CASIA OpenIR  > 毕业生  > 博士学位论文
基于多组学数据的精神分裂症生物标记研究
李昂
Subtype博士
Thesis Advisor刘冰
2020-05
Degree Grantor中国科学院自动化研究所
Place of Conferral中国科学院自动化研究所
Degree Name工学博士
Degree Discipline模式识别与智能系统
Keyword精神分裂症 抗精神病药物 纹状体 精准诊疗 多巴胺 多基因风险
Abstract

精神分裂症是一种病因不明、高异质性和高致残性的重性精神障碍。精神分 裂症患者的主要临床表征是具有偏执的妄想和幻觉,以及注意、记忆和执行控制 等认知功能缺陷。精神分裂症的全球患病率约为 1%,给众多患者及家庭造成巨 大的精神痛苦和生活压力,给社会及国家带来沉重的经济负担。近几十年,精神 分裂症等相关精神疾病的治疗未取得突破性进展,患者预期寿命也没有得到显 著提升。精神分裂症的临床综合征定义太过宽泛并且其评估依赖医生的主观经 验,缺乏精准、有效和具有生物学意义的诊疗指标和治疗靶点是造成这种僵局的 主要原因之一。面对这一困境,我们需要能够对经典疾病诊断框架有所突破,能 够在更深层次的机理上理解精神疾病病理,特别是能在个体层次、可量化的生物 标记有新的发现。已有研究发现及假说提示,纹状体可能是精神分裂症的核心病 理性脑区,其功能在精神分裂症患者中存在病理性改变;同时众多临床案例也表 明,作为针对精神分裂症的主要治疗手段,抗精神疾病药物的药理都不同程度地 依赖于针对纹状体突触前多巴胺 D2 受体的拮抗机制。基于以上证据,我们合理 地推测:发现纹状体功能的有效异常表征,很有可能帮助我们找到精神分裂症精 准诊疗生物标记的钥匙。为此,我们基于多中心大样本的汉族人群精神分裂症数 据库,从多水平多组学的角度系统探索了精神分裂症纹状体功能异常的影像学 表征,发现了一种全新的神经影像学标记 — 纹状体功能损伤 (functional striatal abnormalities,FSA),可能是精神疾病精准诊疗的有效生物标记之一。本文从患 者纹状体的功能特征、FSA 的发现及诊疗模型构建、基于 FSA 的疾病分层、FSA 的跨诊断意义、FSA 的生物学意义等五个方面总结了我们所发现的这一全新的、 个体化的、可泛化的、具有生物学意义的精神分裂症生物标记相关研究结果,并 以此为基础探讨了其在临床个体化诊断治疗及研究精神疾病病理方面的潜在重 要意义。因此,本文的五个主要研究内容和贡献如下:

1. 全面分析精神分裂症患者纹状体功能的病理性特征

基于多中心大样本的精神分裂症神经影像数据,提取纹状体区域功能特征,揭示纹状体功能在精神分裂症患者中跨中心稳定的影像学表征。通过纹状体脑 区的比例低频振幅 (fractional amplitude of low frequency fluctuation, fALFF)、局 部一致性 (Regional Homogeneity, ReHo) 和纹状体内部功能连接(intra-striatal functional connectivity)、纹状体到全脑功能连接 (extra-striatal functional connectivity),我们从多层次系统研究了纹状体局部信号,纹状体系统内部环路 和纹状体到全脑环路在精神分裂症患者中的病理性改变。

2. 提出 FSA(functional striatal abnormalities,纹状体功能损伤)的概念及构建诊疗预测模型

利用以上所提取的高维度、多层次的纹状体功能特征数据,我们创新性地提 出了纹状体功能损伤 FSA 的概念,试图在组间统计分析的基础上,通过利用机 器学习技术,对精神分类症患者的纹状体功能损伤做出定量化评估,以在个体层 次衡量纹状体功能相对健康状态到严重病理性损伤之间的程度(FSA 分数)。在 此基础上,构建了相应的诊疗模型,并对所构建的模型进行了严格的跨中心交叉 验证,结果发现 FSA 分数可以较准确地在不同独立影像中心区分健康对照和精 神分类症患者,并与抗精神病药物的短期治疗效果显著相关。同时,我们还发现 了 FSA 分数对抗精神病药物预后的预测能力,以及与药物的 Meltzer 比例(五羟 色胺 5-HT2A 与多巴胺 D2 的亲和力之比)具有显著的相关性。

