An EEG-based Study on Alzheimer’s Disease for Early Stage Detection

Abstract

Alzheimer's disease (AD) is most common type of dementia, characterized by cognitive impairment and loss of memory. Due to its progressive and neurodegenerative properties, it is believed that early detection and intervention may be greatly helpful in therapeutic success. By performing CSF testing, or PET imaging, the existence of current biomarker for detecting early-stage AD can be ascertained. However, these tests are expensive and/or highly invasive. Even presymptomatic AD patients do not have any symptom associated with cognitive function. Electroencephalography (EEG) has many advantages in that it is completely non-invasive and cost-effective. In this thesis, we investigated the difference between task-related EEG patterns of early-stage AD group (presymptomatic and prodromal AD) and that of normal elderlies. About 30 people in each clinical disease state had participated in the visual oddball and 1-back working memory task. Recording EEG, subjects were instructed to press button for a target stimulus and not to respond for a non-target stimulus for the assessment of behavior performance. After the experiments, we analyzed the behavioral results and EEG data. As a result, there was no significant difference in behavior performance such as accuracy and mean response time between normal and early-stage AD group. EEG analysis for the visual oddball task showed that the alpha (8-13 Hz) ERD at Pz channel for early-stage AD group was decreased than that for normal control group. In addition, alpha ERD appears later in both early-stage AD group. It shows that there are changes in brain activities that were induced AD progression. Therefore, we demonstrate that decline in cognition and attention strength.