The Clinical Hospital of Chengdu Brain Science Institute, MOE Key Lab for Neuroinformation, School of Life Science and Technology, University of Electronic Science and Technology of China, Chengdu, 610054, China.
Department of Medical Technology, Cangzhou Medical College, Cangzhou, 061001, PR China.
Neuropsychologia. 2021 Jul 16;157:107865. doi: 10.1016/j.neuropsychologia.2021.107865. Epub 2021 Apr 21.
Conduct disorder (CD) is a psychiatric condition characterized by severe aggressive and antisocial behaviors. Prior neuroimaging work reported that CD is associated with abnormal resting-state local intrinsic brain activity (IBA). However, few studies detected the time-varying brain activity patterns in CD. In this study, eighteen adolescent patients with CD and 18 typically developing controls underwent resting-state functional magnetic resonance imaging scans. We then compared the dynamic characteristics of IBA by calculating the dynamic regional homogeneity (dReHo) through a sliding-window approach between the two groups, and the correlations between the dReHo variability and clinical symptoms in CD were further examined. Moreover, the statistical between-group differences in dReHo were selected as classification features to help distinguish CD patients from controls by adopting a linear support vector machine (SVM) classifier. CD patients showed increased dReHo variability in the left precuneus, right postcentral gyrus, right precentral gyrus, left middle cingulate gyrus, and left paracentral lobule compared to controls, and dReHo variability in the left precuneus was significantly positively associated with impulsiveness scores in CD patients. Importantly, SVM combined with the leave-one-out cross-validation method results demonstrated that 75% (p < 0.001) subjects were correctly classified with sensitivity of 61% and specificity of 89%. Our results provided the initial evidence that CD is characterized by abnormal dynamic IBA patterns, giving novel insights into the neuropathological mechanisms of CD. Further, our findings exhibited that the dReHo variability may distinguish CD patients from controls with high accuracy.
品行障碍（CD）是一种以严重攻击性行为和反社会行为为特征的精神疾病。先前的神经影像学研究报告称，CD 与静息状态下大脑局部固有脑活动（IBA）异常有关。然而，很少有研究检测到 CD 患者的时变脑活动模式。在这项研究中，18 名青少年 CD 患者和 18 名正常发育对照者接受了静息态功能磁共振成像扫描。然后，我们通过滑动窗口方法计算动态区域同质性（dReHo），比较两组之间 IBA 的动态特征，并进一步研究 CD 中 dReHo 可变性与临床症状之间的相关性。此外，将统计上两组间 dReHo 的差异作为分类特征，通过采用线性支持向量机（SVM）分类器来帮助区分 CD 患者和对照者。与对照组相比，CD 患者的左侧楔前叶、右侧中央后回、右侧中央前回、左侧扣带回中部和左侧旁中央小叶的 dReHo 可变性增加，并且 CD 患者的左侧楔前叶的 dReHo 可变性与冲动性评分呈显著正相关。重要的是，SVM 结合留一法交叉验证方法的结果表明，75%（p<0.001）的受试者被正确分类，敏感性为 61%，特异性为 89%。我们的研究结果提供了初步证据，表明 CD 的特征是异常的动态 IBA 模式，为 CD 的神经病理学机制提供了新的见解。此外，我们的发现表明 dReHo 可变性可以以很高的准确率区分 CD 患者和对照者。
