Analysis of EEG coherence according to the onset of Alzheimer's disease

Alzheimer's disease (AD) is one of the most common neurodegenerative diseases, the effects of which can be slowed down by an early and correct diagnosis. Greater diagnostic accuracy depends on a better understanding of the disease. In this regard, electroencephalography is a promising low-cost tool to explore brain activity and support the diagnosis of AD. When analyzing EEG signals, age is a factor that is not usually taken into account despite being one of the main risk factors. In this work, an analysis of brain connectivity is proposed in the different frequency bands by analyzing the study groups (healthy control and Alzheimer's disease) in two separate processes according to the age of the patients (U65 and A65). The results obtained suggest that changes in brain connectivity due to AD are different depending on whether AD is present before or after the age of 65 years. Furthermore, it is observed that theta, high Alpha and high Beta are more affected for the U65 group, mainly in the frontal and parietal areas; and low Alpha, low Beta and, especially gamma are more affected for the A65 group in the frontotemporal and frontoparietal areas. Therefore, the importance of dividing the analysis by age lies in its ability to identify changes in brain connectivity based on the age at which AD occurs. Finally, using the information obtained in the analysis, a classification framework is performed to validate the EEG coherence analysis with an overall performance of 87.8 % accuracy, 88.8 % sensitivity, and 87.2 % specificity.