Developing learning slope scores for the repeatable battery for the assessment of neuropsychological status
Developing learning slope scores for the repeatable battery for the assessment of neuropsychological status
复制标题
DOI：
10.1080/23279095.2020.1791870
复制
发表时间：
2020-07-10
影响因子：
1.7
通讯作者：
Hammers, Dustin B.
中科院分区：
文献类型：
--
作者：
Spencer, Robert J.;Gradwohl, Brian D.;Hammers, Dustin B.
初始学习和学习斜率通常被认为是学习的重要定性方面，但用于评估神经心理状态的可重复组合（RBANS）包含两个独立的指标。传统的计算学习斜率的方法涉及最后一次试验和第一次试验之间的差异得分，该差异得分被称为原始学习得分（RLS）。然而，这种方法并没有考虑到最初的试验一的性能，并产生天花板效应，惩罚有效的第一个学习者。我们提出了一种计算学习分数的替代方法，该方法考虑了初始学习性能，称为学习比（LR），我们比较了这些方法的心理测量和预测特性。从列表学习和故事记忆子测试的性能被用来创建指数，并通过组合列表学习和故事记忆计算综合学习分数。样本包括289名退伍军人（平均年龄= 65.9 [12.6]，教育= 13.3 [2.4]），其中大多数是男性，接受包括RBANS在内的神经心理学评估。结果表明，与RLS相比，LR表现出与记忆标准测量值的上级相关性，LR综合评分AUC = 0.81（0.76-0.87）比RLS综合评分AUC = 0.70（0.64-0.76）更好地区分有和无神经认知诊断的患者。我们的结论是，从RBANS的分数可以计算初始学习和学习斜率，LR方法计算学习是上级RLS在这个老老兵样本。
Initial learning and learning slope are often acknowledged as important qualitative aspects of learning, but the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) contains discrete indices for neither. The traditional method of calculating learning slope involves a difference score between the last trial and first trial, which is referred to as raw learning score (RLS). However, this method does not account for initial Trial One performance and produces a ceiling effect that penalizes efficient first learners. We propose an alternative method of calculating learning score that accounts for initial learning performance, called learning ratio (LR), and we compared the psychometric and predictive properties of these methods. Performances from the List Learning and Story Memory subtests were used to create the indices, and composite learning scores were calculated by combining List Learning and Story Memory. The sample included 289 military veterans (mean age = 65.9 [12.6], education = 13.3 [2.4]), most of whom were male, undergoing neuropsychological assessments that included the RBANS. Results indicated that LR demonstrated superior correlations with criterion measures of memory when compared with RLS, and the LR composite score better discriminated between those with and without a neurocognitive diagnosis, AUC = 0.81 (0.76-0.87), than the RLS composite, AUC = 0.70 (0.64-0.76). We concluded that scores from the RBANS can be computed for initial learning and learning slope and that the LR method of calculating learning is superior to RLS in this older veteran sample.