; fax: þ31408889340. E-mail addresses: [email protected] (E.J. Meijer), [email protected] (P. Andriessen). http://dx.doi.org/10.1016/j.ejpn.2014.08.003 1090-3798/© 2014 European Paediatric Neurology Society. Published by Elsevier Ltd. All rights reserved.
e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 7 8 0 e7 8 9
1.
Introduction
Computer assisted analyses of neonatal EEG may be used as a physiologic biomarker of developmental neural plasticity1 and neurological outcome.2e5 Frequency- and timedependent signal processing strategies of EEG recordings provide a methodology to investigate neuronal network maturation in terms of differentiation and integration of short-distance versus long-distance neuronal connections throughout the brain.1,6 An important long-distance neuronal connection e linking associative areas of both cerebral hemispheres and facilitating functional interhemispheric connectivity e is the corpus callosum (CC). In the perinatal period the development of the corpus callosum is mainly governed by the development and increase of the number of axons and not by myelination.7e10 Kostovic et al. and Fransson et al. demonstrated that both cortico-cortical and thalamocortical connections are formed during the second half of gestation,11,12 and it is well established that all of these connections play an important role in synchrony and interhemispheric connections.13 In several studies the association between functional connectivity and the development of the corpus callosum has been made. A significant correlation between the size of the neonatal CC in Magnetic Resonance Imaging (MRI) and the motor outcome at the age of 9e10 years has been demonstrated.14 In preterm infants at term age15,16 diffusion tensor imaging and fiber tracking have been used to quantify maturation by using the diffusion properties and fiber directions in the CC, showing a correlation between anisotropy and white matter injury and with outcome. In most cases MRI techniques cannot be used immediately after birth, since the newborn needs to be stabilized and sedated before undergoing an MRI exam, and this technique does not allow for continuous monitoring and is prone to artifacts. Functional connectivity can be investigated using fMRI, however Omidvarnia et al.17 suggest that essential dynamics of early network activity in the human brain may not be directly reflected in resting-state functional MRI signals. To provide early identification of compromised connectivity, we therefore choose to study quantitative EEG techniques. EEG coherence and burst synchrony detection are potential candidates for bed-side monitoring of functional interhemispheric connectivity in preterm infants.18,19 Preterm infants characteristically show a very discontinuous EEG pattern, consisting of short periods of high amplitude activity (bursts), alternated by longer periods of relative quiescence (interburst intervals).10 Grieve et al. compared prematurely born, extremely low-birth weight infants at corresponding term age with full-term controls. Compared to term infants, low-birth weight infants manifested regional differences in EEG functional connectivity at term age, with significantly reduced interhemispheric coherences in the frontal polar and parietooccipital regions.19 Kuks et al. demonstrated a significantly reduced interhemispheric coherence in full-term infants with agenesis of the CC.20 Gonzalez et al. showed that functional brain connectivity in term and preterm neonates presents regional differences that depend on maturation.21 Koolen et al. demonstrated that features can be extracted from the
781
EEG to distinguish between physiological and pathological asymmetry in the developing neonatal brain.22 Hence, EEG coherence may reflect functional brain connectivity and may be of predictive value for adverse neurological outcome in later life. Before using coherence parameters as potential outcome predictor, it is essential to investigate maturational trends for preterm infants with normal outcome, and to define the parameters that can be monitored.23 Therefore, the goal of this study was to investigate maturational changes in interhemispheric EEG functional connectivity in preterm infants with normal outcome at five years of age, using coherence analysis between homologous EEG channels. In contrast to other studies, we investigate the effect of both gestational age and postnatal age to study the potential difference in coherence caused by ex utero versus in utero brain maturation.24 We hypothesize that the coherence increases with gestational age and also with postnatal age due to increased development of anatomical brain connectivity with maturation.
2.
Methods
2.1.
Study subjects
After approval of the hospital's ethics committee, 32 preterm infants were recruited from the neonatal intensive care unit xima Medical Center in Veldhoven, the (NICU) of the Ma Netherlands, between May of 2006 and July of 2007. Infants were enrolled after written informed consent from both parents. All infants had a gestational age (GA) less than 37 weeks and showed a normal neurological follow-up after two and five years. The mean GA ± SD for the group was 30.3 ± 2.9 weeks. During the recordings patients did not receive medication that is known to influence the EEG background pattern, except caffeine. For 18 infants 3 or more repeated EEGs with a weekly interval were recorded. For this group the mean GA was 28.8 ± 1.3 weeks. For further details of the study group we refer to previous work.25e27
2.2.
