Brain volumes, cognitive dysfunction, and socioeconomic status in adults with and without Alzheimer’s disease: an exploratory data analysis
Alberto Guevara-Tirado, School of Human Medicine, Universidad Científica del Sur, Lima, Peru
Background: Neurodegeneration in Alzheimer’s dementia could be correlated with cognitive manifestations and socioeconomic factors. Objective: To analyze the correlation between intracranial and brain volume and the socioeconomic and cognitive states of adults with and without Alzheimer’s disease (AD). Method: An analytical and cross-sectional study of secondary data from the open access imaging studies collection (OASIS), made up of 146 adults with dementia and 190 without dementia. The variables were sex, age, intracranial volume (eTIV), normalized brain volume (nWBV), educational level, socioeconomic status, mini mental health test (MMSE), and clinical dementia classification (CDR). The students’ t-test, Mann-Whitney U, Spearman correlation, and multilayer perceptron were used. Results: Mean years of education, socioeconomic status, eTIV, and nWBV were lower in adults with AD. In adults without AD, nWBV was moderately correlated with age (Rho: –0.581) and socioeconomic status (Rho: 0.235). In adults with AD, nWBV was moderately correlated with age (Rho: –0.406), slightly with CDR (Rho: –0.285), moderately with MMSE (Rho: 0.401), while eTIV was moderately correlated with socioeconomic status (Rho: 0.467) and years of study (Rho: 0.409). The multilayer perceptron correctly predicted 91.50% of AD cases and correctly ruled out 74.80% of them. Its area under the curve was 0.862, indicating an acceptable classification model. Conclusions: In adults with AD, brain volume correlates with cognitive function, while eTIV correlates with socioeconomic status and years of study. Brain and intracranial volume, socioeconomic status, years of study, and sex were predictors of dementia using neural network analysis.
Keywords: Alzheimer disease. Cerebrum. Mental status and dementia tests. Socioeconomic factors. Correlation of data.
