Eeg Signals Analysis Of Diabetic Patients With Amnestic Mild Cognitive Impairment

The incidence of diabetes and cognitive impairment complications increased year byyear, seriously affecting the patients’ life, and there is an enormous as well as financialburden on individuals, their families and society as a whole. Diabetes is an important riskfactor for cognitive decline and dementia in elder, however, the complex relationshipbetween diabetes and aging mechanisms needs to be further revealed. Since the1980s,people begin to pay close attention to the relationship between type2diabetes anddementia, but has not been solved. Electroencephalogram(EEG) is the window forunderstanding the brain, so EEG signals was used to the research of revealing damagemechanisms of the functions in diabetics, tracking the patients with diabetes afterrehabilitation treatment, so as to deepen the cognition and understanding of diabetesinduced impairment mechanism and improvement, and to guid the rehabilitation therapyof diabetic cognitive impairment is of great significance.This paper was devoted to studythe changes of EEG signals and event-related EEG signals in amnestic mild cognitiveimpairment (aMCI) with diabetes, and the mainly research work was as following:Firstly, based on the linear analysis of EEG, the brain activity and the synchronizationinter-and intra-regions were studied with power spectrum and coherence methods, andthe correlations between the EEG linear indicators and neuropsychological scores wereanalized. We explored the role of resting-state EEG indicators in the assessment ofcognitive function, then proposed that the cognitive impairment in diabetic patientsexhibited reduced brain activity and weakened synchronization, and the EEG analysiswere more objective.Secondly, based on the non-linear analysis of EEG, the differences of permutationentropy and weighted permutation entropy values between the amnestic cognitiveimpairment diabetic patients and the normal control group, the sensitivity to the parameterchanges of the two indicators were studied. The correlations between entropy andneuropsychological scores were introduced. We proposed that the weighted permutationentropy is more robust to noise, it has practical value in distincting the aMCI from normal diabetic patients.Thirdly, based on the event-related EEG signals, the feature of ERPs and discretewavelet coherence were extracted. These features, such as time, area, frequency band andexperimental paradigm, were normalized to0.1to0.9, and then were combined to extratcanonical variables applying the minimal-redundancy-maximal-relevance analysis basedon mutual information, thus made classification of two groups. Those results showed thatthe effect of n-back paradigm on gamma band was more significant, which was associatedwith the reduced function of working memory in aMCI diabetes. Noted the study on therole of Kanizsa illusions figure in distinguishing the two group patients remains to furtherconducted.Finally, the EEG signals of health college students was analyzed using the powerspectrum and global synchronization index methods. It was found that the Pilates trainingmade the brain neural network more active, and the synchronization of frontal andtemporal regions reduced. This is a preliminary study of Pilates intervention on brainfunction, providing a preliminary basis for Pilates’ applying to degenerative brain diseaseand cognitive dysfunction rehabilitation interventions.