Prediction Error Signaling and Reinforcement Learning in Schizophrenia

精神分裂症的预测误差信号和强化学习

• 批准号: 8263410
• 负责人: Erin Connor Dowd
• 金额: $ 4.72万
• 依托单位: WASHINGTON UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-07-01 至 2013-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8263410
• 关键词: Activities of Daily Living; Address; Anhedonia; Antipsychotic Agents; Area; Attenuated; Basal Ganglia; Behavior; Behavioral; Brain; Clinical; Code; Computer Simulation; Corpus striatum structure; Cues; Data; Development; Disease; Dopamine; Dopamine D1 Receptor; Dopamine D2 Receptor; Dopamine Receptor; Dose; Electrophysiology (science); Feedback; Functional Magnetic Resonance Imaging; Future; Goals; Impairment; Individual; Individual Differences; Learning; Measures; Mediating; Modeling; Motivation; Negative Reinforcements; Neural Network Simulation; Operant Conditioning; Outcome; Pathway interactions; Patient Self-Report; Patients; Performance; Pharmaceutical Preparations; Positive Reinforcements; Psychological reinforcement; Quality of life; Relative (related person); Resistance; Rewards; Schizophrenia; Severities; Signal Transduction; Stimulus; Symptoms; System; Task Performances; Testing; Time; Translating; Translations; Work; behavioral impairment; dosage; experience; hedonic; improved; interest; mesolimbic system; neuroimaging; performance tests; pleasure; relating to nervous system; response; reward processing

DESCRIPTION (provided by applicant): The long-term objective of this proposal is to understand the relationship between altered processing of rewards and symptoms of anhedonia (a reduced experience of pleasure) and amotivation in individuals with schizophrenia. Motivational impairments are critical features of schizophrenia that significantly impact functional capacity and are resistant to treatment. A growing body of data suggests that while in-the-moment hedonic experience is intact in schizophrenia, patients may be impaired in their ability to translate rewarding experiences into future goal-directed behavior. Essential to this ability is a capacity for reinforcement learning, which strengthens actions that result in rewarding outcomes and suppresses those that do not. This proposal aims to determine whether individuals with schizophrenia show impairments in reinforcement learning and its related neural activity, and whether individual differences in these impairments are related to symptoms of anhedonia and amotivation. Converging data from electrophysiology, neuroimaging, and computational modeling implicates the mesolimbic dopamine system in reinforcement learning, suggesting that it codes reward prediction errors that gradually integrate outcomes over several trials. This system is of particular interest in schizophrenia, given evidence of altered dopamine release in the striatum in patients with this illness. The work proposed here uses fMRI in conjunction with a computational model of reinforcement learning to examine prediction error-related neural activity during an instrumental learning paradigm in patients and controls. If prediction error signaling is disrupted in schizophrenia, patients would be expected to show reduced prediction error-related neural activity in mesolimbic areas such as the striatum, as well as behavioral impairments in reinforcement learning. If these disruptions in prediction error signaling contribute to symptoms of anhedonia and amotivation, patients who are higher in these symptoms would be expected to show larger reductions in prediction error activity and poorer task performance. Furthermore, because antipsychotic medications that block dopamine receptors in the striatum may disrupt reinforcement learning in schizophrenia, an additional aim of this proposal is to examine the relationship between individual differences in prediction error signaling/task performance and antipsychotic type and dosage. If dopamine receptor antagonism interferes with reinforcement learning, patients experiencing higher levels of dopamine receptor antagonism would be expected to show attenuated prediction error signaling and impaired task performance. An improved understanding of the relationship between dopaminergic prediction errors and reinforcement learning in schizophrenia, as well as their relationship to clinical symptoms and potential medication effects, may contribute to the development of targeted therapies to address these clinically important but currently under- treated symptoms.

描述（由申请人提供）：本提案的长期目标是了解精神分裂症患者奖励处理的改变与快感缺乏症状（快乐体验减少）和动机丧失之间的关系。动机障碍是精神分裂症的关键特征，显著影响功能能力，并且对治疗有抵抗力。越来越多的数据表明，虽然精神分裂症患者的即时享乐体验是完整的，但患者将奖励体验转化为未来目标导向行为的能力可能会受损。这种能力的关键是强化学习的能力，它加强了导致奖励结果的行动，并抑制了那些没有的行动。该提案旨在确定精神分裂症患者是否在强化学习及其相关神经活动中表现出损伤，以及这些损伤的个体差异是否与快感缺失和动力丧失的症状有关。来自电生理学、神经成像和计算建模的融合数据暗示了中脑边缘多巴胺系统在强化学习中的作用，这表明它编码了奖励预测错误，这些错误逐渐整合了几次试验的结果。这一系统在精神分裂症中特别令人感兴趣，因为有证据表明患有这种疾病的患者纹状体中的多巴胺释放发生了改变。这里提出的工作使用功能磁共振成像结合强化学习的计算模型，以检查预测错误相关的神经活动在患者和对照组的工具学习范式。如果预测错误信号在精神分裂症中被破坏，预计患者将在中脑边缘区域（如纹状体）显示出减少的预测错误相关神经活动，以及强化学习中的行为障碍。如果预测错误信号的这些中断导致快感缺乏和动力不足的症状，则这些症状较高的患者预计会显示预测错误活动的较大减少和较差的任务表现。此外，由于阻断纹状体多巴胺受体的抗精神病药物可能会破坏精神分裂症的强化学习，本提案的另一个目的是检查预测错误信号/任务表现与抗精神病药物类型和剂量之间的个体差异。如果多巴胺受体拮抗作用干扰强化学习，那么经历更高水平的多巴胺受体拮抗作用的患者预计会显示出减弱的预测错误信号和受损的任务表现。对精神分裂症中多巴胺能预测错误和强化学习之间的关系以及它们与临床症状和潜在药物作用之间的关系的更好理解，可能有助于开发靶向治疗来解决这些临床重要但目前治疗不足的症状。