Department of Psychiatry, University of North Carolina at Chapel Hill, United States of America.
Department of Psychiatry, University of North Carolina at Chapel Hill, United States of America.
Dev Cogn Neurosci. 2023 Dec;64:101314. doi: 10.1016/j.dcn.2023.101314. Epub 2023 Oct 14.
There is strong evidence that the functional connectome is highly related to the white matter connectome in older children and adults, though little is known about structure-function relationships in early childhood. We investigated the development of cortical structure-function coupling in children longitudinally scanned at 1, 2, 4, and 6 years of age (N = 360) and in a comparison sample of adults (N = 89). We also applied a novel graph convolutional neural network-based deep learning model with a new loss function to better capture inter-subject heterogeneity and predict an individual's functional connectivity from the corresponding structural connectivity. We found regional patterns of structure-function coupling in early childhood that were consistent with adult patterns. In addition, our deep learning model improved the prediction of individual functional connectivity from its structural counterpart compared to existing models.
有强有力的证据表明，功能连接组与大龄儿童和成年人的白质连接组高度相关，尽管对于儿童早期的结构-功能关系知之甚少。我们对在 1、2、4 和 6 岁时进行纵向扫描的儿童（N=360）和成人（N=89）的比较样本进行了皮质结构-功能耦合的研究。我们还应用了一种新的基于图卷积神经网络的深度学习模型和一个新的损失函数，以更好地捕捉个体间的异质性，并从相应的结构连接预测个体的功能连接。我们发现，在儿童早期就存在与成人模式一致的结构-功能耦合的区域模式。此外，与现有的模型相比，我们的深度学习模型提高了从个体的结构连接预测其功能连接的能力。
