Neuroplasticity-Based Treatment to Address State Representation Failures in People with Early Psychosis

项目概要：项目4 项目4的目的是研究计算信息的精确处理，以改善两个 早期精神病的状态表征功能障碍的形式：1）知觉的状态估计过程 输入水平，我们将通过听觉辨别训练的目标; 2）状态表征的稳定性 听觉信息，我们将通过听觉工作记忆训练的目标。参加者将被抽中 在项目3中，他们将在基线时接受行为和EEG功能磁共振成像评估 经过6个月的常规护理，使他们的初始特征和临床轨迹将 知道的将根据状态估计过程的EEG指数（额顶叶θ）对参与者进行分层 DPX编码的功率），我们认为约60%的受试者存在这种情况，并随机分配到 两种训练策略。我们的目标不是进行治疗效果的研究比较这两个 干预措施。相反，我们试图使用来自吸引子网络模型的预测来测试 基于神经可塑性的精确治疗，针对两种不同的信息处理病理， 精神病，最终目标是改善状态表征过程和认知。 在目标1中，我们将研究每个受试者组中拟合吸引子网络模型的参数变化，拟合到 DPX和Bandit Task行为数据在训练后立即和3个月后，我们将评估 参数变化是否反映了恢复性或补偿性修改。我们还将检验假设 陈述表征过程和认知表现在以下受试者中表现出更大的改善： 接受针对其状态估计参数的训练。在目标2中，我们将研究如何具体 吸引子网络模型中的参数变化与测量索引中的神经生理变化有关 活动时间，兴奋-抑制平衡和系统噪声，以确定哪些变化是最重要的 预示着认知能力的提高。将采用因果发现分析来确定因果关系 在计算参数，行为数据，神经生理学指标，治疗分配，和一个- 年临床轨迹。

PROJECT SUMMARY: PROJECT 4 The purpose of PROJECT 4 is to investigate computationally-informed precision treatments to improve two forms of state representation dysfunction in early psychosis: 1) State estimation processes at the perceptual input level, which we will target through auditory discrimination training; 2) State representation stability of auditory information, which we will target through auditory working memory training. Participants will be drawn from PROJECT 3, where they will have been assessed with behavioral and EEG-fMRI measures at baseline and after 6 months of usual care, so that their initial characteristics and clinical trajectory will be known. Participants will be stratified on an EEG index of state estimation processes (fronto-parietal theta power at DPX encoding), which we posit to be present in ~60% of subjects, and randomly assigned to one of the two training strategies. Our goal is not to perform a treatment efficacy study comparing these two interventions. Rather, we seek to use predictions derived from attractor network models to test the effects of neuroplasticity-based precision treatments targeting two distinct information processing pathologies in early psychosis, with the ultimate goal of improving state representation processes and cognition. In Aim 1, we will investigate parameter changes in the fit attractor network models in each subject group, fit to DPX and Bandit Task behavioral data immediately after training and 3 months later, and we will assess whether parameter changes reflect restorative or compensatory modifications. We will also test the hypothesis that state representation processes and cognitive performance show greater improvement in subjects who received training tailored to their state estimation parameter. In Aim 2, we will examine how specific parameter changes in attractor network models relate to neurophysiological changes in measures indexing activity timing, excitatory-inhibitory balance, and system noise, in order to identify which changes are the most predictive of improved cognition. Causal discovery analyses will be employed to identify causal relationships among computational parameters, behavioral data, neurophysiologic indices, treatment assignment, and one- year clinical trajectories.