fMRI-based characterization of brain activity patterns during weight loss and body weight maintenance: relation to hormonal mechanisms regulating energy homeostasis

肥胖是21世纪的主要流行病之一。然而，导致这种情况的神经机制仍然知之甚少。最近的研究表明，激素因素（如瘦素）和多巴胺能脑区（如纹状体）的活动之间存在联系，纹状体与奖励感知和刺激反应习惯学习有关。以往的研究表明，背纹状体在习惯学习和食物线索反应中都起着重要作用。肥胖发展和维持的第二个关键区域似乎是与目标导向行为以及行为的启动和抑制有关的前额叶皮层。人们怀疑，在高热量食物刺激后，中边缘和前馈连接区域（如眶额皮层和背外侧皮层）的激活/失活是由瘦素水平调节的，这可能是维持肥胖的神经机制之一。因此，该项目的第一个目标是调查与减肥和可能的体重反弹相关的生活方式干预如何与食物线索反应的变化和执行功能的变化（通过延迟满足任务来测量）相关，以及这些变化如何由激素和心理因素介导。这些目标将通过测量与食物相关的行为任务中的功能磁共振成像（fMRI）信号来实现。另一个需要解决的问题是，对与紊乱相关的神经元过程的分析是否可以从多元模式识别程序或分类器的应用中受益。这些多变量算法能够比传统的单变量技术更有效地提取信息，因为它们可以考虑到大脑活动空间模式中包含的全部信息。多变量方法在信号检测方面往往比单变量方法更敏感，这使得它们非常适合应用于基于功能磁共振成像激活模式的神经或精神诊断。这个项目的独特方法包括同时调查导致肥胖的关键机制，激素和神经，因此可能会大大提高我们对这种情况的理解。

Obesity is one of the major epidemics of the 21st century. However, neural mechanisms contributing to this condition remain poorly understood. Recent approaches suggest a link between hormonal factors such as leptin and activity of dopaminergic brain regions such as the striatum involved in the perception of reward and stimulus-response habit learning. Previous studies indicate that the dorsal striatum plays an important role in both habit learning and food cue-reactivity. A second key area in the development and maintenance of obesity appears to be the prefrontal cortex linked to goal-directed behavior and initiation and inhibition of behavior. It is suspected that activation/deactivation of mesolimbic and feedforwardly connected regions such as orbitofrontal and dorsolateral cortex following highcalorie food-cue stimulation is modulated by leptin levels and may be one of the neural mechanisms contributing to the maintenance of obesity. Therefore, a first goal of this project is to investigate how a lifestyle intervention associated with weight loss and possible weight regain relates to changes in food cue-reactivity and changes in executive functioning as measured by a delay of gratification task, and how these changes are mediated by hormonal as well as psychological factors. Theses goals will be achieved by measuring functional magnetic resonance imaging (fMRI) signals during food-related behavioral tasks. A further question that will be addressed is whether or not the analysis of disorder-associated neuronal processes can benefit from the application of multivariate pattern recognition procedures or classifiers. These multivariate algorithms are able to extract information much more efficiently than conventional univariate techniques because they can take into account the full information contained in the spatial pattern of brain activity. Multivariate approaches tend to be more sensitive in signal detection than univariate approaches which in turn makes them highly suitable for the application in neurological- or psychodiagnosis based on fMRI activation patterns. The unique approach of this project consists in investigating simultaneously both key mechanisms, hormonal and neural, known to contribute to obesity, and hence may substantially advance our understanding of this condition.