Development of neuronal connections relies on proper neuronal activity, and it starts during the time when early preterm babies are treated in the neonatal intensive care units. While synchrony has been a key element in visual assessment of neonatal EEG signals, there has been no unambiguous definitions for synchrony, and no objective measures available for neonatal signals. Estimation of phase locking value (PLV) has been an established paradigm in adults, but many unique characteristics of the neonatal EEG have precluded its applicability in them. In the present paper, we developed the existing PLV-based methods further to be applicable for neonatal signals at two different temporal scales, oscillations and events, where the latter refers technically to quantitating phase synchrony (PS) between band-specific amplitude envelopes (bafPS). In addition, we present a measure for quantitation based on assessing cumulative proportion of time with statistically significant synchrony between the given signal pair. The paper uses real EEG examples and the prior neurobiological knowledge in the process of defining optimal parameters in each step of the procedure. Finally, we apply the method to a set of dense array EEG recordings from very early preterm babies, recorded at conceptional age of less than 30 weeks. By comparing PS and bafPS from babies without and with major cerebrovascular lesion, we show that the effects of brain lesions may be selective both in space and in frequency. These findings do by nature escape visual detection in the conventional EEG reading, however they have intriguing correlates in the current concept of how somatosensory networks are thought to develop and/or become disorganized in the early preterm babies.
Copyright © 2012 Elsevier Inc. All rights reserved.