EEG multiscale entropy dynamics in mild cognitive impairment and early Alzheimer's disease

Abstract

Mild cognitive impairment (MCI) is a neurological condition that is often the early stage of Alzheimer's disease (AD). This pilot study explores event-related multiscale entropy (MSE) measures as features for effectively discriminating between normal aging, MCI, and AD participants. Thirty two-channel scalp EEG records recorded during a working memory task from 43 age-matched participants (mean age 75.7 years)-17 normal controls (NC), 16 MCI, and 10 early ADare examined. Multiscale entropy curves are computed for responses during the working memory task. Support vector machine models are constructed to perform binary discriminations among the three groups. Leave-one-out cross-validation accuracies of 87.9% (p-value <;1.322E-4) for MCI vs. NC, 88.9% (p-value <;2.886E-5) for AD vs. NC, and 92.3% (p-value <;4.910E-6) for MCI vs. AD are achieved. Results demonstrate links between event-related multiscale entropy dynamics of EEG and short-term memory deficits.