A systematic review on early prediction of Mild cognitive impairment to alzheimers using machine learning algorithms

Abstract

Background

A person consults a doctor when he or she is suspicious of their cognitive abilities. Finding patients who can be converted into Alzheimer's in the future is a difficult task for doctors. A person's dementia can be converted into several types of dementia conditions. Among all dementia, Alzheimer's is considered to be the most dangerous as its rapid progression can even lead to the death of an individual. Consequently, early detection of Alzheimer's would help in better planning for the treatment of the disease. Thereby, it is possible to reduce the progression of the disease. The application of Machine Learning algorithms is useful in accurately identifying Alzheimer's patients. Advanced Machine Learning algorithms are capable of increasing the performance classification of future AD patients. Hence, this study is made on a number of previous works from 2016 onwards on Alzheimer's detection. The aspects such as the country of the participants, modalities of data used and the features involved, feature extraction methods used, how many follow-up data were used, the period of Mild Cognitive Impairment to Alzheimer's Disease converters predicted, and the various machine learning models used in the previous studies of Alzheimer's detection are reviewed in this study. This review helps a new researcher to know the features and Machine Learning models used in the previous studies for the early detection of Alzheimer's. Thus, this study also helps a researcher to critically evaluate the literature on Alzheimer's disease detection very easily as the paper is organized according to the various steps of the Machine Learning process for Alzheimer's detection in a simplified manner.