标题： 自闭症跨脑网络的共同进化平衡状态：生物标志物发现的神经功能框架 显示英文标题

标题： Co-evolutionary Balance State of the Autism inter-Brain Network: A Neurofunctional Framework for Biomarker Discovery

摘要： 自闭症谱系障碍（ASD）是一种神经发育状况，其特征是社交交流缺陷和重复行为；然而，客观的神经生理生物标志物仍然缺乏。 我们提出了一种共进化平衡范式，通过一个哈密顿量来量化网络层面的能量，该哈密顿量整合了通过低频波动分数幅值（fALFF）和静息状态功能连接（FC）测量的区域活动。 对93名患有 ASD 的成年男性和93名匹配对照组的静息状态 fMRI 数据分析显示，经验网络的能量低于1000个保持拓扑结构的零模型（配对 t = -4.12，p 小于等于 1e-4）。 ASD 参与者表现出更负的整体大脑能量（t = -3.239，p = 0.0015），这由增加的一致链接和减少的不平衡同构模式驱动。 子网络分析表明，在假发现率校正后，默认模式网络的能量更大（p 小于 0.016），并且默认模式网络、显著性和背侧注意网络之间的能量增强（p 小于 0.032）。 能量指标和网络间连通性与自闭症诊断访谈修订版和自闭症诊断观察量表的严重程度评分相关（绝对相关性大于等于 0.29，p 小于 0.02）。 使用包括模体比例、全局节点链接对齐、网络间 fALFF 加权和 FC 强度、子网络磁化和成对能量在内的九个主要特征的 k 最近邻分类器在平衡灵敏度和特异性下达到了 79% 的准确率。 这些结果表明，共进化能量能够检测可解释的网络中断，并为 ASD 分类建立了一个稳健的框架。 显示英文摘要

摘要： Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by deficits in social communication and repetitive behaviors; however, objective neurophysiological biomarkers remain lacking. We propose a coevolutionary balance paradigm that quantifies network level energy via a Hamiltonian integrating regional activity measured by fractional amplitude of low frequency fluctuations (fALFF) and resting state functional connectivity (FC). Analysis of resting state fMRI data from 93 adult males with ASD and 93 matched controls revealed that empirical networks showed lower energy than 1000 topology preserving null models (paired t = -4.12, p less than or equal to 1e-4). Participants with ASD exhibited more negative whole brain energy (t = -3.239, p = 0.0015), driven by increased agreement links and reduced imbalanced same motifs. Subnetwork analysis indicated greater energy in the Default Mode Network after false discovery rate correction (p less than 0.016) and enhanced energy between the Default Mode, Salience and Dorsal Attention networks (p less than 0.032). Energy metrics and inter network connectivity correlated with Autism Diagnostic Interview Revised and Autism Diagnostic Observation Schedule severity scores (absolute correlation greater than or equal to 0.29, p less than 0.02). A k nearest neighbors classifier using nine principal features including motif proportions, global node link alignment, inter network fALFF weighted and FC strengths, subnetwork magnetization and pairwise energy achieved an accuracy of 79 percent with balanced sensitivity and specificity. These results demonstrate that coevolutionary energy detects interpretable network disruptions and establishes a robust framework for ASD classification.