Evaluation of Electroencephalography Signals in Alzheimer’s Disease Using Coherence Analysis and Persistent Homology

<strong>Objective:</strong> This study aimed to use a new approach, namely persistent homology, to analyse electroen cephalogram (EEG) coherence and identify the alterations in brain connectivity in patients with Alzheimer’s disease (AD). <strong>Materials and Methods: </strong>We applied persistent homology to the distance maps that we created using the EEG coherence values from five different frequency bands in order to determine if there are disruptions specific to these bands in patients diagnosed with AD. <strong>Results: </strong>Our findings revealed that the features extracted using persistent homology were significantly different in two bands (delta and theta) between AD patients and subjects in the healthy control (HC) group. Furthermore, the machine learning algorithms using these topological features achieved accurate classification results. This suggests that persistent homology may be a useful adjunct in the diagnosis of AD. <strong>Conclusion:</strong> We have demonstrated the potential of persistent homology in identifying AD-related changes in brain connectivity, which are the most clearly present in the theta and delta bands. Larger datasets should be used in future research to determine the clinical relevancy of this method.