Autism spectrum disorder is a common neurodevelopmental condition that manifests as a disruption in sensory and social skills. Although it has been shown that the brain morphology of individuals with autism is asymmetric, how this differentially affects the structural connectome organization of each hemisphere remains under-investigated. We studied whole-brain structural connectivity-based brain asymmetry in individuals with autism using diffusion magnetic resonance imaging obtained from the Autism Brain Imaging Data Exchange initiative. By leveraging dimensionality reduction techniques, we constructed low-dimensional representations of structural connectivity and calculated their asymmetry index. Comparing the asymmetry index between individuals with autism and neurotypical controls, we found atypical structural connectome asymmetry in the sensory and default-mode regions, particularly showing weaker asymmetry towards the right hemisphere in autism. Network communication provided topological underpinnings by demonstrating that the inferior temporal cortex and limbic and frontoparietal regions showed reduced global network communication efficiency and decreased send-receive network navigation in the inferior temporal and lateral visual cortices in individuals with autism. Finally, supervised machine learning revealed that structural connectome asymmetry could be used as a measure for predicting communication-related autistic symptoms and nonverbal intelligence. Our findings provide insights into macroscale structural connectome alterations in autism and their topological underpinnings.

中文翻译：

---

自闭症患者的全脑结构连接体不对称

自闭症谱系障碍是一种常见的神经发育疾病，表现为感觉和社交技能的破坏。尽管已经证明自闭症患者的大脑形态是不对称的，但这如何对每个半球的结构连接组组织产生不同的影响仍有待研究。我们使用自闭症脑成像数据交换计划获得的扩散磁共振成像研究了自闭症患者基于全脑结构连接的大脑不对称性。通过利用降维技术，我们构建了结构连通性的低维表示并计算了它们的不对称指数。比较自闭症患者和神经典型对照组之间的不对称指数，我们发现感觉和默认模式区域存在非典型的结构连接组不对称性，特别是自闭症患者的右半球表现出较弱的不对称性。网络通信通过证明自闭症患者的下颞叶皮层、边缘和额顶叶区域显示出整体网络通信效率降低以及下颞叶和外侧视觉皮层的发送-接收网络导航减少，从而提供了拓扑基础。最后，监督机器学习表明，结构连接组不对称性可以用作预测与交流相关的自闭症症状和非语言智力的指标。我们的研究结果提供了对自闭症宏观结构连接体改变及其拓扑基础的见解。