Shan Yuanyuan, Zhang Lin, Zhang Peng, Xu Yan, Wang Jun, Yang Mingshu, Cheng Guoqiang
Department of Pediatric Intensive Care Unit, National Children's Medical Center, Children's Hospital of Fudan University, Shanghai, China.
Department of Radiology, Children's Hospital of Fudan University Xiamen Branch, Xiamen, China.
Sci Rep. 2025 Jul 9;15(1):24620. doi: 10.1038/s41598-025-10127-6.
The objective of this study was to identify relevant quantitative parameters to distinguish premature infants with presence of brain injury from conventional electroencephalography (EEG) and predict short-term neuromotor developmental outcomes. This is a prospective cohort study of newborns at 34 weeks' gestation or earlier. Multichannel EEG recordings were performed within 24 h after birth. The total power (TP), absolute and relative band power (ABP and RBP), alpha/theta ratio (ATR), alpha/delta + theta ratio (ADTR), 95% spectral edge frequency (SEF), approximate entropy (ApEn), coherence and brain symmetry index (BSI) were calculated using the Auto-Neo-EEG signal processing system. Neonates were divided into two groups: with and without brain injury, and clinical outcomes of general movements (GMs) assessment at three months were available for analysis. This study comprised 43 and 65 premature neonates with and without brain injury, respectively. Premature neonates with brain injury had significantly lower TP, ABP-δ, ABP-α, RBP-δ and coherence than those without brain injury (all p values < 0.05). The area under curve (AUC) of TP, ABP-δ, ABP-α, RBP-δ and coherence for predicting brain injury was 0.749, 0.830, 0.721, 0.799 and 0.743, respectively. Preterm infants with brain injury had significantly lower GMs optimality scores (15.6 ± 6.7) than those without brain injury (28.4 ± 8.3) (P = 0.019). For 43 preterm infants with brain injury, TP (P = 0.023) and ABP-δ (P = 0.030) were positively correlated with GMs optimality scores; while coherence (P = 0.039) was the opposite. Compared with those without brain injury, preterm infants with brain injury tended to have reduced spectral power, accompanied by impaired brain network connectivity, and delayed short-term motor development. Automated quantitative EEG (qEEG) analysis provides predictive value for the occurrence of brain injury and outcomes in preterm neonates, among which ABP-δ presented the best predictive performance.
本研究的目的是从常规脑电图（EEG）中识别出相关定量参数，以区分存在脑损伤的早产儿，并预测短期神经运动发育结局。这是一项对妊娠34周或更早的新生儿进行的前瞻性队列研究。出生后24小时内进行多通道EEG记录。使用自动新生儿EEG信号处理系统计算总功率（TP）、绝对和相对频段功率（ABP和RBP）、α/θ比值（ATR）、α/δ+θ比值（ADTR）、95%频谱边缘频率（SEF）、近似熵（ApEn）、相干性和脑对称指数（BSI）。新生儿分为两组：有脑损伤组和无脑损伤组，并对三个月时的全身运动（GMs）评估的临床结局进行分析。本研究分别纳入了43例和65例有脑损伤和无脑损伤的早产儿。有脑损伤的早产儿的TP、ABP-δ、ABP-α、RBP-δ和相干性显著低于无脑损伤的早产儿（所有p值<0.05）。TP、ABP-δ、ABP-α、RBP-δ和相干性预测脑损伤的曲线下面积（AUC）分别为0.749、0.830、0.721、0.799和0.743。有脑损伤的早产儿的GMs最优性评分（15.6±6.7）显著低于无脑损伤的早产儿（28.4±8.3）（P=0.019）。对于43例有脑损伤的早产儿，TP（P=0.023）和ABP-δ（P=0.030）与GMs最优性评分呈正相关；而相干性（P=0.039）则相反。与无脑损伤的早产儿相比，有脑损伤的早产儿往往频谱功率降低，伴有脑网络连接受损和短期运动发育延迟。自动定量脑电图（qEEG）分析为早产儿脑损伤的发生和结局提供了预测价值，其中ABP-δ的预测性能最佳。
