7 - EEG-Derived Functional Connectivity as a Biomarker of HIE Severity in Neonates

Publication Number: 1006.7

Soheila Norasteh, University of Texas at Arlington, DALLAS, TX, United States; Srinivas Kota, University of Texas Southwestern Medical School, Dallas, TX, United States; Yu-Lun Liu, University of Texas Southwestern Medical School, Dallas, TX, United States; Rong Zhang, UTSW Medical Center, Dallas, TX, United States; Hanli Liu, University of Texas at Arlington, Arlington, TX, United States; Lina Chalak, Childrens Health Dallas UT Southwestern, DALLAS, TX, United States

Background: Hypoxic-ischemic encephalopathy (HIE) is a leading cause of neonatal death and long-term neurological disability. Although therapeutic hypothermia offers neuroprotection, it must begin within six hours after birth, making early identification critical. The evolving nature of injury and variability in clinical assessment highlight the need for objective bedside biomarkers to enable timely and personalized treatment.

Objective: This study tested the hypothesis that EEG-derived functional connectivity can serve as an objective marker for early and accurate assessment of HIE severity. Specifically, mean phase coherence (MPC) reflecting scalp-wide EEG synchronization was analyzed to reveal progressive alterations in brain network organization across control, mild, and moderate-to-severe HIE groups.

Design/Methods: Two hours (2-h) of artifact-free EEG data were obtained from eight scalp electrodes of (Fz, C3, Cz, C4, P3, P4, O1, O2) by preprocessing and then segmenting into 10-second (10-s) epochs. MPC was computed based on phase differences between every pair of electrodes within the 1-2 Hz frequency band. MPCs were computed between all electrode pairs to form an 8×8 connectivity matrix for all available segments for each neonate. Three global connectivity metrics (i.e., global efficiency, mean strength, and mean clustering coefficient) were derived from each matrix and averaged across all segments per neonate. Group differences in each of these metrics among controls, mild, and moderate-to-severe HIE were assessed using the Wilcoxon rank-sum test with Holm-Bonferroni correction.

Results: Connectivity analysis included 59 full-term newborns (8 controls, 27 mild, 24 moderate-to-severe HIE). The moderate-to-severe group exhibited significant higher values than the other two groups in all three global network metrics (global efficiency, mean strength, and mean clustering coefficient). Newborns with moderate HIE showed increased cortical synchronization, indicating functional hyperconnectivity associated with brain injury (Figure 1). This finding aligns with prior studies reporting elevated local network metrics in HIE neonates and may reflect a compensatory mechanism, whereby intact brain regions enhance communication to preserve essential functions.

Conclusion(s): Our findings support that EEG-based connectivity metrics effectively capture alterations in brain network organization linked to HIE severity. These results underscore the potential of EEG connectivity as a noninvasive and quantitative biomarker for early, objective assessment of HIE for timely neuroprotective interventions.

Table 1. Maternal and Neonatal Characteristics


Figure 1
Group comparisons of EEG-derived connectivity metrics across control, mild, and moderate-to-severe HIE groups using the Wilcoxon rank-sum test with Holm-Bonferroni correction. (a) Comparison of global efficiency, with p = 0.038 for moderate-to-severe versus both mild and control groups. (b) Comparison of mean strength, with p = 0.020 and p = 0.026 for moderate-to-severe versus mild and control groups, respectively. (c) Comparison of mean clustering coefficient, with p = 0.023 and p = 0.041 for moderate-to-severe versus mild and control groups, respectively.