MCI identification based on EEG signals during a cognitive test
Résumé
Electroencephalography (EEG) is a valuable method for diagnosing neurological conditions such as mild cognitive impairment (MCI), which is a risk factor for dementia. Based on wearable EEG technologies, recent works have been proposed for the early detection of neurocognitive disorders. Cognitive tests performed during EEG recording have lately emerged as one of the most promising techniques for assisting in the identification of various mental aberrations. In this paper, we suggest a novel method based on machine learning (ML) technology during the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) test for MCI and healthy patients. Yet, the CERAD test has now grown into a popular screening tool for various kinds of dementia, including MCI. Our data were collected from 17 elderly german patients which are previously diagnosed by neurologists as MCI and healthy. To improve the performance of the used algorithms and to overcome overfitting, we applied the principal component analysis (PCA) method as a feature reduction and selection step before applying classification. The proposed approach shows promising results in terms of categorization findings for MCI identification during the CERAD cognitive test. Based on the random forest (RF) algorithm, the best accuracy test is equal to 92.16%.