Prediction of MMSE Scores for Cognitive Impairment

Prediction of MMSE Scores for Cognitive Impairment: A Machine Learning Analysis of Oral Health and Demographic Data in Individuals Over 60 Years of Age

This study aims to explore the potential of using machine learning (ML) algorithms to predict cognitive status, specifically MMSE scores, based on oral health and demographic data. The objective is to evaluate the effectiveness of various ML models and identify the most relevant oral health indicators for predicting MMSE scores of 30 (normal cognition) or ≤26 (cognitive impairment) in individuals aged 60 and above.

This cross-sectional study utilizes oral health and demographic data from two existing cohort studies: the European collaborative study Support Monitoring and Reminder Technology for Mild Dementia (SMART4MD) and the Swedish National Study on Aging and Care (SNAC-B). Participants aged 60 years or older will be included in the analysis. The data will be used to develop and evaluate machine learning models for predicting cognitive status.

Objectives:

1. Primary Objective: To assess the potential of oral health parameters for binary classification of MMSE scores (30 vs. ≤26).
2. Secondary Objective: To identify the most influential oral health parameters contributing to cognitive impairment predictions.
3. Tertiary Objective: To compare the performance of Random Forest (RF), Support Vector Machine (SVM), and CatBoost (CB) classifiers in predicting MMSE scores using oral health data.

Inclusion Criteria:

• Individuals aged 60 years or older.
• Participants with recorded oral health parameters and MMSE scores of either 30 or ≤26.

Exclusion Criteria:

• Individuals with MMSE scores of 27, 28, or 29, as these scores represent a transition phase between normal cognition and cognitive impairment, which could introduce variability.
• Individuals younger than 60 years.