NCS-FO: Understanding the computations the brain performs during choice

这项建议试图通过采取跨学科的方法来理解大脑如何支持和编码决策。研究人员计划使用参与者做出复杂决定时记录的神经活动。然后，这种大脑活动可以被捕捉到一个模型中，该模型对每个大脑区域如何对决策做出贡献做出特定的预测。反过来，研究人员希望使用这个模型来测试关于这些区域的神经刺激如何影响决策的具体假设。关于大脑如何支持决策的知识可以为教育课程的发展提供宝贵的见解。信息最好以一种与大脑协同工作的方式呈现给学生。此外，来自该项目的见解可能会为人工智能的发展提供洞察。这一潜在的应用具有深远的伦理影响，研究人员打算在这项提案中进行研究。最后，这项工作侧重于评估决策背后的价值，这是经济学的基础。具体地说，估值基于经验维持和/或更新我们对收益或损失的理解。过度冒险或厌恶风险的行为可能会导致次优决策，并与神经心理障碍有关。拟议的研究分为三个同时进行的目标。首先，研究人员建议确定数据的哪些特征(例如，神经活动的幅度)提供有关决策过程的信息。在第二项研究中，研究人员的目标是建立解码器，将这些特征明确地结合在一起，以预测这个人会做出哪个决定。在最终目标中，研究人员的目标是开发一种控制算法，预测一个人的决定，然后通过应用神经刺激来改变决定进行干预。这种闭环刺激可能被证明对开发干预措施很重要，例如，这种干预可能有助于减少强迫症赌徒的冒险行为。或者，对于一些过于厌恶风险的人(从不离开家)，在不同类型的刺激下可以降低这种对风险的厌恶。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This proposal seeks to understand how the brain supports and encodes decision making by taking an interdisciplinary approach. The researchers plan to use neural activity recorded while participants make complex decisions. This brain activity can then be captured in a model that makes particular predictions about how each brain area contributes to decision-making. In turn, the researchers hope to use that model to test specific hypothesis about how decision-making is affected by neurostimulation to those areas. Knowledge of how the brain supports decision making can provide invaluable insight for the development of educational curricula. Information is best presented to students in a way that works in concert with the brain. Further, insights from this project may provide insight into the development of artificial intelligence. This potential application has profound ethical implications that the researchers aim to study in this proposal. Finally, this work focuses on assessing the value underlying decision making which is fundamental to economics. Specifically, valuation maintains and/or updates our understanding of gains or losses based on experience. Overly risky or risk-averse behavior can lead to suboptimal decisions and have been linked to neuropsychological disorders. The proposed research is split into three simultaneous objectives. In the first, the researchers propose to identify which features of the data (e.g., amplitude of the neural activity) provide information about the decision-making process. In the second, the researchers aim to build decoders that explicitly combine those features to predict which decision the person will make. In the final objective the researchers aim to develop a control algorithm which predicts the person’s decision and then intervenes by applying neurostimulation to alter that decision. This closed loop stimulation may prove important for developing interventions that could, for example, help reduce risk seeking in compulsive gamblers. Alternatively, in some people that are too risk averse (never leave the house), at different type of stimulation could reduce this aversion to risk.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.