Examining Cognitive Factors for Alzheimer's Disease Progression Using Computational Intelligence

Area of Science:

• Neurology
• Artificial Intelligence
• Cognitive Science

Background:

• Alzheimer's disease (AD) progression prediction is complex due to numerous patient features.
• While diagnosis using pathological markers is common, predicting AD advancement via cognitive elements, especially with AI, is less explored.

Purpose of the Study:

• To evaluate Alzheimer's Disease Assessment Scale-Cognitive 13 (ADAS-Cog-13) items to identify key cognitive factors influencing AD progression.
• To apply machine learning and feature selection techniques for predicting AD advancement using cognitive data.

Main Methods:

• A methodology combining machine learning and feature selection was developed.
• The approach was tested on data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) repository, focusing on ADAS-Cog-13 cognitive items.
• Ten-fold cross-validation and various classification/feature selection techniques were employed.

Main Results:

• The decision tree algorithm yielded the best classification models for predicting AD progression using cognitive items.
• Key predictors of AD advancement identified were memory and learning features: word recall, delayed word recall, and word recognition.
• Using the C4.5 algorithm with these three cognitive items (excluding demographics) resulted in 82.90% accuracy, 87.60% sensitivity, and 78.20% specificity.

Conclusions:

• Memory and learning components of the ADAS-Cog-13 test are crucial for predicting Alzheimer's disease progression.
• Machine learning, particularly decision tree algorithms, can effectively identify cognitive markers for AD advancement.
• The findings highlight the potential of specific cognitive tests in forecasting disease trajectory, aiding in early intervention strategies.