Signal Analysis Of Dementia Based On Neural Network Oscillations

The appearance of Alzheimer’s disease (AD) causes great pain and burden to families and society. The early diagnostic criteria of AD has not been found and the exact etiology and pathogenesis is unclear. Now people according to find the widespread presence of senile plaques and neurofibrillary tangles in cortex regions to get diagnosed. And there are no effective methods at home or abroad for the treatment of AD. Electroencephalogram (EEG) is the important external manifestation to Judge the brain’s structure and function normal or not. Based on the form or function of the brain, cognitive function is an important aspect of brain function, therefore, exploring the EEG characteristics of the elderly and the relationship between cognitive function and EEG is important for the early diagnosis of senile cognitive decline and its physiological basis.In this paper, the global synchronization index (GSI) method and parallel factor (PARAFAC) analysis method analyzed the mild cognitive impaired (MCI) patients and normal control (NC) group respectively. Global synchronization index method is a coincident index method which can integrate the entire brain regions to evaluate the multivariate time series recorded at the same time. Finally, according to all eigenvalues get a GSI which reflects the synchronization’s size. Compared with the previous algorithm,it integrates more informations and has other advantages, such as high sensitivity, small calculation-error and strong time-stability. This method can be applied to analysis the complex, spatial extension and unstable systems. The purpose of parallel factor analysis is decompose the three-dimensional tensor generated by the continuous wavelet transform into the two-dimensional information on the time domain, frequency domain and space domain. Then take advantage of the statistical methods to reveal the EEG characteristics on time domain, frequency domain and space domain. This method for the decomposition of multi-channel EEG on time, space and frequency can extract the meaningful and significant physiological activitiesreliably and uniquely. The results show that after the global synchronization index method analyze two sets of data, the GSI value of mild cognitive impairment patients is significant differences with the GSI value of NC at δ,α and β3, and the values of the GSI at three bands significantly related to the severity of symptoms. Compared with NC,after parallel factor analysis, the advantage rhythm of MCI patients was significantly lower and the size of the corresponding energy also declined.