Predicting and Evaluating Cognitive Status in Aging Populations Using Decision Tree Models.

Area of Science:

• Neurology
• Artificial Intelligence
• Biostatistics

Background:

• Cognitive impairment, including normal cognition (NC), mild cognitive impairment (MCI), and Alzheimer's disease (AD), presents a significant public health challenge.
• Accurate identification is crucial for timely intervention and management.

Purpose of the Study:

• To develop and validate a machine learning model for distinguishing between NC, MCI, and AD.
• To identify key cognitive assessment variables for predicting these conditions.

Main Methods:

• A recursive partitioning tree model was trained using ARMADA data and NIH Toolbox assessments.
• Demographic and clinical variables were incorporated to predict cognitive status.
• Model performance was evaluated using AUC, precision, recall, F1 score, and cross-validation.

Main Results:

• The model achieved high performance metrics, with macro-AUC of 0.92 (training) and 0.89 (testing).
• Key predictors included the Picture Sequence Memory Test and List Sorting Working Memory Test.
• Cross-validation demonstrated 70.22% accuracy and a Kappa of 0.52.

Conclusions:

• Machine learning models can effectively differentiate between NC, MCI, and AD using a limited set of assessments.
• This approach offers a valuable, data-driven tool to support clinical decision-making in cognitive impairment diagnosis.
• Further validation across diverse populations is recommended to enhance generalizability.