健康与遗传心理学研究室知识库

精神分裂症不仅是一种精神疾病，也是一种认知疾病。认知缺陷是精神分裂症的核心特征，精神分裂症患者的认知异质性高。精神分裂症患者的大脑功能异常不仅表现为局部加工异常，还表现为大规模网络间的失调。传统的方法使用静息态功能磁共振或者简单重复的任务刺激来研究精神分裂症患者的脑功能，静息态功能磁共振主要体现被试内源性自发活动的神经信号，而简单重复的任务刺激把刺激局限在一个非常特定而狭小的范围，并不能很好的诱导出精神分裂症患者的大脑活动异常。本系列研究创新性的采用自然刺激范式来研究精神分裂症患者的脑功能异常。自然刺激范式是指给被试呈现复杂动态的视频或者/和听觉刺激，这些刺激模拟真实世界发生的事件，目的是唤起被试在应对真实世界的大脑反应。本系列研究使用各种创新的方法来量化精神分裂症患者在观看自然刺激时的大脑反应并与健康对照进行比较。

本系列研究分为五个部分，分别考察首发精神分裂症患者的大脑局部功能异常、脑岛新的分区方法、基于脑岛的种子点功能连接分析、基于全脑种子点的功能连接分析以及基于全脑种子点的被试间功能相关分析。

研究一考察首发精神分裂症患者的大脑局部功能异常。被试间相关（inter-subject correlation, ISC）是指对于观看视频刺激的不同被试间的同一脑区的时间序列的相关性。ISC 提供一种无模型的方法来量化自然场景感知期间的大脑反应，不需要定义先验的刺激模型。这样的数据驱动分析对于解决诸如社会感知等研究问题是高效的，因为刺激空间过于高维，无法为合适的特征模型生成完整的先验假设集。ISC 能够滤除内源性神经信号和非神经信号的影响，反映刺激唤起的神经信号的作用。结果显示，与健康对照组相比，首发精神分裂症患者在观看视频刺激时，双侧楔前叶、后扣带回、前扣带回、颞顶联合区、楔前回、颞上回、额中回以及脑岛在内的大脑广泛区域的ISC显著降低。视觉皮层高的ISC 反映精神分裂症患者和健康对照一样都认真观看视频刺激。突显网络和默认网络的核心节点出现异常意味着患者在自我和社会加工、记忆提取以及对内外刺激分配注意时在这些核心区域有异常活动的迹象。

基于研究一的结果，研究二和三主要考察脑岛分区以及基于脑岛种子点的功能连接分析。ISC 的研究发现，大面积的双侧脑岛呈现出巨大的组间活动异质性，而右侧脑岛表现出最高的组间功能异质性。结果发现，脑岛在精神分裂症患者和健康对照中都呈现前后脑岛功能分离的趋势，精神分裂症患者的分离程度相对要低一些。基于脑岛种子点的功能连接研究发现，双侧前脑岛和后脑岛与感觉运动皮层、视觉皮层、前扣带皮层、后扣带皮层、背侧前额叶皮层和脑岛本身存在很高的功能连接。组间功能连接统计分析表明，对于左侧的前脑岛和后脑岛，组间的功能连接差异分布在双侧的突显网络；对于右侧的前脑岛和后脑岛，组间的功能连接差异分布在右侧的突显网络。这些结果表明脑岛在处理突显刺激并分配注意力出现异常。

研究四探讨在自然刺激范式下的基于全脑种子点的功能连接。组间功能连接比较后发现，在精神分裂症患者中，默认网络内的节点与扣带回、脑岛、感知觉皮层以及视觉皮层等区域存在更弱的功能连接。感知觉运动皮层节点的功能连接几乎都与扣带回的异常有关，尤其是前扣带回。对于扣带回任务控制网络、听觉网络和记忆提取网络，组间功能连接差异主要集中在右侧的脑岛，也即是突显网络的核心节点之一。对于额顶控制网络节点，组间功能连接差异表现在扣带回、楔前叶以及脑岛等区域。后扣带回和楔前叶是默认网络的核心节点，而脑岛是突显网络的核心节点。这些结果再次验证默认网络、任务控制网络以及突显网络之间的联系和异常，以及三个网络之间的异常的信息交流模式。对于突显网络，组间功能连接表现出大规模的网络间的失调，特别是与默认网络、感知觉运动皮层、任务控制网络之间的功能连接失调。总体上，我们可以得出这样的结果，在自然刺激范式下，精神分裂症患者和健康对照在默认网络、感知觉运动皮层、任务控制网络以及突显网络之间存在连接异常，并且几乎所有的连接异常都指向脑岛的异常。

