Neural, hormonal and behavioral mechanisms of long-term weight maintenance

肥胖由于其高患病率和严重的医疗后果是一个主要的全球健康问题。肥胖治疗的关键问题是，在饮食引起的体重减轻后，体重经常反弹，甚至可能导致体重水平高于基线水平。这种影响只有通过长期观察研究才能得到确认。这种体重循环过程的神经机制及其与激素和能量稳态行为参数的关系在很大程度上是未知的。该项目旨在同时评估在参与饮食干预研究后2至3年的时间间隔内，最初肥胖受试者与体重相关的神经、激素和行为参数。神经活动将使用功能磁共振成像结合磁共振波谱、血液样本、行为协议和多变量分析技术来测量。该项目有两个主要目标。首先，它旨在描述基于神经、激素和行为参数的体重维持机制。其次，它旨在从这些测量中得出适合长期身体维护的纵向预测因子。关于大脑活动，我们假设体重变化将通过参与调节非稳态食物摄入的与奖励相关的大脑区域（例如纹状体、眼窝额叶皮层和脑岛）的活动变化来反映。我们假设这些区域的活动也可以预测未来的体重变化。此外，我们提出了一个假设，即与食物相关的自我控制相关的区域（例如，背外侧前额叶皮层）的活动将与给定的体重有关，并包含对未来体重变化的预测信息。为了实现这些目标，将应用高灵敏度的多变量模式识别分析技术。这些方法还可以识别因素之间的微妙关系，这些关系只能通过考虑数据的共变结构（在大多数传统研究中应用的单变量分析方法忽略了这一点）来实现。由于结合了多模式数据采集和最先进的分析技术，该项目有望大大提高我们对持续肥胖控制的长期机制的理解。

Obesity is a major worldwide health problem due to its high prevalence and severe medical consequences. The key problem in the treatment of obesity is the frequent weight regain following a dietary induced weight loss that can even lead to weight levels above the level of baseline. Such effects will be recognized only with long-term observational studies. The neural mechanisms underlying such weight cycling processes and their relation to hormonal and behavioural parameters of energy homeostasis are largely unknown. The proposed project aims to simultaneously assess neural, hormonal, and behavioural parameters related to body weight in initially obese subjects during an extended interval of 2 to 3 years after they participated in the dietary intervention study. Neural activity will be measured using functional magnetic resonance imaging in combination with magnetic resonance spectroscopy, blood samples, behavioural protocols, and multivariate analysis techniques. The project has two major goals. First, it aims to characterize the mechanisms of weight maintenance based on neural, hormonal, and behavioural parameters. Second, it aims to derive suitable longitudinal predictors for long-term body maintenance from these measures. Regarding brain activity, we hypothesize that body weight changes will be reflected by changes in activity of rewardrelated brain areas involved in the regulation of non-homoeostatic food-intake (e.g., the striatum, the orbitofrontal cortex, and the insula). We assume that activity in these areas will also be predictive of future weight changes. Moreover, we put forward the hypothesis that activity in areas involved in food-related self-control (e.g., the dorsolateral prefrontal cortex) will be related to a given body weight and contain predictive information for future weight changes. To address these goals, highly sensitive multivariate pattern recognition analysis techniques will be applied. These methods can also identify subtle relations between factors that are only accessible by taking into account the co-variation structure of the data (which is ignored by univariate analysis methods applied in most traditional studies). Due to the combination of multimodal data acquisition and state-of-the-art analysis techniques, the project promises to substantially advance our understanding of long-term mechanisms of sustained obesity control.