焦虑，尤其是严重而持久的焦虑，会严重干扰人们日常生活的一系列认知功能，从基本的注意、记忆到高级的社会决策、经济决策等等。例如，大量研究表明焦虑和人们在经济决策中的风险偏好有关。然而焦虑为什么会影响到风险决策，如何影响，又造成了怎样的认知和神经机制的变化？这些问题一直困扰着当今世界的心理学家、神经科学家以及经济学家。通过回顾近年来关于焦虑个体认知加工、情绪加工的理论模型与神经回路的研究以及关于风险决策相关的的认知理论与神经基础的研究，我们认为特质焦虑个体的风险决策可能与其变化的情绪机制有关。因而，我们试图通过一系列的特质焦虑相关行为及脑成像研究来对上述问题进行研究。本研究首先采用经典的框架效应实验范式，通过操纵经济决策中的潜在情绪背景（即抉择选项被描述为获益或损失），来考查特质焦虑和风险决策反应模式间的关系。研究发现，高焦虑个体表现出放大的框架效应，特质焦虑者的风险偏好受潜在情绪背景调节，当情绪背景为正性（即获益）时，高特质焦虑者表现出更高的风险厌恶或风险回避，而情绪为负性时，高特质焦虑者表现出更高的风险追求。这一研究结果表明特质焦虑者在风险决策中的行为反应模式与风险决策中潜在的情绪信息有关，而与风险偏好本身无关。随后本研究采用功能磁共振成像技术对这一现象的神经机制进行检测。研究发现，在框架效应下，个体的特质焦虑水平与基于杏仁核的“情绪”系统的活动强度呈显著正相关，而与基于背侧扣带前回的“理性”系统的活动强度呈显著负相关，而腹内侧前额叶特质焦虑与这两个系统的交互作用中有重要枢纽作用。即高焦虑个体夸大的框架效应和其功能亢进的情绪神经系统及功能减退的理性神经系统有关。此结果为高焦虑个体在面对风险时更为情绪化的行为表现进一步提供了神经水平的证据。那么特质焦虑个体在面对风险时为什么会表现出这种过于情绪化的决策行为？这种对情绪信息过度敏感的原因何在？是否来自于对损失的过度担心？最后我们采用损失厌恶的实验范式和功能磁共振成像技术，通过控制风险水平和调节背景信息的情绪水平（即获益/损失比），来试图对这一问题进行回答。研究发现，高焦虑群体表现出了对损失的过度厌恶，同时对应着腹内侧前额叶的过度激活和杏仁核的活动异常。这一结果表明，特质高焦虑群体对风险的过度回避或追求，本质上可能是对损失的过度厌恶及理性-情绪相关神经系统的异常活动。通过系统地考查特质焦虑者风险决策模式的认知和神经机制，我们发现了焦虑之所以影响风险决策是由于其对情绪信息的敏感性，其风险偏好随着情绪信息的变化而变化，而这种敏感性来自焦虑个体对损失的过度厌恶，同时伴随着基于杏仁核等情绪神经系统的功能亢进和基于背侧扣带前回等理性脑区的功能减退，以及腹内侧前额叶、纹状体等价值编码相关神经系统的紊乱。本研究为人类情绪和决策的研究以及临床焦虑障碍的评估和治疗提供了证据。

Anxiety, especially when chronic and persistent, can be hugely disruptive to cognitive function, ranging from attention and memory to social and economic decision making. While the mechanisms for anxiety’s impact on basic cognitive and emotional functions are more established, the nature of interaction of these functions in higher level processes (e.g., decision making) is less clear. A large body of research suggests that anxiety is associated with risk aversion in economic decision making. However, it is unclear whether anxiety specifically impacts the cognitive or emotional processes, neural mechanisms, or all of these, in relation to decision making. Reviewing recent studies of the behavioral, neurocognitive, and emotional aspects of risk decision making in trait anxiety, we hypothesize that risk decision making in trait anxiety is predominantly a function of altered emotional processing. More specifically, we posit a shift towards heightened use of emotional information in decision making. To examine this hypothesis, we conducted a series of behavioral and brain imaging experiments to explore the influence of emotion on decision making in consideration of trait anxiety. In the current study, we manipulated the emotional context in economic decision making by using a framing effect paradigm. The paradigm manipulates the context of the risk presentation such that choices are framed as potential gains or potential losses, despite actually being comparable in probabilistic outcome. Our results indicated an amplification of the framing effect with increased trait anxiety; the risk preference of anxious individuals was modulated by the emotional context. When the risk was presented competing against a gain frame (positive emotional context), high anxious individuals exhibited higher risk aversion compared to the low anxious group. When the option was presented against a loss frame (negative emotional context), high anxious individuals showed greater risk seeking compared to individuals with low anxiety. The findings suggest that risk preferences of anxious individuals were associated with the emotional context, rather than the risk per se, in the decision making.Using functional magnetic resonance imaging, we found that under the framing effect, trait anxiety was positively correlated with activity of an amygdala based emotional system, but negatively correlated with activity of a dorsal anterior cingulate cortex (dACC) based analytic system. Further we found that the ventral medial prefrontal cortex may function as a hub region between the emotional (amygdala-based) and analytic (dACC-based) systems. The results suggest that the amplified framing effect in trait anxiety is underpinned by a hyper-active emotional and a hypoactive analytic system in the brain. These findings provide neural evidence for the emotional behavior in trait anxiety.To clarify the nature of the anxious vulnerability to emotional context, we conducted an additional experiment, controlling for risk level and manipulating the level of emotional valence (e.g. the ratio between gain and loss) by using a loss aversion task and simultaneous functional magnetic resonance imaging. The results showed that the high anxious group exhibited greater loss aversion than the low anxious group. These behavioral findings were accompanied by relatively more activity in the ventral medial prefrontal cortex, anterior insula and striatum, and relatively less activity in the amygdala. These findings suggest that risk avoidant or risk seeking behavior in elevated anxiety is likely a consequence of increased loss aversion and abnormalities in the underlying emotional and reward neural systems.Through systemically investigating the neurocognitive mechanisms of risk decision making in trait anxiety, we found that anxious individuals are more vulnerable to the emotional context in decision making, wherein the risk preference is changed by this information. Furthermore, either risk avoidant or risk seeking, essentially, is a result of heightened loss aversion in trait anxiety during decision making. This elevated loss aversion is underpinned by a hyper-active amygdala based emotional system, a hypoactive analytic dACC based system, and alterations to the subjective value processing in the ventral medial prefrontal cortex and striatum. Our findings provide evidence for human emotion and decision making studies and clinical application of anxiety disorders.