Mungas Dan, Reed Bruce R, Kramer Joel H
Department of Neurology, School of Medicine, University of California, Davis, Sacramento 95817, USA.
Neuropsychology. 2003 Jul;17(3):380-92. doi: 10.1037/0894-4105.17.3.380.
Item response theory methods were used to derive psychometrically sophisticated measures of global cognition, memory, and executive function. Goals were that these measures (a) could be derived from commonly used neuropsychological tests, (b) would have linear measurement properties, and (c) would be psychometrically matched. Scale development was based on a sample of 400 older individuals with cognitive function ranging from normal to demented. Scales were reasonably matched with linear measurement over an ability range relevant to many important clinical applications. Cognitively normal, mild impairment, and dementia participant groups differed on baseline measures and rate of decline. Association of measures with quantitative structural magnetic resonance imaging variables followed expected patterns. This approach to scale development may have applications for other neuropsychological assessment problems.
项目反应理论方法被用于得出在心理测量学上更为精细的整体认知、记忆和执行功能测量指标。目标是这些测量指标：（a）可以从常用的神经心理学测试中得出；（b）具有线性测量属性；（c）在心理测量学上相互匹配。量表开发基于400名认知功能从正常到痴呆的老年人样本。在与许多重要临床应用相关的能力范围内，量表与线性测量合理匹配。认知正常、轻度认知障碍和痴呆参与者组在基线测量和衰退率方面存在差异。测量指标与定量结构磁共振成像变量之间的关联遵循预期模式。这种量表开发方法可能适用于其他神经心理学评估问题。
