Investigation of EEG Signal for Diagnosis of Mild Cognitive Impairment and Alzheimer's Disease


Oltu B., Aksahin M. F. , Kibaroglu S.

Medical Technologies Congress (TIPTEKNO), İzmir, Turkey, 3 - 05 October 2019, pp.495-498 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/tiptekno.2019.8895256
  • City: İzmir
  • Country: Turkey
  • Page Numbers: pp.495-498
  • Gazi University Affiliated: No

Abstract

Alzheimer's disease (AD), is the most common form of dementia that occurs late in life and is characterized by intellectual regression and neuropsychiatric behavioral disorders. Mild cognitive impairment (MCI) is a transition between normal aging and dementia, and approximately 10-15% of MCI patients progress to Alzheimer type dementia each year. Neurological examination, blood tests, cognitive tests, imaging methods, and analysis of cerebrospinal fluid are performed for diagnosis of these diseases. Such techniques are not practical for large populations at risk, moreover test results are not sufficient to diagnose with high accuracy. In the proposed study, a decision support system that can make an automatic diagnosis is developed for objective and high accuracy diagnosis of AD and MCI. In the developed algorithm, electroencephalography (EEG) signals were filtered and continuous wavelet transform (CWT) was applied to the filtered signals. Mean and standard deviations of the SDD coefficients and scalogram values were extracted as attributes, and with the subtracted attributes disease groups were classified using "Subspace K Nearest Neighbor (Subspace KNN)". The subspace KNN classifier was trained with an accuracy of 88.9% and the trained system tested. As a result of this test, it was concluded that the disease groups can be seperated with 94.44% accuracy rate.