The Neural Mechanisms of Peer Influence on Risky Decision-Making

 描述（由申请人提供）：危险行为，如吸烟，吸毒和酗酒，以及青少年性行为都受到同龄人行为的严重影响。虽然人们对参与风险决策的神经网络有很多了解，但这项研究在很大程度上是在现实生活中做出风险决策时缺乏丰富的社会信息网络的情况下进行的。鉴于大量证据表明，在做出风险决策时，有关同伴选择的信息非常重要，因此了解这些信息如何在神经水平上整合到风险决策中，特别是在处理这些信息时有什么差异，这对公共卫生具有重要意义。 信息导致一些人比其他人更容易受到同伴影响。这项研究的目的是通过让50名14 - 19岁的青少年在扫描仪中执行一项新的真实赌注社交赌博任务来回答这些问题。在这项任务中，参与者将对一系列冒险的赌博做出选择，在每次试验中，他们都会收到关于之前完成任务的四个朋友的错误信息。使用报告的风险接受百分比作为参数调制器，我们将识别大脑中BOLD激活编码他人选择行为信息的区域。其次，我们将根据每个受试者在任务中的选择计算他们的同伴敏感性分数，反映他们的选择在多大程度上受到其他人报告的行为的影响。然后，我们将使用基于机器学习的回归技术来生成回归权重的空间模式，该空间模式可以预测每个参与者对同伴影响的行为敏感性。这些回归权重在大脑不同区域的强度反映了该区域对他人选择的编码强度如何预测个体的同伴易感性得分。这些数据使我们能够：1）识别对他人选择信息的编码模式，这些模式最强烈地预测了对同伴影响的高度敏感性，以及2）生成可以在新的参与者样本中使用的神经特征，以预测他们对同伴影响的敏感性水平。然后，我们将从大脑数据中得出的同伴易感性预测与参与者及其同伴的真实冒险模式进行比较。我们相信，这些研究结果可以为未来的工作奠定基础，研究这些网络的发展变化，以及这些网络中的反应如何通过暴露于不同类型的同龄人或通过旨在降低特定领域对同龄人影响的敏感性的潜在干预措施来改变。 这些研究途径将有助于加速制定更精确和有效的干预战略，以减少在与健康相关的冒险领域受同伴影响的可能性。

 DESCRIPTION (provided by applicant): Risky behaviors such as smoking, drug and alcohol use, and teenaged sexual intercourse are all heavily influenced by the behaviors of one's peers. While much is known about the neural networks involved in risky decision-making, this research has largely been conducted in the absence of the rich network of social information available when making risky decisions in real-life. Given the large body of evidence suggesting the importance of information about peers' choices when making risky decisions, it is of great public-health relevance to understand how the integration of this information into the risky decision takes place at the neural level, and, particularly, what differences in processing of this information lead some people to be more susceptible to peer influence than others. The proposed research aims to answer these questions by having a sample of 50 14-19 year olds perform a novel real-stakes social gambling task in the scanner. In this task, participants will make choices about a series of risky gambles, on each trial receiving false information about the decisions of four of their friends who previously completed the task. Using the reported percentage of risk acceptance as a parametric modulator, we will identify regions in the brain where BOLD activation encodes information about others' choice behavior. Second, we will calculate a peer susceptibility score for each subject based on their choices in the task, reflecting how much their choices were swayed by the reported actions of others. We will then use a machine-learning based regression technique to generate a spatial pattern of regression weights that can predict each participant's behavioral susceptibility to peer influence. The strength of these regression weights in different areas of the brain reflects how strongly encoding of others' choices in that region predicts individuals' peer susceptibility scores. These data allow us to: 1) Identify patterns of encoding of others' choice information that are most strongly predictive of being highly susceptible to peer influence, and 2) Generate a neurologic signature that can be used in new samples of participants to predict their level of susceptibility to peer influence. We will then compare predictions of peer susceptibility derived from the brain data to patterns of real-world risk-taking of the participant and their peers. We believe that thes findings can set the stage for future work examining developmental changes in these networks, and how responses in these networks can be altered by exposure to different types of peers or by potential interventions designed to reduce susceptibility to peer influence in specific domains. These avenues of research will allow for the accelerated development of more precise and effective intervention strategies to reduce susceptibility to peer influence in the domain of publi-health relevant risk-taking.