Data acquisition
Starting at the end of the first week of life, digital EEG recordings (NicoletOne; Viasys Healthcare, Conshohocken, PA, USA) were acquired. After skin preparation (Nuprep Gel, D.O. Weaver, Aurora, CO, USA) Ag/AgCl cup electrodes, filled with a conductive paste (Ten20, D.O. Weaver, Aurora, CO, USA) were placed on the scalp according the international reduced 10-20 montage system.28 Impedance of the electrodes was maintained below 10 kU. The digital EEG signal was sampled at 256 Hz. Artifacts related to repositioning the infant or replacement of an electrode when impedance was above 10 kU were removed (
Official Journal of the European Paediatric Neurology Society
Original article
Functional connectivity in preterm infants derived from EEG coherence analysis E.J. Meijer a, K.H.M. Hermans a,b, A. Zwanenburg c, W. Jennekens d, H.J. Niemarkt e, P.J.M. Cluitmans f, C. van Pul a, P.F.F. Wijn a,b, P. Andriessen e,g,* xima Medical Center, Veldhoven, The Netherlands Clinical Physics, Ma Applied Physics, Eindhoven University of Technology, Eindhoven, The Netherlands c Biomedical Engineering, Maastricht University, Maastricht, The Netherlands d Medical Physics and Technology, Maasstad Hospital, Rotterdam, The Netherlands e xima Medical Center, Veldhoven, The Netherlands Neonatal Intensive Care Unit, Ma f Medical Electrical Engineering, Eindhoven University of Technology, The Netherlands g Faculty of Health, Medicine and Life Science, University of Maastricht, Maastricht, The Netherlands a
b
article info
abstract
Article history:
Objective: To quantify the neuronal connectivity in preterm infants between homologous
Received 26 May 2014
channels of both hemispheres.
Received in revised form
Methods: EEG coherence analysis was performed on serial EEG recordings collected from
12 August 2014
preterm infants with normal neurological follow-up. The coherence spectrum was divided
Accepted 16 August 2014
in frequency bands: dnewborn(0e2 Hz), qnewborn(2e6 Hz), anewborn(6e13 Hz), bnewborn(13 e30 Hz). Coherence values were evaluated as a function of gestational age (GA) and post-
Keywords:
natal maturation.
Preterm infant
Results: All spectra show two clear peaks in the dnewborn and qnewborn-band, corresponding to
EEG
the delta and theta EEG waves observed in preterm infants. In the dnewborn-band the peak
Coherence
magnitude coherence decreases with GA and postnatal maturation for all channels. In the
Maturation
qnewborn-band, the peak magnitude coherence decreases with GA for all channels, but increases
Computer-assisted signal
with postnatal maturation for the frontal polar channels. In the bnewborn-band a modest
processing
magnitude coherence peak was observed in the occipital channels, which decreases with GA. Conclusions: Interhemispherical connectivity develops analogously with electrocortical maturation: signal intensities at low frequencies decrease with GA and postnatal maturation, but increase at high frequencies with postnatal maturation. In addition, peak magnitude coherence is a clear trend indicator for brain maturation. Significance: Coherence analysis can aid in the clinical assessment of the functional connectivity of the infant brain with maturation. © 2014 European Paediatric Neurology Society. Published by Elsevier Ltd. All rights reserved.
xima Medical Center, Neonatal Intensive Care Unit, P.O. Box 7777, 5500 MB Veldhoven, The Netherlands. Tel.: * Corresponding author. Ma þ31408889350; fax: þ31408889340. E-mail addresses: [email protected] (E.J. Meijer), [email protected] (P. Andriessen). http://dx.doi.org/10.1016/j.ejpn.2014.08.003 1090-3798/© 2014 European Paediatric Neurology Society. Published by Elsevier Ltd. All rights reserved.
e u r o p e a n j o u r n a l o f p a e d i a t r i c n e u r o l o g y 1 8 ( 2 0 1 4 ) 7 8 0 e7 8 9
1.
