Work

P4‐057: Neuropsychological domains in mild cognitive impairment and normal aging: Results from the NACC UDS study

Year: 2009
Type: article
Abstract: Factor analysis is used to determine the relationship between cognitive measures and construct validity in neuropsychological batteries. Correlated items form factors suggesting distinct cognitive domains. Few studies have evaluated how factor loadings change over time within groups of cognitively matched individuals. We used data from the National Alzheimer's Coordinating Center (NACC) to evaluate change over time in two groups, one cognitively normal and the other empirically defined as mildly cognitively impaired. Data were collected at 29 NIA funded Alzheimer's Disease Centers. Each center administers the Unified Data Set (UDS) battery, a standardized set of neuropsychological tests. Two groups were analyzed: a normal group consisted of n=4,038 cognitively normal individuals defined as having a CDR global score=0.0. A subsample of this group, n=2,012 was evaluated one year later. A second group consisted of n=3,391 individuals with mild cognitive impairment (MCI) defined as having a CDR global score=0.5; n=1,310 were evaluated one year later. Factor analyses were performed separately on each group. Baseline factors and weights were determined. Using scores from the second evaluation, followup factor loadings were predicted. Correlations between factor loadings at the first and second time points were then computed. The normal group analysis yielded three primary factors after varimax rotation (executive/language, memory, and attention) which explained 56.8% of the variance. Loadings at each time point were highly correlated. The MCI group analysis yielded five primary factors (verbal retrieval, sensorimotor/executive, auditory attention, auditory working memory, and orientation) which explained 75.2% of the variance. Loadings at each time point in this group were also highly correlated. Factor analyses of the UDS neuropsychological battery performed in a group with normal cognition and one with MCI yielded differing factor structures as might be anticipated. Both groups were remarkably stable over one year. In a range of cognition from normal to MCI, factor loadings and thus constructs measured will vary, reflecting different underlying relationships among cognitive domains. Further work is needed to determine if constructs or domains become more distinct in severely cognitively compromised groups and if there is variation by dementia subtype.
Institution: Duke Medical Center
Language: English
Field: Medicine
Open Access status: bronze
OpenAlex ID: W2068722373
FWCI:
Cites:
Cited by:
Related to: 10