3. 基于 FSA 的疾病亚型分析

在发现及验证了 FSA 对于精神分裂症分类及预后的效应后,我们继续进行 了拓展研究,发现纹状体功能损伤在不同独立中心的精神分裂症患者中均存在 显著增加的个体差异。基于这一差异,我们尝试将患者分成不同的疾病亚型,揭 示了不同亚型存在显著的临床特性、大脑形态学的差异。进而我们验证了这种新 的影像学标记不对抗精神病药物的治疗状态及功能核磁共振头动状态等指标敏 感。这些结果表明 FSA 可能作为一种稳定的、个体化的生物标记对精神分类症 患者进行临床及病理分层。

4. 探索 FSA 在精神疾病跨诊断方面的意义 

在前三个研究的基础上,我们基于多中心的不同精神疾病神经影像数据集 ,进一步研究了纹状体功能损伤是否在双相情感障碍、抑郁症、多动症及强迫 症患者中存在跨诊断病理学意义。通过分析国内外四个独立影像中心同时包含 精神分裂症、双相情感 障碍患者及健康对照的数据,我们发现了该纹状体功能损伤与双相情感障碍显 著相关,而且这种效应量显著弱于精神分裂症患者。然而,我们并未在抑郁症 、多动症及强迫症患者中观测到精神分裂症维度显著的纹状体功能损伤。该结 果提示纹状体损伤可以作为潜在的精神疾病跨诊断的生物标记。

5.理解FSA背后的生物学意义

在发现并多方面验证了 FSA 可能是精神分裂症有效的生物标记之后,我们 进而结合多组学的数据试图对这种纹状体功能损伤对应的生物学意义进行一定 的机制探讨。主要的技术手段是结合正电子发射扫描 (Positron emission tomography, PET),单光子发射计算机化断层显像 (Single photon emission computed tomography, SPECT) 和艾伦人脑转录组图谱 (Allen Human Brain Atlas)数据开展相应的分析。通过结合以上多组学数据以及相应的生物信息学分 析,我们发现 FSA 与纹状体局部功能损伤模式在空间模式上具有稳定的相关 性;同时,我们还发现纹状体局部功能损伤模式与精神分裂症多基因风险表达 模式、以及多巴胺系统分子分布存在显著的空间对应关系,尤其是多巴胺 D2 受体和多巴胺综合合成物。通过跨模态空间分析和基因富集分析,我们找出了 一系列与纹状体功能异常显著相关的基因通路。这些结果可以为我们理解精神 分裂症纹状体功能损伤的生物学机制提供重要帮助。

Other Abstract

Schizophrenia is a fatal mental disorder with unknown etiology, high heterogeneity, and high disability, in which people have the main clinical features of paranoid delusions and hallucinations, as well as cognitive deficits in attention, memory, and executive control. The global morbidity rate of schizophrenia is about 1%, resulting in tremendous mental pain and life pressure to numerous patients and their families, and bringing about the heavy economic burden to society and the country. In recent decades, there is little breakthrough in the treatment of schizophrenia and other psychiatric disorders, and there is no significant improvement in terms of patients’ life expectancy. One of the main reasons for this stalemate is the lack of precise, effective and biologically meaningful treatment targets and diagnostic indicators, which due to the clinical syndrome of schizophrenia is defined too generally, and its evaluation relies on the subjective experience of psychiatists. Solving the puzzle requires to reconsider the conventional disease diagnosis framework and to understand the pathology of psychiatric disorders deeply, especially to develop individual-level biomarkers. Despite of clinicians’ intuitive experience as the main basis for the clinical diagnosis of psychiatry and no effective biomarkers available, existing research literature and hypotheses suggest that the striatum may be the core pathological brain region of schizophrenia, and its function may have pathologically changed in patients with schizophrenia. In the meantime, antipsychotic medication is the main treatment option for schizophrenia pharmacologically, which depends on the antagonistic mechanism of the striatal presynaptic dopamine D2 receptor. Given above, it is reasonable to hypothesize that the striatal dysfunction is potentially a useful biomarker with application of diagnosis and prognosis. In view of this, basing on the multi-center large sample in Chinese Han population schizophrenia database, we proposed and developed a new neuroimaging marker called functional striatal abnormalities (FSA).