最后研究五考察基于264 个功能种子点的被试间功能相关 （ISFC, inter-subject functional correlation）组间差异。在更宏观的尺度上来考察由刺激唤起的信号在两组被试间是否存在活动的一致性。和基于全脑种子点的功能连接不同，ISFC 滤除被试内的内源性神经信号的影响，也滤除由扫描仪带来的非神经信号的影响。结果显示，被试间功能相关变异主要集中在精神分裂症患者的感知觉运动网络、扣带回任务控制网络、听觉网络、视觉网络、额顶控制网络、突显网络和皮层下网络的小部分节点与健康对照的默认网络间的活动一致性上有差异。这些结果表明，相对于健康对照的默认网络的活动一致性，精神分裂症患者的脑网络在与默认网络进行信息交流时有异常，也即是默认网络作为信息整合中心，在精神分裂症患者中活动的异质性很大。本系列研究确立了突显网络、默认网络以及任务控制网络之间的信息交流的异常。突显网络在认知加工中负责分配对内外刺激的注意，而默认网络是更高级的信息整合中心，任务控制网络在执行各项任务控制时也出现异常。我们推测，这三大网络的信息交流的异常是精神分裂症患者信息加工的典型特征，为之后确定诊断疾病提供非常重要的生物标记物。

总体上，本系列实验从自然刺激的首发精神分裂症患者如何感知刺激入手，发现患者在很多高级的认知整合的区域存在变异，如扣带回、楔前叶、颞顶联合区以及脑岛。由于脑岛这研究一中表现出最大的组间差异，因为我们把脑岛作为种子点单独进行研究，结果发现，脑岛的分化在精神分裂症患者中更大，而以脑岛为关键节点的突显网络的内部相比健康被试也存在更低的功能连接强度。在研究局部异常的基础上，实验还考察基于全脑种子点的全脑功能连接的组间差异，结果显示，精神分裂症患者几乎所有的大脑功能网络相比健康被试都存在失连接。由于功能连接分析方法并不能排除内源性信号的影响，于是我们采用被试间功能相关的方法来探讨大脑网络在组与组之间的活动一致性，结果发现，精神分裂症患者在几乎有所的功能网络间的活动一致性都较低。

