Repeated Stochastic Processes in Single Trial EEG-Analysis - Statistical Theory and Methods

单次试验脑电图分析中的重复随机过程 - 统计理论和方法

• 批准号: 215899693
• 负责人: Professor Dr. Jan Beran
• 金额: --
• 依托单位: Arbeitsgruppe Prof. Dr. Beran
• 依托单位国家: 德国
• 项目类别: Research Units
• 财政年份: 2012
• 资助国家: 德国
• 起止时间: 2011-12-31 至 2015-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/215899693?language=en
• 关键词: Repeated; Stochastic; Processes; Single; Trial

There is a growing interest among psychologists and behavioural economists in data generated by repeated observations of a stochastic process. Projects TP2, TP6 aim at answering complex questions in the context of dual processing (automatic and controlled processes) based on data of this type. The standard approach of averaging ERPs usually requires a large number of homogeneous trials. This can be problematic when dealing, for instance, with microeconomic decisions where the number of available repetitions is often small. Moreover, trial averaging ignores variability of automatic and controlled processes across experimental trials. Single-trial analysis is therefore the appropriate approach in the context of this RU. Since the data structure in single trials is more complex and available data sets are relatively small, sophisticated and parsimonious statistical methodology is needed to extract the essential information. In this project this methodology will be developed by considering functional data analysis (FDA) and independent component analysis (ICA) in the context of suitably defined stochastic processes. The data-analytical focus will be on EEG data (ERPs) and their predictive power in modelling human decision making (TP2, TP6). The statistical methods will in corporate temporal and/or spatial dependence structures within the realm of FDA and ICA. This will provide more insight into the process of decision making and its neural correlates

心理学家和行为经济学家对通过反复观察随机过程产生的数据的兴趣越来越大。项目TP2，TP6旨在根据此类数据在双重处理（自动和受控过程）的背景下回答复杂问题。平均ERP的标准方法通常需要大量同质试验。例如，处理可用重复数量通常很少的微观经济决策时，这可能是有问题的。此外，试验平均忽略了在实验试验中自动和受控过程的可变性。因此，在此RU的背景下，单审判分析是适当的方法。由于单个试验中的数据结构更为复杂，并且可用的数据集相对较小，因此需要进行复杂且需要简约的统计方法来提取基本信息。在此项目中，将在适当定义的随机过程中考虑功能数据分析（FDA）和独立组件分析（ICA）来开发该方法。数据分析的重点将放在脑电图数据（ERP）及其在建模人类决策时的预测能力（TP2，TP6）。统计方法将在FDA和ICA领域内的公司时间和/或空间依赖结构中进行。这将为决策过程及其神经相关性提供更多的了解