Complexity Analysis of Alzheimer Disease EEG Data through Multiscale Permutation Entropy

The number of patients suffering from Alzheimer’s disease (AD) is rapidly increasing. A variety of sophisticated techniques have been proposed to early detect the precursors of AD to help predict the conversion from Mild Cognitive Impairment (MCI) to AD. The complexity of EEG signals is believed to face an average reduction during the course of the disease. This paper reports preliminary studies on multi-scale entropy analysis of real EEG recordings. In particular, in this research it is introduced the concept of Multi-Scale Permutation Entropy, by comparing its abilities to the well known Multi-Scale Sample Entropy. Limited to the analyzed tracks, the results show that the severity of the disease reflects in the dynamic complexity. Both indexes could develop for copying with multi-channel EEG signals.