Abstracts

Rationale: The human brain undergoes significant age-specific anatomical and physiological changes over the course of early development. Functional connectivity network analysis provides a tractable, activity-based measure to identify dynamic relationships between developing neuronal populations (J Child Neurol, 2011;26(4):488-500). Over the last decade, these approaches have been used extensively to identify abnormalities in normal and diseased populations (J Neurosci, 2010;30(30):10076-85) . Here we characterize the developing functional brain networks in infancy and childhood using EEG. This work provides a foundation upon which to better understand normal brain development and ultimately, the impact of abnormal neurophysiology on these developing networks. Methods: We evaluated sleep rhythms and functional network connectivity patterns in 334 neuro-typical infants and children (age 0-10 years). 19-channel EEGs were manually reviewed, artifacts rejected, and state-specific epochs >100sec were selected for analysis. Data were referenced to a bipolar and physical reference (C2) and filtered for broad band analysis (1-55 Hz). Spectra were computed using the multitaper method allowing a frequency resolution of 1 Hz. To infer functional networks, prepared EEG data were divided into discrete 1sec windows for coupling analysis. The maximal cross-correlation between all electrode pairs was calculated, allowing a lag up to 100ms. To identify edges, we applied an analytic significance testing procedure to identify network edges with statistical confidence and a linear step-up false detection rate controlling procedure to correct for multiple comparisons (Phys Rev E, 2009;79(6 Pt 1):061916). To minimize the impact of volume conduction, edges identified with maximum correlation at zero phase lag were ignored (detailed methods in: J Neurosci 2012: 32(8)2703-13). For each window, significant cross-correlation values were represented as a network in the form of an undirected binary adjacency matrix. Graph metrics and similarity measures were then computed for each age-group and compared between groups. Results: We show that the topology of functional networks changes significantly over infancy in a stereotyped fashion. These changes are most marked in the first few months of life and plateau by childhood. Conclusions: This work demonstrates that cortical functional connectivity networks change dramatically over early development, in a manner consistent with the overall arc of neurodevelopment. These changes are evident within routine EEG recordings across a time span of months and years. Functional network connectivity analysis may provide a sensitive, age-specific clinical tool to track and better understand the pathophysiological impact of altered neurophysiology during critical periods of neurodevelopment.