Schizophrenia is not only a mental illness, but also a cognitive disease. Cognitive impairment is the core feature of schizophrenia, and the cognitive heterogeneity of patients with schizophrenia is high. The brain dysfunction of schizophrenic patients is not only manifested as abnormal local processing, but also the disorder among large-scale networks. The traditional method using resting state fMRI or simple repetitive tasks stimulation to study brain function in patients with schizophrenia, resting state fMRI mainly reflect subject’s endogenous neural spontaneous activity, and simple repetitive tasks to stimulate the stimulus confined in a very specific and narrow scope, may not be very good induction of abnormal brain activity in patients with schizophrenia. This series of studies is an innovative use of the natural stimulus paradigm to study brain dysfunction in patients with schizophrenia. The natural stimulus paradigm refers to the video or/and auditory stimuli that present the subject with complex dynamics. These stimuli simulate events in the real world, with the purpose of evoking the subject's brain response to the real world. This series of studies used a variety of innovative methods to quantify brain responses to natural stimuli in schizophrenia and to compare them with healthy controls.
This series of studies is divided into five parts, respectively investigating the local brain dysfunction in patients with first-episode schizophrenia, the new insula partitioning method, the functional connectivity analysis of seeds based on insula, the functional connectivity analysis based on whole-brain seeds, and the intersubject functional correlation analysis based on whole-brain seeds. The first study explored the local brain dysfunction in patients with first episode schizophrenia. Inter-subject correlation (ISC) refers to the correlation of time series in the same brain region between different subjects viewing video stimulation. ISC provides a model-free approach to quantify brain responses during natural scene perception without the need to define a priori stimulus model. Such data-driven analysis is efficient for solving research problems such as social perception, because the stimulus space is too high dimensional to generate a complete set of prior hypotheses for an appropriate feature model. ISC filters out the effects of endogenous neural signals and non-neural signals, reflecting the role of neural signals evoked by stimulation. The results showed that compared with healthy controls, patients with first episode schizophrenia showed significantly lower ISC in a wide range of brain regions including bilateral precuneus, posterior cingulate gyrus, anterior cingulate gyrus, temporoparietal junction, precuneus gyrus, superior temporal gyrus, middle frontal gyrus, and insula when viewing video stimulation. A high ISC in the visual cortex reflects video stimulation that schizophrenics and healthy controls looked at carefully as well. Abnormalities in the core nodes of the salient network and the default network indicate abnormal activity in these core regions during self-processing and social processing, memory retrieval, and attention allocation to internal and external stimuli.
Based on the results of study 1, studies 2 and 3 mainly investigated insula subdivision and functional connectivity analysis based on insula seeds. The ISC study found that large bilateral insula showed high inter-group activity heterogeneity, while the right insula showed the highest inter-group functional heterogeneity. The results showed that the insula showed a trend of anterior and posterior insula function separation in both schizophrenic patients and healthy controls. Functional connectivity studies based on insula seeds found that bilateral anterior insula and posterior insula had high functional connectivity with sensorimotor cortex, visual cortex, anterior cingulate cortex, posterior cingulate cortex, dorsal prefrontal cortex and insula itself. Statistical analysis of intergroup functional connectivity showed that for the left anterior insula and posterior insula, the functional connectivity differences were distributed in bilateral prominence networks. For the right anterior insula and the right posterior insula, functional connectivity differences between groups were distributed in the right highlighting network. These results suggest that the insula is abnormal in processing salient stimuli and allocating attention.
Study 4 explored functional connectivity based on whole-brain seeds in the natural stimulus paradigm. After comparing the functional connections between groups, it was found that the nodes in the default network had weaker functional connectivity with the cingulate cortex, insula, sensory and perceptual cortex and visual cortex. Functional connectivity of sensory motor cortex nodes are almost all related to abnormal cingulate gyrus, especially anterior cingulate gyrus. For the cingulate task control network, auditory network and memory extraction network, functional connectivity differences between groups were mainly concentrated in the right insula, which was one of the core nodes of the salient network. For the nodes of the frontal and parietal control network, the differences in functional connections between groups were manifested in cingulate gyrus, precuneus and insula. Posterior cingulated gyrus and precuneus are the core nodes of the default network, while insula is the core node of the salient network. These results again validate the default network, task control network, and the communication patterns that highlight the connections and exceptions between networks, and the anomalies among the three networks. For the salience networks, the functional connectivity between groups show large-scale dissonance between networks, especially thos between groups and default networks, sensorimotor cortex and task control networks. Overall, we can conclude that under the natural stimulus paradigm, schizophrenia patients and healthy controls had abnormal connections between the default network, sensorimotor cortex, taskcontrol network, and salience network, and that almost all of these connections pointed to abnormalities in the insula.e Finally, we investigated the inter-subject functional correlation (ISFC) group based on 264 functional seeds. On a more macroscopic scale, the activity consistency of the stimulus evoked signals between the two groups was examined. Unlike functional connectivity based on wholebrain seeds, ISFC filters out the effects of endogenous neural signals in the subjects, as well as the effects of non-neural signals brought about by the scanner. The results showed that the intergroup functional variation was mainly focused on the difference of activity consistency between sensory motor network, cingulate task control network, auditory network, visual network, frontal and parietal control network, salience network and a small part of subcortical network and the default network of healthy controls. These results show that compared with the default network activity consistency of healthy controls, the brain network of patients with schizophrenia has abnormalities in the information communication with the default network, that is, as the information integration center, the default network has great heterogeneity in the activity of patients with schizophrenia. This series of studies establishes anomalies in the communication between salient networks, default networks, and task-control networks. Highlighting that the network is responsible for allocating attention to internal and external stimuli in cognitive processing, while the default network is a more advanced information integration center, the task control network also shows abnormalities in the implementation of various task control. We conclude that the abnormal information exchange in these three networks is a typical feature of information processing in patients with schizophrenia, and provides a very important biomarker for later diagnosis of the disease.
Overall, this series of experiments, starting with how first-episode schizophrenics with natural stimuli perceive stimuli, found variations in many areas of higher cognitive integration, such as the cingulate gyrus, precuneus, temporoparietal junction, and insula. Since the insula showed the largest inter-group differences in this study, because we used the insula as a seed point to conduct a separate study, it was found that insula differentiation was greater in patients with schizophrenia, and the insula as a key node of the highlighted network also had lower functional connectivity intensity compared with healthy subjects. On the basis of local abnormalities, the experiment also examined the differences between groups in whole-brain functional connections based on whole-brain seed points, and the results showed that almost all the brain functional networks in patients with schizophrenia were disconnected compared with healthy subjects. Since the functional connectivity analysis method could not exclude the influence of endogenous signals, we used the intergroup functional correlation method to explore the activity consistency of brain networks between groups, and found that the activity consistency of schizophrenia patients in almost all functional networks was low.