Improving Early Diagnosis of Alzheimer's Disease Using Synchrony Measures.

It is well-known that Alzheimer's disease causes changes on the electroencephalography of the patients. However those changes are difficult to parameterize. In this paper a new ratio between synchrony in. and a band is investigated in order to get an early diagnosis of Mild Alzheimer's patients. The presented ratio is compared using two types of classifiers, Linear Discriminant Analysis and Artificial Neural Networks, with values of synchrony in the standard frequency bands. Presented results improve using the ratio in the linear classifier. Using the non-linear classifier, best results are obtained using synchrony measures in. and a band simultaneously.