CRCNS: Neurocomputational Study of Reward-Related Decision-Making & Uncertainty

CRCNS：奖励相关决策的神经计算研究

• 批准号: 9916054
• 负责人: Angela Yu
• 金额: $ 23.85万
• 依托单位: UNIVERSITY OF CALIFORNIA, SAN DIEGO
• 依托单位国家: 美国
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-09-30 至 2022-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9916054
• 关键词: Address; Affect; Alcohol or Other Drugs use; Anhedonia; Animals; Anxiety; Area; Bayesian Method; Behavior; Behavioral; Brain; Brain region; Cardiac; Choice Behavior; Cognitive; Complex; Control Locus; Data; Decision Making; Detection; Disease; Drug abuse; Drug usage; Environment; Functional Magnetic Resonance Imaging; Goals; Human; Individual Differences; Instruction; Learning; Life; Measures; Mediation; Modeling; Monitor; Motivation; Neurobiology; Neurocognitive; Neurosciences; Outcome; Physiological; Play; Prevention; Probability; Process; Psychological reinforcement; Psychology; Questionnaires; Research; Rewards; Signal Transduction; Social Interaction; Structure; Substance Addiction; Substance abuse problem; System; Uncertainty; Update; Weather; Work; base; control theory; depressive symptoms; experimental study; ideal observer (Bayesian); innovation; neuroregulation; relating to nervous system; respiratory; response; theories; trait

Humans and animals often make decisions under uncertainty, whereby each decision affects not only the immediate reward gain but also longer-term information gain. While important advances have been made in understanding human learning and decision-making, there is still a lack of understanding of the different motivational factors that come into play when the behavioral context confers systematically varying amounts of reward and information gain. This project tackles this problem using a combination of sophisticated cognitive modeling, innovative behavioral experiments, fMRI data, physiological (pupillometry, cardiac, and respiratory) data, and psychiatric measures (questionnaires addressing depressiveness, anxiety, anhedonia, locus of control, pessimism, and substance abuse). The objectives are (1) to develop a statistically grounded and neurobiologically informed theory for how different motivational factors (immediate reward, long-term reward, reduction of uncertainties, and random stochasticity) jointly influence human decision making; (2) use this theoretical framework to guide the understanding of how different brain regions, in particular neuromodulatory systems, work separately and conjointly to implement behavioral choices in response to the reward and informational structure of the environment; (3) characterize individual differences in terms of motivations, subjective monitoring of uncertainties, neural and physiological responses, and psychiatric profile. This work builds on multiple theoretic approaches: Bayesian ideal observer, reinforcement learning, Markov decision process, and control theory; and multiple neuroscientific research areas: learning, information seeking, confidence, decision making, change-point detection. It will advance an integrated understanding of computational theory, neuro-cognitive processes, behavioral manifestations, physiological signals, and psychiatric traits in choice behavior under uncertainty. It will help to clarify how different cortical and subcortical (especially neuromodulatory) brain regions differentially and cooperatively contribute to reward- and information-based learning, decision making, and exploration. These outcomes can be expected to contribute to advancements in basic scientific understanding of brain circuits, mechanisms, and functions related to the use and abuse of addictive substances, as well as their prevention and treatment. RELEVANCE (See instructions): Drug use and abuse often involve alterations in reward learning, decision-making, and uncertainty-related processing. This project contributes to basic computational and neurobiological understanding of these processes in the healthy brain, and may help to elucidate how these processes go awry in substance use and addiction disorders.

人类和动物经常在不确定的情况下做出决定，因此每个决定不仅会影响 即时的奖励收益，以及长期的信息收益。虽然已经取得了重要进展 在理解人类学习和决策方面，仍然缺乏对不同的 当行为环境赋予系统性变化时， 奖励和信息收益。该项目使用以下组合来解决此问题： 复杂的认知建模，创新的行为实验，功能磁共振成像数据，生理 （瞳孔测量，心脏和呼吸）数据和精神病学措施（问卷调查， 抑郁、焦虑、快感缺乏、控制点、悲观和药物滥用）。目标 （1）发展一个基于统计学和神经生物学的理论， 激励因素（即时奖励，长期奖励，减少不确定性，随机 随机性）共同影响人类决策;（2）使用这个理论框架来指导 了解不同的大脑区域，特别是神经调节系统如何单独工作， 联合实施行为选择，以响应奖励和信息结构的 环境;（3）在动机方面描述个体差异， 不确定性、神经和生理反应以及精神状况。这项工作建立在多个 理论方法：贝叶斯理想观测器，强化学习，马尔可夫决策过程， 控制理论;和多个神经科学研究领域：学习，信息寻求，信心， 决策，变化点检测。它将促进对计算的综合理解， 理论、神经认知过程、行为表现、生理信号和精神特征 在不确定性下的选择行为。这将有助于澄清不同的皮质和皮质下（特别是 神经调节）脑区差异和合作有助于奖励-和 基于信息的学习、决策和探索。这些成果预计将 有助于促进对大脑回路、机制和功能的基本科学理解 与成瘾物质的使用和滥用有关的问题，以及预防和治疗成瘾物质的问题。 相关性（参见说明）： 药物使用和滥用通常涉及奖励学习，决策和不确定性相关的改变。 处理.这个项目有助于对这些基本的计算和神经生物学的理解 这可能有助于阐明这些过程在物质使用中是如何出错的 和成瘾性障碍。