Ma Xueling, Wu Xiushuang, Shi Yuan
Department of Neonatalogy, Children's Hospital of Chongqing Medical University, Chongqing, China.
National Clinical Research Center for Child Health and Disorders, Chongqing, China.
Front Pediatr. 2020 Jul 23;8:412. doi: 10.3389/fped.2020.00412. eCollection 2020.
To investigate the changes of dynamic functional connectivity (DFC) in late preterm infants, and assess whether these changes are associated with the indicators measuring the maturity of neonates. Resting-state fMRI (rs-fMRI) data of eligible neonates was acquired with a 3.0-T MRI scanner in the Department of Radiology, Daping Hospital, Army Medical University (Chongqing, China). After the selection of functional connectivity networks obtained by independent component analysis (ICA), a sliding-window approach was used to cluster all the windows into different states. Then the differences of temporal properties of DFC between groups were compared, and the association between these temporal properties and the degree of maturity was also explored in each state. Eventually, 34 late preterm and 37 term neonates were included in the final analysis. Based on their data, 5 components were located in 5 networks: default-mode (DMN), dorsal attention (DAN), auditory (AUD), sensorimotor (SMN), and visual (VN). Then four reoccurring state patterns of functional connectivity were identified with the k-means clustering method. The late preterm group dwelled significantly longer in State III (late preterm: 33.57 ± 37.64 s, term: 18.50 ± 11.71 s; = 0.03), which was characterized by general weaker connectivity between networks. Also, the correlation analysis shows the degree of maturity is negatively correlated to the dwell time and fractional windows in State III. Our findings suggested that compared with term infants, late preterm infants preferred to stay in a state with general weak connectivity between networks, but this preference declined as maturity increased.
为了研究晚期早产儿动态功能连接性（DFC）的变化，并评估这些变化是否与测量新生儿成熟度的指标相关。在陆军军医大学（中国重庆）大坪医院放射科，使用3.0-T MRI扫描仪获取符合条件的新生儿静息态功能磁共振成像（rs-fMRI）数据。在选择通过独立成分分析（ICA）获得的功能连接网络后，采用滑动窗口方法将所有窗口聚类为不同状态。然后比较组间DFC时间特性的差异，并在每个状态下探索这些时间特性与成熟度之间的关联。最终，34名晚期早产儿和37名足月儿被纳入最终分析。基于他们的数据，5个成分位于5个网络中：默认模式（DMN）、背侧注意（DAN）、听觉（AUD）、感觉运动（SMN）和视觉（VN）。然后用k均值聚类方法识别出四种反复出现的功能连接状态模式。晚期早产组在状态III中的停留时间明显更长（晚期早产：33.57±37.64秒，足月儿：18.50±11.71秒；P = 0.03），其特征是网络之间的连接普遍较弱。此外，相关性分析表明，成熟度与状态III中的停留时间和分数窗口呈负相关。我们的研究结果表明，与足月儿相比，晚期早产儿更倾向于停留在网络之间连接普遍较弱的状态，但随着成熟度的增加，这种偏好会下降。
