Prediction of Mild Cognitive Impairment in Older Adults

This dataset consists of 845 cognitively healthy Spanish individuals, aged 65 to 87 years at baseline, all living independently at home and free from significant psychiatric, neurological, or systemic disorders. The data were collected with the aim of improving the early detection and prevention of mild cognitive impairment (MCI) and dementia. To achieve this, all participants underwent a comprehensive assessment protocol, typically completed within four hours, with appropriate breaks provided.

The full assessment included a semi-structured clinical interview, as well as neurological and neuropsychological evaluations. Consequently, the dataset contains information on 219 variables, organized into four categories: Sociodemographic, Self-reported, Medical Examination, and Cognitive Assessment.

This dataset was utilized to develop four eXtreme Gradient Boosting (XGBoost) models of increasing complexity. The models were trained and evaluated using robust preprocessing techniques, including multiple imputation for handling missing data and the Synthetic Minority Oversampling Technique (SMOTE) for class balancing. Three versions of the dataset are provided here: the original dataset, the dataset with multiple imputations, and the balanced dataset.