Introduction
Computer assisted analyses of neonatal EEG may be used as a physiologic biomarker of developmental neural plasticity1 and neurological outcome.2e5 Frequency- and timedependent signal processing strategies of EEG recordings provide a methodology to investigate neuronal network maturation in terms of differentiation and integration of short-distance versus long-distance neuronal connections throughout the brain.1,6 An important long-distance neuronal connection e linking associative areas of both cerebral hemispheres and facilitating functional interhemispheric connectivity e is the corpus callosum (CC). In the perinatal period the development of the corpus callosum is mainly governed by the development and increase of the number of axons and not by myelination.7e10 Kostovic et al. and Fransson et al. demonstrated that both cortico-cortical and thalamocortical connections are formed during the second half of gestation,11,12 and it is well established that all of these connections play an important role in synchrony and interhemispheric connections.13 In several studies the association between functional connectivity and the development of the corpus callosum has been made. A significant correlation between the size of the neonatal CC in Magnetic Resonance Imaging (MRI) and the motor outcome at the age of 9e10 years has been demonstrated.14 In preterm infants at term age15,16 diffusion tensor imaging and fiber tracking have been used to quantify maturation by using the diffusion properties and fiber directions in the CC, showing a correlation between anisotropy and white matter injury and with outcome. In most cases MRI techniques cannot be used immediately after birth, since the newborn needs to be stabilized and sedated before undergoing an MRI exam, and this technique does not allow for continuous monitoring and is prone to artifacts. Functional connectivity can be investigated using fMRI, however Omidvarnia et al.17 suggest that essential dynamics of early network activity in the human brain may not be directly reflected in resting-state functional MRI signals. To provide early identification of compromised connectivity, we therefore choose to study quantitative EEG techniques. EEG coherence and burst synchrony detection are potential candidates for bed-side monitoring of functional interhemispheric connectivity in preterm infants.18,19 Preterm infants characteristically show a very discontinuous EEG pattern, consisting of short periods of high amplitude activity (bursts), alternated by longer periods of relative quiescence (interburst intervals).10 Grieve et al. compared prematurely born, extremely low-birth weight infants at corresponding term age with full-term controls. Compared to term infants, low-birth weight infants manifested regional differences in EEG functional connectivity at term age, with significantly reduced interhemispheric coherences in the frontal polar and parietooccipital regions.19 Kuks et al. demonstrated a significantly reduced interhemispheric coherence in full-term infants with agenesis of the CC.20 Gonzalez et al. showed that functional brain connectivity in term and preterm neonates presents regional differences that depend on maturation.21 Koolen et al. demonstrated that features can be extracted from the
781
EEG to distinguish between physiological and pathological asymmetry in the developing neonatal brain.22 Hence, EEG coherence may reflect functional brain connectivity and may be of predictive value for adverse neurological outcome in later life. Before using coherence parameters as potential outcome predictor, it is essential to investigate maturational trends for preterm infants with normal outcome, and to define the parameters that can be monitored.23 Therefore, the goal of this study was to investigate maturational changes in interhemispheric EEG functional connectivity in preterm infants with normal outcome at five years of age, using coherence analysis between homologous EEG channels. In contrast to other studies, we investigate the effect of both gestational age and postnatal age to study the potential difference in coherence caused by ex utero versus in utero brain maturation.24 We hypothesize that the coherence increases with gestational age and also with postnatal age due to increased development of anatomical brain connectivity with maturation.
2.
Methods
2.1.
Study subjects
After approval of the hospital's ethics committee, 32 preterm infants were recruited from the neonatal intensive care unit xima Medical Center in Veldhoven, the (NICU) of the Ma Netherlands, between May of 2006 and July of 2007. Infants were enrolled after written informed consent from both parents. All infants had a gestational age (GA) less than 37 weeks and showed a normal neurological follow-up after two and five years. The mean GA ± SD for the group was 30.3 ± 2.9 weeks. During the recordings patients did not receive medication that is known to influence the EEG background pattern, except caffeine. For 18 infants 3 or more repeated EEGs with a weekly interval were recorded. For this group the mean GA was 28.8 ± 1.3 weeks. For further details of the study group we refer to previous work.25e27
2.2.
Data acquisition
Starting at the end of the first week of life, digital EEG recordings (NicoletOne; Viasys Healthcare, Conshohocken, PA, USA) were acquired. After skin preparation (Nuprep Gel, D.O. Weaver, Aurora, CO, USA) Ag/AgCl cup electrodes, filled with a conductive paste (Ten20, D.O. Weaver, Aurora, CO, USA) were placed on the scalp according the international reduced 10-20 montage system.28 Impedance of the electrodes was maintained below 10 kU. The digital EEG signal was sampled at 256 Hz. Artifacts related to repositioning the infant or replacement of an electrode when impedance was above 10 kU were removed (
