MRS validation of computational metabolic modeling of human brain function to determine energetic disruptions underlying fMRI-derived functional connectivity in degenerative or psychiatric disorders

静息状态下的功能磁共振成像已经暗示了静息状态下人脑中高功能连接的枢纽，例如，的 默认模式网络（DMN），与其他地区相比。在本提案中，我们将直接评估 代谢在支持高连通性方面的作用，并评估代谢功能障碍是否会导致连通性 在神经退行性疾病中观察到的DMN中枢的减少（例如，老年痴呆症）和神经精神疾病。 健康人脑中功能连接的复杂性质具有非常高的能量需求， 通过葡萄糖氧化产生的高ATP（CMRglc（ox））满足的需求。使用H[ C] MRS，我们发现， 1 13 神经元CMRglc（ox）随多巴胺能神经传递呈线性变化，静息时大部分脑区CMRglc（ox）随多巴胺能神经传递呈线性变化， 大脑皮层的能量产生主要用于信号传递。然而，有一个很大的一部分精力用于 管家需要，如突触发生和维持膜电位。鉴于与 能量学和信号传导，我们惊奇地发现，使用定量PET成像，DMN中的总CMRglc（ox） 类似于具有较低fMRI衍生连接性的区域。 对这一悖论的一个潜在解释是，DMN与其他皮质区域相比，其中枢具有更大的 它们的总能量的一部分专门用于信令，而不是非信令，从而使DMN集线器更 容易出现功能性能量衰竭。可替换地，已经提出了更高的非信令需求（例如， 突触重塑）存在于DMN中。为了回答这个对口译有重大意义的新问题， 静息态fMRI数据，并研究能量代谢和组织成分功能障碍如何影响 功能，需要新的测量和计算工具。 为了解决这一挑战，我们将开发一个计算模型来计算信号和非信号 能源成本。该模型使用来自个体受试者的组织成分数据， 分辨率MRI。通过与1H[13C]进行比较，将在啮齿动物模型和人类中验证该模型。 MRS，它可以唯一地测量神经能量学的信号和非信号成分。的 将在高功能连接中测量和计算信号与非信号的相对比率 在健康的年轻人和老年人的DMN区域和控制低连接皮层区域。我们 假设高功能连接区域将具有更大比例的能量产生 致力于信号传导，并且这一部分将随着年龄的增长而下降。一旦计算预算模型 开发，并验证，它将提供一个强大的非侵入性的工具，研究皮层的变化， 能量学和组织组成，导致fMRI衍生的连接性丧失，以及潜在的 用于评估预后和治疗的新的临床生物标志物。

Resting-state fMRI has implicated hubs of high functional connectivity in the resting human brain, e.g., the default mode network (DMN), compared to other regions. In this proposal we will directly assess the role of metabolism in supporting high connectivity and assess whether metabolic dysfunction leads to connectivity reduction in DMN hubs seen in neurodegenrative (e.g., aging, Alzheimer's) and neuropsychiatric disorders. The complex nature of functional connectivity in healthy human brain has very high-energy demands, a need that is met by high ATP yielded from glucose oxidation (CMRglc(ox)). Using H[ C] MRS, we found that 1 13 neuronal CMRglc(ox) changes linearly with glutamatergic neurotransmission and that at rest most of the cerebral cortex's energy production is devoted to signaling. However there is a significant fraction of energy devoted to housekeeping needs such as synaptogenesis and maintaining membrane potentials. Given the tight link of energetics and signaling we found surprisingly, using quantitative PET imaging, that total CMRglc(ox) in the DMN is similar to regions with lower fMRI-derived connectivity. A potential explanation for this paradox is that the hubs in DMN vs. other cortical regions have a greater fraction of their total energy devoted to signaling than to nonsignaling, thus making the DMN hubs more vulnerable to functional energy failure. Alternatively it has been proposed that higher nonsignaling needs (e.g., synaptic remodeling) is present in DMN. To answer this novel question with large implications for interpreting resting-state fMRI data and to study how dysfunction of energy metabolism and tissue composition impacts function, there is a need for novel measurement and computational tools. To address this challenge we will develop a computational model to calculate signaling and nonsignaling energy costs. The model uses data from individual subjects on tissue composition obtained from high- resolution MRI. The model will be validated in both a rodent model and humans by comparison with 1H[13C] MRS, which can uniquely measure the signaling and nonsignaling components of neuroenergetics. The relative ratios of signaling to nonsignaling will be measured and calculated in high functional connectivity regions of the DMN and control low connectivity cortical regions in healthy young and elderly adults. We hypothesize that regions of high functional connectivity will have a greater fraction of energy production devoted to signaling and that this fraction will decline with age. Once the computational budget model is developed, and validated, it will provide a powerful noninvasive tool for studying alterations in cortical energetics and tissue composition that lead to loss of fMRI-derived connectivity as well as potentially as a novel clinical biomarker for assessing prognosis and treatment.