Abstract：
OBJECTIVE: Dispersion is a form of intra-individual variability across neuropsychological tests that has been shown to predict cognitive decline. However, few studies have investigated the stability and predictive utility of both across- and within-domain dispersion. The current study aims to fill these gaps in the literature by examining multiple indices of dispersion in a longitudinal clinical sample of individuals diagnosed with mild cognitive impairment (MCI) at baseline.
METHODS: Two hundred thirty-eight MCI patients from a cognitive disorders clinic underwent testing at baseline and after approximately 1.5 years. Linear regression was used to examine whether baseline across- and within-domain dispersion predicted cognitive decline in individuals whose diagnostic classification progressed to dementia (i.e., MCI-Decline) and those who retained an MCI diagnosis at follow-up (i.e., MCI-Stable). Cognitive decline was operationalized dichotomously using group status and continuously using standardized regression-based (SRB) z-scores.
RESULTS: Dispersion variables at baseline and follow-up were positively correlated in both groups, with the exception of within-domain executive functioning and language dispersion in the MCI-Decline group. None of the dispersion variables predicted diagnostic conversion to MCI. Using SRB z-scores, greater across-domain dispersion predicted greater overall cognitive decline at follow-up, but this was not the case for within-domain variables with the exception of visuospatial skills.
CONCLUSIONS: Results suggest that across- and within-domain dispersion are relatively stable across time, and that across-domain dispersion is predictive of subtle cognitive decline in patients with MCI. However, these results also highlight that findings may differ based on the tests included in dispersion calculations.
METHODS: Two hundred thirty-eight MCI patients from a cognitive disorders clinic underwent testing at baseline and after approximately 1.5 years. Linear regression was used to examine whether baseline across- and within-domain dispersion predicted cognitive decline in individuals whose diagnostic classification progressed to dementia (i.e., MCI-Decline) and those who retained an MCI diagnosis at follow-up (i.e., MCI-Stable). Cognitive decline was operationalized dichotomously using group status and continuously using standardized regression-based (SRB) z-scores.
RESULTS: Dispersion variables at baseline and follow-up were positively correlated in both groups, with the exception of within-domain executive functioning and language dispersion in the MCI-Decline group. None of the dispersion variables predicted diagnostic conversion to MCI. Using SRB z-scores, greater across-domain dispersion predicted greater overall cognitive decline at follow-up, but this was not the case for within-domain variables with the exception of visuospatial skills.
CONCLUSIONS: Results suggest that across- and within-domain dispersion are relatively stable across time, and that across-domain dispersion is predictive of subtle cognitive decline in patients with MCI. However, these results also highlight that findings may differ based on the tests included in dispersion calculations.
摘要：
目的：离散是神经心理学测试中个体内变异性的一种形式，已被证明可以预测认知能力下降。然而,很少有研究研究了跨域和域内分散的稳定性和预测效用。本研究旨在通过检查基线诊断为轻度认知障碍（MCI）的个体的纵向临床样本中的多个分散指数来填补文献中的这些空白。
方法：来自认知障碍诊所的238名MCI患者在基线和大约1.5年后接受了测试。线性回归用于检查基线跨域和域内分散是否预测诊断分类进展为痴呆的个体的认知下降（即，MCI-下降)和那些在随访中保留MCI诊断的人(即，MCI稳定）。使用组状态和连续使用基于标准化回归（SRB）的z评分将认知下降二分法进行操作。
结果：两组基线和随访时的离散变量呈正相关，除了MCI-Decline组中的域内执行功能和语言分散。没有一个离散变量预测诊断转化为MCI。使用SRBz分数，更大的跨域离散度预测随访时更大的整体认知能力下降，但对于领域内变量，情况并非如此，除了视觉空间技能。
结论：结果表明，整个域和域内分散在时间上相对稳定，并且跨域离散度可以预测MCI患者的细微认知功能下降。然而,这些结果还强调,根据离散度计算中包含的测试,结果可能会有所不同.
方法：来自认知障碍诊所的238名MCI患者在基线和大约1.5年后接受了测试。线性回归用于检查基线跨域和域内分散是否预测诊断分类进展为痴呆的个体的认知下降（即，MCI-下降)和那些在随访中保留MCI诊断的人(即，MCI稳定）。使用组状态和连续使用基于标准化回归（SRB）的z评分将认知下降二分法进行操作。
结果：两组基线和随访时的离散变量呈正相关，除了MCI-Decline组中的域内执行功能和语言分散。没有一个离散变量预测诊断转化为MCI。使用SRBz分数，更大的跨域离散度预测随访时更大的整体认知能力下降，但对于领域内变量，情况并非如此，除了视觉空间技能。
结论：结果表明，整个域和域内分散在时间上相对稳定，并且跨域离散度可以预测MCI患者的细微认知功能下降。然而,这些结果还强调,根据离散度计算中包含的测试,结果可能会有所不同.