This paper is the summary of our research results on the goup-level striatal dysfunction in schizophrenia, the establishment of the FSA model and its applications, the clincal stratification potential of FSA, the transdiagnostic implications of FSA, the possible biological significance of FSA .These five aspects summarize what we found this new, individualized, generalized, biological significance of schizophrenia biomarkers related research results, and on this basis, discussed its clinical individualized diagnosis and treatment and research of mental illness pathological aspects of potential significance. The five primary research contents and contributions of this paper are as follows:

1. Group-level striatal functional features in individuals with schizophrenia

Based on multi-center large-sample schizophrenia neuroimaging data, the functional features of the striatum were extracted to reveal reproducible imaging features of the striatum in schizophrenic patients. Multi-level systematic discussion of pathological changes in the regional signal, internal circuit of striatal system and corticostriatal to the whole brain circuit in patients with schizophrenia was performed through the ratio of low-frequency amplitude (fALFF), regional homogeneity (ReHo) and inter- and extra- striatal functional connectivity.

2.The discovery of the FSA (functional striatal abnormalities) and the establishment of the FSA model

Based on the high-dimensional, multi-level functional features of the striatum extracted from research 1 and between-group analysis, the concept of FSA was further proposed using machine learning. An individualized biomarker FSA score was developed in combination with the machine learning technology for quantitative assessment of striatum functional impairment in patients with schizophrenia between the relatively healthy state of striatum and severe pathological damage. After strict cross-validation, we found this biomarker can accurately distinguish healthy volunteers from individuals with schizophrenia in different independent imaging centers, and it is significantly correlated to the short-term antipsychotic medication effects. Further, we also found that its ability to predicting the prognosis of antipsychotics is related to the drug's Meltzer ratio (the ratio of the affinity of serotonin 5-HT2A to dopamine D2).

3. The patient stratification stragety based on FSA

After discovering and verifying the effect of FSA on the classification and the prognosis prediction of schizophrenia, we also found that schizophrenia patients with striatal dysfunction in different independent centers showed significant individual differences. We tried to map patients into different disease subtypes, revealing that significant clinical characteristics and differences exist in brain morphology between different subtypes. Then, it is verified that this new imaging marker is not sensitive to the treatment status of antipsychotic drugs and the state of functional MRI head motion. These results indicate that FSA may be used as a stable and individualized biomarker for clinical and pathological stratification of schizophrenia patients.

4. The transdiagnostic implications of FSA

Based on the research above and the multi-center neuroimaging data sets of different mental diseases, whether striatal dysfunction has transdiagnostic significance in patients with bipolar disorder, depression, ADHD and OCD have been further investigated by our research results. By analyzing the data of four independent imaging centers including individuals with schizophrenia, bipolar disorder and healthy controls, we revealed that striatal dysfunction is significantly correlated with bipolar disorder, and the extent of FSA scores are significantly lower than those in healthy controls, and higher than schizophrenia. However, we have not observed significant striatal dysfunction in individuals with depression, ADHD, and OCD. This result suggests that striatal dysfunction can be used as a biomarker for neurodevelopment disorders in psychiatry.

5. The potential biological underpingings for FSA

After revealing and validating FSA as a stable and promising biomarker for schizophrenia, we further discussed this striatal dysfuncion for its possible biological underpingings. Our primary technical approach was to combine positron emission tomography (PET), single photon emission computed tomography (SPECT) with Allen Human Brain Atlas data. We mainly adopted specific research methods as follows: we constructed a polygenic risk expression pattern in the striatum based on the schizophrenia risk reported in the previous large-scale genome-wide association study, and we found that its spatial pattern was signifincantly correlated with the regional striatal dysfunction pattern. Meanwhile, we revealed a significant spatial corresponding relation between this pattern and the dopaminergic system, especially for dopamine D2 receptor and dopamine synthesis. Through cross-modal spatial analysis and gene enrichment analysis, several pathways with significant correlation with striatal dysfunction were identified. These results will contribute to our understanding of the physiological mechanism of striatal dysfunction in schizophrenia.

Subject Area人工智能其他学科
Pages121
Funding ProjectNational Basic Research Program of China (973)[2011CB707800] ; National Basic Research Program of China (973)[2011CB707800]
Language中文
Document Type学位论文
Identifierhttp://ir.ia.ac.cn/handle/173211/40385
Collection毕业生_博士学位论文
Recommended Citation
GB/T 7714
李昂. 基于多组学数据的精神分裂症生物标记研究[D]. 中国科学院自动化研究所. 中国科学院自动化研究所,2020.
Files in This Item:
File Name/Size DocType Version Access License
毕业论文最终版本.pdf(23410KB)学位论文 限制开放CC BY-NC-SA
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[李昂]'s Articles
Baidu academic
Similar articles in Baidu academic
[李昂]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[李昂]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.