Comparative analysis of different brain regions using machine learning for prediction of EMCI and LMCI stages of Alzheimer's disease



  • Publication Type: Article / Article
  • Publication Date: 2023
  • Doi Number: 10.1007/s11042-023-16413-7
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, FRANCIS, ABI/INFORM, Applied Science & Technology Source, Compendex, Computer & Applied Sciences, INSPEC, zbMATH
  • Istanbul University Affiliated: No


For early diagnosis of dementia and slowing the progression of Alzheimer's disease (AD), detection of Mild Cognitive Impairment (MCI), which is the first stage of AD, in the early or late stages is crucial. The progression from Early Mild Cognitive Impairment (EMCI) stage to Late Mild Cognitive Impairment (LMCI) stage is not reversible and means that the cognitive condition of the patient gets worse significantly. Therefore, distinguishing the stages of MCI is very important for treatment possibilities. In this paper, it has been aimed to specify which brain regions are affected higher during the progression from EMCI to LMCI. Detection of EMCI stage gives an important opportunity to control the progression and results of the disease. Unfortunately, it is a very challenging classification problem because the changes in the values of biomarkers are generally low during the EMCI and LMCI stages. As a result of this study, we detect and present a combination of features which are the most effective ones for distinguishing the stages of MCI. Atrophy values obtained by magnetic resonance imaging (MRI) are considered as the powerful diagnostic biomarkers for the detection of AD. In this work, atrophy values of 90 EMCI, 38 LMCI and 14 MCI patients have been used. Volume information of 13 different brain regions for each patient were obtained from the ADNI dataset. By using the results of classification algorithms, the mostly affected brain regions on transition process from EMCI to LMCI are determined. Moreover, the classification results indicate the combination of the most effective features. This feature combination can be used as a pattern in the researches about the stages of MCI. Focusing on the brain regions which have more impact on the progression of AD can provide more sensitive analysis of the stages of AD and make possible to control and smooth the effects of it.