Background: Cognitive dysfunction, particularly in Alzheimer’s disease (AD), seriously
affects the health and quality of life of older adults. Early detection can prevent and slow cognitive
Objective: This study aimed at evaluating the role of socio-demographic variables, lifestyle, and
physical characteristics in cognitive decline during AD progression and analyzing the probable
causes and predicting stages of the disease.
Methods: By analyzing data of 301 subjects comprising normal elderly and patients with mild cognitive
impairment (MCI) or AD from six communities in Taiyuan, China, we identified the influencing
factors during AD progression by a Logistic Regression model (LR) and then assessed the
associations between variables and cognition using a Bayesian Networks (BNs) model.
Results: The LR revealed that age, sex, family status, education, income, character, depression, hypertension,
disease history, physical exercise, reading, drinking, and job status were significantly
associated with cognitive decline. The BNs model revealed that hypertension, education, job status,
and depression affected cognitive status directly, while character, exercise, sex, reading, income,
and family status had intermediate effects. Furthermore, we predicted probable cognitive stages of
AD and analyzed probable causes of these stages using a model of causal and diagnostic reasoning.
Conclusion: The BNs model lays the foundation for causal analysis and causal inference of cognitive
dysfunction, and the prediction model of cognition in older adults may help the development
of strategies to control modifiable risk factors for early intervention in AD.