Combining pathological and cognitive tests scores: A novel data analytics process to improve dementia prediction...

Area of Science:

• Neuroscience
• Computer Science
• Medical Diagnostics

Background:

• Dementia is a progressive brain disease affecting many elderly individuals.
• Traditional diagnostic methods for dementia lack optimal accuracy and efficiency.
• Artificial Intelligence (AI) applications show promise in improving dementia detection.

Purpose of the Study:

• To investigate the effectiveness of supervised learning, a data analytics technique, for detecting Alzheimer's disease (AD).
• To develop predictive classification systems for dementia detection by evaluating cognitive tests and biomarkers.
• To enhance the accuracy and efficiency of diagnosing dementia and its subtypes.

Main Methods:

• Utilized supervised learning algorithms for data analysis.
• Evaluated cognitive test scores and cerebrospinal fluid (CSF) biomarkers.
• Developed and tested predictive classification models.
• Applied the data analytics process to real-world data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) repository.

Main Results:

• The developed models demonstrated significant predictive power for AD levels.
• Cognitive test scores were identified as key features for prediction.
• Tauopathy-related biological markers also proved valuable in prediction.
• The AI models achieved improved detection accuracy and efficiency compared to traditional methods.

Conclusions:

• Supervised learning models are effective in predicting Alzheimer's disease.
• Cognitive assessments and specific biomarkers are crucial for accurate AD prediction.
• AI-driven approaches offer a promising future for early and precise dementia diagnosis.