Multi-modal imaging of functional systems in the human brainstem

当涉及到维持我们的生命时，脑干是大脑中最重要的部分。尽管它非常重要，但它在很大程度上被人类神经科学所忽视。造成这种忽视的最重要原因是，标准的神经科学测量方法，如功能磁共振成像（fMRI），在大脑的这一部分表现不佳，这主要是由于生理噪声水平升高。迄今为止，fMRI研究仅限于单个脑干核的基本激活研究，而由于独立成分分析（ICA）的应用问题，对功能连接的研究大多不成功。这种关键方法在脑干中存在严重问题，因为标准方法无法将生理噪声抑制到一定程度，在这种情况下，神经元来源的信号成为数据方差的主要来源。该项目的中心目标是改进申请人最近开发的一种新的脑干-功能磁共振成像方法，使其可以用于可靠地测量单个核的神经元活动以及核间和核-皮层的连通性。新方法使用了一种完全不同的生理噪声抑制方法，可以应用于标准的功能磁共振成像数据集。为了达到这一目标，我们将首先优化和激励之前成功分析中的一些特定参数选择，如ICA分解的维度数或脑干解剖面具的确切形状。随后，申请人将与来自哈佛大学波士顿分校马蒂诺斯生物医学成像中心的学术合作伙伴一起，在当地的7-特斯拉高场MRI扫描仪上获取20-30个高分辨率解剖扫描样本，并开发一种数据融合方法，将这些结构数据与申请人现有的100多个受试者的3-特斯拉功能数据集结合起来。目的是大大提高对脑干激活簇的解剖结构的识别。最后，我们将把改进的方法应用于学术合作伙伴最近获得的三个现有的常见慢性疼痛综合征数据集。这些数据包括纤维肌痛、腰痛和腕管综合征。目的是确定与疼痛调节过程相关的核，并调查这些核是否在慢性疼痛患者中表现出病理改变的连通性。令人感兴趣的问题是，改变的脑干内在连接是否可能是不同疼痛综合征的共同机制，以及新皮质中心还是脑干中心在慢性疼痛中发挥最重要的作用。

The brainstem is the most important part of the brain, when it comes to sustaining our life. Despite its tremendous importance, it has been largely neglected by human neuroscience. The most important reason for this neglect lies in the poor performance of standard neuroscientific measurement methods, like functional magnetic resonance imaging (fMRI), in this part of the brain, which is mainly due to the elevated level of physiological noise. fMRI investigations have so far been restricted to basic activation studies of single brainstem nuclei, while investigations on functional connectivity have been mostly unsuccessful due to problems in the application of independent component analysis (ICA). This key method suffers from severe problems in the brainstem as standard approaches are unable to suppress physiological noise to a point, where signals of neuronal origin become the major source of variance in the data. The central goal of this project is to improve a new brainstem-fMRI approach, recently developed by the applicant, to a point where it can be used to reliably measure neuronal activity of single nuclei as well as inter-nuclear and nucleo-cortical connectivity. The new approach uses a radically different approach of physiological noise suppression and can be applied to standard fMRI datasets. To reach this goal, we will first optimize and motivate some ad hoc choices for parameters in previous successful analyses, like the number of dimensions for the ICA decomposition or the exact shape of the brainstem anatomical mask. Afterwards, the applicant together with the academic partner from the Martinos Center for Biomedical Imaging of Harvard University Boston will acquire a sample of 20-30 high-resolution anatomical scans on the local 7-Tesla high-field MRI scanner and develop a data fusion approach to combine these structural data with an existing 3-Tesla functional dataset of more than 100 subjects acquired by the applicant. The aim is to greatly improve the identification of anatomical structures underlying activation clusters in the brainstem. Finally, we will apply the improved method to three existing datasets of common chronic pain syndroms recently acquired by the academic partner. These include data on fibromyalgia, low back pain and carpal tunnel syndrome. The aim is to identify nuclei of relevance for pain modulatory processes and investigate, whether these nuclei show pathologically altered connectivity in chronic pain patients. Questions of interest are, if altered brainstem intrinsic connectivity may be a common mechanism to different pain syndromes, and whether neocortical or brainstem centers play the most important role in chronic pain.