Functional impairment in daily activity is a cornerstone in distinguishing the clinical progression of dementia. Multiple indicators based on neuroimaging and neuropsychological instruments are used to assess the levels of impairment and disease severity; however, it remains unclear how multivariate patterns of predictors uniquely predict the functional ability and how the relative importance of various predictors differs.

In this study, 881 older adults with subjective cognitive complaints, mild cognitive impairment (MCI), and dementia with Alzheimer’s type completed brain structural magnetic resonance imaging (MRI), neuropsychological assessment, and a survey of instrumental activities of daily living (IADL). We utilized the partial least square (PLS) method to identify latent components that are predictive of IADL.

The result showed distinct brain components (gray matter density of cerebellar, medial temporal, subcortical, limbic, and default network regions) and cognitive–behavioral components (general cognitive abilities, processing speed, and executive function, episodic memory, and neuropsychiatric symptoms) were predictive of IADL. Subsequent path analysis showed that the effect of brain structural components on IADL was largely mediated by cognitive and behavioral components. When comparing hierarchical regression models, the brain structural measures minimally added the explanatory power of cognitive and behavioral measures on IADL.

Our finding suggests that cerebellar structure and orbitofrontal cortex, alongside with medial temporal lobe, play an important role in the maintenance of functional status in older adults with or without dementia. Moreover, the significance of brain structural volume affects real-life functional activities via disruptions in multiple cognitive and behavioral functions.

日常活动中的功能障碍是区分痴呆临床进展的基石。基于神经影像学和神经心理学仪器的多项指标用于评估损伤程度和疾病严重程度；然而，目前尚不清楚预测因子的多元模式如何独特地预测功能能力以及各种预测因子的相对重要性如何不同。

在这项研究中，881 名患有主观认知障碍、轻度认知障碍 (MCI) 和阿尔茨海默病型痴呆的老年人完成了脑结构磁共振成像 (MRI)、神经心理学评估和日常生活工具活动 (IADL) 调查。我们利用偏最小二乘 (PLS) 方法来识别可预测 IADL 的潜在成分。

结果显示不同的大脑成分（小脑、内侧颞叶、皮质下、边缘和默认网络区域的灰质密度）和认知行为成分（一般认知能力、处理速度和执行功能、情景记忆和神经精神症状） IADL 的预测。随后的路径分析表明，大脑结构成分对 IADL 的影响主要由认知和行为成分介导。在比较分层回归模型时，大脑结构测量最小地增加了认知和行为测量对 IADL 的解释力。

我们的研究结果表明，小脑结构和眶额皮质以及内侧颞叶在患有或不患有痴呆症的老年人的功能状态维持中发挥着重要作用。此外，大脑结构体积的重要性通过破坏多种认知和行为功能来影响现实生活中的功能活动。