The analysis of EEG signals of AD and MCI

Proposal details

Title: The analysis of EEG signals for diagnosis of Alzheimer’s disease and mild cognitive impairment based on synchronization and classification methods
Research Area(s): Brain Modeling
Development and Aging
Brain Imaging
Thinking and Cognition
Background: The synchronization of EEG signals is a signature of brain function and a bio-marker for the early diagnosis of brain diseases such as mild cognitive impairment and Alzheimer’s diseases (Aarabi, Wallois, & Grebe, 2008; Darvas, Ojemann, & Sorensen, 2009; Maria G Knyazeva, et al., 2012; M.G. Knyazeva, et al., 2010; Rudrauf, et al., 2006; Stam, Jones, Nolte, Breakspear, & Scheltens, 2007). And the classification of EEG signals also is a important method used in diagnosing mild cognitive impairment and Alzheimer’s diseases (Joseph Mcbride,et al., 2013; Tiago H Falk, et al., 2012; Trambaiolli LR,et al., 2011; V. Podgorelec,2012). We recently completed a study in diagnosing mild cognitive impairment with small sample based on several synchronization methods. And we classified the mild cognitive impairment sample and control sample based on machine learning methods, and has achieved significant results.
Aims: In this project, we hope to apply our synchronization methods to diagnosis mild cognitive impairment and Alzheimer’s diseases with large sample from Brain Resource International Database, and study some more effective synchronization methods and obtain the EEG signals of mild cognitive impairment and Alzheimer’s diseases from the database to test the performance of these methods. In order to significantly improve the effectiveness of diagnosis to mild cognitive impairment and Alzheimer’s diseases, while studying the synchronization methods we also want to classify the EEG signals of mild cognitive impairment, Alzheimer’s diseases and normal control from the database based on some effective classification methods optimized by us.
Method: We will select all resting EEG and ERP signals of mild cognitive impairment (MCI), Alzheimer’s diseases (AD) and normal control (NC) and relevant clinical data (various neuropsychological scales and other clinical indicators) from the Brain Resource International Database. From the point of synchronization, we will preprocess the resting EEG and ERP signals of MCI, AD and NC from the database, test our existing methods and modified methods by using the EEG signal, and analyze the difference and correlation. From the point of classification, we will also preprocess the EEG signal, select the length and number of sample, extract the features, select some significant features, train the training sample with several classification models, adjust the parameters of relevant models, test the performance of the models by using testing sample, and compare the classification accuracy, sensitivity and specificity of the models.