ABSTRACT

Reversal learning is frequently used to assess components of executive function that contribute to understanding age-related cognitive differences. Reaction time (RT) is less characterized in the reversal learning literature, perhaps due to the daunting task of analyzing the entire RT distribution, but has been deemed a generally sensitive measure of cognitive aging. The current study extends our prior work to further characterize distributional properties of the reversal RT distribution and to distinguish groups of individuals with fractionated profiles of performance, which may be of clinical importance within the context of cognitive aging. Participant sample included young (n = 43) and community-dwelling, healthy, middle-aged (n = 139) adults. To explore individual differences, recursive partitioning analysis achieved a high classification rate by specifying decision tree rules that split participants into young and middle-aged groups. Mu (μ, efficient RT) was the most successful parameter in distinguishing age groups while sigma ($\sigma )$ and tau ($\tau$, ex-Gaussian indices of intra-individual variability) revealed more subtle individual differences. Accuracy measures did not contribute to separating the groups, suggesting that fractionated components of RT, as opposed to accuracy, can distinguish differences between young and middle-aged participants.

摘要

逆向学习常用于评估执行功能的组成部分，这些组成部分有助于理解与年龄相关的认知差异。在逆向学习文献中，反应时间（RT）的特征较少，这可能是由于分析整个RT分布的艰巨任务所致，但它被认为是认知衰老的一般敏感措施。当前的研究扩展了我们先前的工作，以进一步表征逆向RT分布的分布特征，并区分具有分级表现的个体群体，这在认知衰老的背景下可能具有临床重要性。参与者样本包括年轻人（n = 43）和社区居民，健康，中年（n= 139）成人。为了探索个体差异，递归分区分析通过指定将参与者分为年轻和中年组的决策树规则，实现了很高的分类率。Mu（μ，有效RT）是区分年龄组的最成功参数，而sigma（$\sigma ）$ 和tau（$\tau$，前个体内变异性的高斯指数）显示出更细微的个体差异。准确度测量并不能帮助区分人群，这表明RT的分数成分与准确度相反，可以区分年轻和中年参与者之间的差异。