Department of Radiology, Affiliated Hospital of Jining Medical University, Jining 272000, Shandong Province, China.
Department of Pharmacy, Affiliated Hospital of Jining Medical University, Jining 272000, Shandong Province, China.
Prog Neuropsychopharmacol Biol Psychiatry. 2024 Aug 30;134:111040. doi: 10.1016/j.pnpbp.2024.111040. Epub 2024 May 26.
Modern neuroimaging methods have revealed that autistic symptoms are associated with abnormalities in brain morphology, connectivity, and activity patterns. However, the changes in brain microstructure underlying the neurobiological and behavioral deficits of autism remain largely unknown.
we characterized the associated abnormalities in intracortical myelination pattern by constructing cortical T1-weighted/T2-weighted ratio maps. Voxel-wise comparisons of cortical myelination were conducted between 150 children with autism spectrum disorder (ASD) and 139 typically developing (TD) children. Group differences in cortical T1-weighted/T2-weighted ratio and gray matter volume were then examined for associations with autistic symptoms. A convolutional neural network (CNN) model was also constructed to examine the utility of these regional abnormalities in cortical myelination for ASD diagnosis.
Compared to TD children, the ASD group exhibited widespread reductions in cortical myelination within regions related to default mode, salience, and executive control networks such as the inferior frontal gyrus, bilateral insula, left fusiform gyrus, bilateral hippocampus, right calcarine sulcus, bilateral precentral, and left posterior cingulate gyrus. Moreover, greater myelination deficits in most of these regions were associated with more severe autistic symptoms. In addition, children with ASD exhibited reduced myelination in regions with greater gray matter volume, including left insula, left cerebellum_4_5, left posterior cingulate gyrus, and right calcarine sulcus. Notably, the CNN model based on brain regions with abnormal myelination demonstrated high diagnostic efficacy for ASD.
Our findings suggest that microstructural abnormalities in myelination contribute to autistic symptoms and so are potentially promising therapeutic targets as well as biomarkers for ASD diagnosis.
现代神经影像学方法揭示，自闭症症状与大脑形态、连接和活动模式的异常有关。然而，自闭症的神经生物学和行为缺陷的脑微观结构变化在很大程度上仍不清楚。
我们通过构建皮质 T1 加权/T2 加权比图来描述皮质内髓鞘形成模式的相关异常。在 150 名自闭症谱系障碍（ASD）儿童和 139 名典型发育（TD）儿童之间进行皮质髓鞘的体素-wise 比较。然后，检查皮质 T1 加权/T2 加权比和灰质体积的组间差异与自闭症症状的相关性。还构建了卷积神经网络（CNN）模型，以检查皮质髓鞘这些区域异常在 ASD 诊断中的应用。
与 TD 儿童相比，ASD 组在与默认模式、突显和执行控制网络相关的区域内表现出广泛的皮质髓鞘减少，例如额下回、双侧岛叶、左侧梭状回、双侧海马、右侧距状裂、双侧中央前回和左后扣带回。此外，这些区域中大多数的髓鞘减少与更严重的自闭症症状有关。此外，ASD 儿童在灰质体积较大的区域表现出髓鞘减少，包括左侧岛叶、左侧小脑_4_5、左侧后扣带回和右侧距状裂。值得注意的是，基于髓鞘异常脑区的 CNN 模型对 ASD 具有较高的诊断效能。
我们的研究结果表明，髓鞘的微观结构异常与自闭症症状有关，因此可能是有希望的治疗靶点和 ASD 诊断的生物标志物。
