Nelson Brent G, Bassett Danielle S, Camchong Jazmin, Bullmore Edward T, Lim Kelvin O
Department of Psychiatry, University of Minnesota, Minneapolis, MN 55454, USA.
Department of Physics, University of California, Santa Barbara, CA 93106, USA; Sage Center for the Study of the Mind, University of California, Santa Barbara, CA 93106, USA; Department of Bioengineering, University of Pennsylvania, Philadelphia, PA 19104, USA.
Neuroimage Clin. 2017 May 14;15:439-448. doi: 10.1016/j.nicl.2017.05.007. eCollection 2017.
Schizophrenia is a disease with disruptions in thought, emotion, and behavior. The dysconnectivity hypothesis suggests these disruptions are due to aberrant brain connectivity. Many studies have identified connectivity differences but few have been able to unify gray and white matter findings into one model. Here we develop an extension of the Network-Based Statistic (NBS) called NBSm (Multimodal Network-based statistic) to compare functional and anatomical networks in schizophrenia. Structural, resting functional, and diffusion magnetic resonance imaging data were collected from 29 chronic patients with schizophrenia and 29 healthy controls. Images were preprocessed, and average time courses were extracted for 90 regions of interest (ROI). Functional connectivity matrices were estimated by pairwise correlations between wavelet coefficients of ROI time series. Following diffusion tractography, anatomical connectivity matrices were estimated by white matter streamline counts between each pair of ROIs. Global and regional strength were calculated for each modality. NBSm was used to find significant overlap between functional and anatomical components that distinguished health from schizophrenia. Global strength was decreased in patients in both functional and anatomical networks. Regional strength was decreased in all regions in functional networks and only one region in anatomical networks. NBSm identified a distinguishing functional component consisting of 46 nodes with 113 links (p < 0.001), a distinguishing anatomical component with 47 nodes and 50 links (p = 0.002), and a distinguishing intermodal component with 26 nodes (p < 0.001). NBSm is a powerful technique for understanding network-based group differences present in both anatomical and functional data. In light of the dysconnectivity hypothesis, these results provide compelling evidence for the presence of significant overlapping anatomical and functional disruption in people with schizophrenia.
精神分裂症是一种思维、情感和行为紊乱的疾病。失连接假说认为，这些紊乱是由于大脑连接异常所致。许多研究已经确定了连接差异，但很少有研究能够将灰质和白质的研究结果统一到一个模型中。在这里，我们开发了一种基于网络的统计方法（NBS）的扩展方法，称为NBSm（多模态基于网络的统计方法），以比较精神分裂症患者的功能和解剖网络。我们收集了29名慢性精神分裂症患者和29名健康对照者的结构、静息功能和扩散磁共振成像数据。对图像进行预处理，并提取90个感兴趣区域（ROI）的平均时间序列。通过ROI时间序列的小波系数之间的成对相关性估计功能连接矩阵。在进行扩散束描记后，通过每对ROI之间的白质流线计数估计解剖连接矩阵。计算每种模态的全局和区域强度。NBSm用于寻找区分健康人和精神分裂症患者的功能和解剖成分之间的显著重叠。患者在功能和解剖网络中的全局强度均降低。功能网络中所有区域的区域强度均降低，而解剖网络中只有一个区域的区域强度降低。NBSm识别出一个由46个节点和113条链接组成的区分性功能成分（p < 0.001）、一个由47个节点和50条链接组成的区分性解剖成分（p = 0.002）以及一个由26个节点组成的区分性跨模态成分（p < 0.001）。NBSm是一种强大的技术，可用于理解解剖和功能数据中基于网络的组间差异。根据失连接假说，这些结果为精神分裂症患者存在显著的解剖和功能重叠性破坏提供了有力证据。
