Making the connection: Methods to infer functional connectivity in brain studies

在帕金森病(PD)这一常见的神经退行性疾病中，脑细胞丢失与临床残疾之间存在不一致之处，因为PD存在实质性的代偿机制，将整体损害降至最低。因此，对帕金森病患者大脑活动变化的测量是疾病和代偿过程的产物，尽管这些影响对于理解残疾的机制至关重要，但分离这些影响仍然是一个挑战。由于越来越多的证据表明，由于大脑区域之间动态连接模式的改变，正常和疾病的大脑功能不同，因此利用目前可用于非侵入性评估PD对象的技术，即fMRI和EEG，开发评估连接的模型至关重要。我们的具体目标是将疾病效应从代偿机制中分离出来，并在3个主要领域确定药物(L-多巴)对帕金森病的影响：连接性、活动的空间范围和时间动力学。UBC、SFU和UNC的这一联合倡议代表了临床医生-治疗和执行PD功能磁共振研究的科学家(McKeown和Huang博士)、在大脑分割和形状分析方面具有专业知识的科学家(Abugharbieh和Beg博士)以及网络建模和统计推理(Wang博士)-之间的新合作。(拟议的研究包括以下技术部分：1.)扩展新开发的动态因果模型(DCM)以实现大脑的有效连接；提供fMRI激活的空间特征的定量分析；确定网络模型对人工绘制的感兴趣区域的错误指定的敏感性；以及4.)将盲源分离(BSS)方法扩展到脑电数据，以推断脑区之间的瞬时同步。这种跨学科的研究努力具有重大影响，因为它将促进对疾病进展和代偿性变化的理解，所开发的方法将广泛适用于其他大脑疾病。此外，加拿大制药业将从区分新疗法的神经保护和症状效应的方法中受益匪浅。

In Parkinson's disease (PD), a common neurodegenerative disorder, there is discordance between brain cell loss and clinical disability because substantial compensatory mechanisms occur in PD, minimizing overall impairment. Thus measures of altered brain activity in PD are a product of both disease and compensatory processes, and disentangling these effects, while vital for understanding the mechanisms of disability, remains a challenge. As there is increasing evidence that normal and diseased brain functioning differ by altered patterns of dynamic connectivity between brain regions, it is critical to develop models assessing connectivity  using current technologies available for non-invasively assessing PD subjects, i.e. fMRI and EEG. Our specific aims are to isolate disease effects from compensatory mechanisms and to determine the effects of medication (L-Dopa) in PD across 3 main domains: Connectivity, Spatial Extent of Activity, and Temporal Dynamics. This joint initiative between UBC, SFU and UNC represents a new collaboration between clinician-scientists treating and performing fMRI research in PD (Drs. McKeown and Huang); scientists with expertise in brain segmentation and shape analysis (Drs. Abugharbieh and Beg), as well as network modeling & statistical inference (Dr. Wang).  The proposed research consists of the following technical components: 1.) Extend the newly developed dynamic causal models (DCM) for brain effective connectivity; 2.) Provide quantitative analysis of the spatial characteristics of fMRI activation; 3.) Determine the sensitivity of network models to mis-specification of manually-drawn regions of interest; and 4.) Expand Blind Source Separation (BSS) methods to the EEG data to infer transient synchronization between brain areas. This cross-disciplinary research effort has significant impact because it will advance understanding of both disease progression and the compensatory changes, and the developed methods will be widely applicable to other brain diseases.  Also, the Canadian pharmaceutical industry would substantially benefit from a means to differentiate between neuroprotective and symptomatic effects of novel therapies.