Predictive models for Alzheimer's disease diagnosis and MCI identification: The use of cognitive scores and artificial intelligence algorithms

Abstract

The paper presents a comprehensive study on predictive models for Alzheimer's disease (AD) and mild cognitive impairment (MCI) diagnosis, implementing a combination of cognitive scores and artificial intelligence algorithms. The research includes detailed analyses of clinical and demographic variables such as age, education, and various cognitive and functional scores, investigating their distributions and correlations with AD and MCI. The study utilizes several machine-learning classifiers, comparing their performance through metrics like accuracy, precision, and area under the ROC curve (AUC). Key findings include the influence of gender on AD prevalence, the potential protective effect of education, and the significance of functional decline and cognitive performance scores in the models. The results demonstrate the effectiveness of ensemble methods and the robustness of the models across different data subsets, highlighting the potential of artificial intelligence in enhancing diagnostic accuracy for Alzheimer's disease and mild cognitive impairment.

Cette étude explore l’application des algorithmes d’apprentissage automatique pour le diagnostic de la maladie d’Alzheimer (MA) et l’identification de la détérioration cognitive légère (DCL), en utilisant des scores cognitifs parmi d’autres variables cliniques et démographiques. Nous décrivons notre méthodologie, incluant la collecte de données, le prétraitement, la sélection des caractéristiques, et l’utilisation de divers classificateurs d’apprentissage machine. Les résultats mettent en évidence l’efficacité des méthodes d’ensemble dans la prédiction de la MA et de la DCL, discutent des implications de ces résultats pour le diagnostic précoce et l’intervention, et suggèrent des directions pour les recherches futures.