Exploration of the relationship between brain function and cortical development in the human preterm period with simultaneous EEG-fMRI

在动物模型中，同步神经元活动的自发爆发在设置早期皮质回路的关键发育过程中发挥了重要作用。实验性的破坏会导致健康皮层组织的永久性丧失，这表明这可能代表了自闭症等病症的一个关键病理过程，自闭症起源于生命早期，与皮层功能障碍有关。尽管在早产儿的头皮脑电图（EEG）上可以看到类似的神经活动爆发，但由于空间特异性的限制，它们与皮层发育没有直接关系。同时获得的EEG和功能性MRI [fMRI]数据克服了这些限制，并且关键地可以与来自其他MRI对比的数据相结合。这可以提供潜在的大脑解剖结构的详细特征[例如，皮质折叠和关键发育区域，如基板]，并允许分析与关键深部脑结构的关系[即，丘脑和小脑]和MR衍生的组织微结构测量[如神经突密度和方向分散指数]这个项目将测试早产儿自发爆发性神经活动定位于发育中的大脑区域的假设。除了50个现有的同步EEG-fMRI新生儿数据集（世界上唯一的数据集）外，还将收集新的数据，以提供不同年龄大脑发育的更深入表征。将使用机器学习分类算法来识别EEG数据中的自发爆发事件，然后将源定位算法和传统的fMRI时间序列分析相结合来识别空间关系。由此产生的空间图可以与其他受试者特异性和人群水平的MRI数据对齐。

In animal models, spontaneous bursts of synchronized neuronal activity play an instructiverole in key developmental processes that set early cortical circuits. Experimental disruptionleads to permanent loss of healthy cortical organization, suggesting this may represent a keypathological process in conditions like autism which originate in early life and are associatedwith cortical dysfunction.Although similar bursts of neural activity are seen on scalp electroencephalography [EEG] inpreterm human infants, they cannot be directly related to cortical development due tolimitations in spatial specificity. Simultaneously acquired EEG and functional MRI [fMRI] dataovercomes these restrictions and crucially can be combined with data from other MRIcontrasts. This can provide detailed characterisation of the underlying brain anatomy [eg,cortical folding and key developmental regions like the subplate], and allows analysis of therelationship with key deep brain structures [ie, thalamus and cerebellum] and MR derivedmeasures of tissue microstructure [such as neurite density and orientation dispersion index].This project will test the hypothesis that spontaneous bursting neural activity localises todeveloping brain regions in preterm infants. In addition to 50 existing simultaneous EEG-fMRIneonatal datasets [the only dataset in the world], new data will be collected to provide adeeper characterisation of brain development at different ages. A machine learningclassification algorithm will be used to characterise spontaneous bursting events within theEEG data and then a combination of source localization algorithms and traditional fMRItimeseries analysis used to identify spatial relationships. Resulting spatial maps can then bealigned to other subject-specific and population level MRI data.