A MIMIC Model Approach to Research in Geriatric Neuropsychology: The Case of Vascular Dementia

Abstract

The goal of the current study is to demonstrate a new methodology that can be used in neuropsychological research concerning differential diagnosis research. The multiple indicators, multiple causes (MIMIC) model is a latent variable methodology which can examine group differences on individual tests while controlling group differences in global cognitive impairment. As a demonstration, neuropsychological data from 217 dementia patients were incorporated into a MIMIC model to examine the influence of cerebrovascular disease (CVD) upon (1) dimensions of global cognitive impairment and (2) upon individual tests after controlling for global impairment. The presence of CVD in dementia (i.e., vascular dementia [VaD]) was not significantly related to dimensions of global impairment. In addition, the presence of CVD within dementia did not significantly contribute to impairment on 9 out of 10 neuropsychological tests/subscales examined after controlling global cognitive impairment. These results are discussed in the context of current vascular dementia research, and are focused primarily upon the MIMIC model methodology and suggestions for its use in future research.