Evaluation Of Mild Cognitive Impairment Based On Eeg Signalanalysis And Intervention System Design

Mild cognitive impairment is a chronic disease threatening the health of the elderly,which creates an intermediate state mainly between the normal aging and Alzheimer's dementia patients,compared with normal aging,there is impairment of memory as well as cognition,but it has not reached the criterion of dementia.Combining the research based on scale assessment and EEG evaluation,an assessment system including mild cognitive impairment evaluation and early intervention is designed,which is able to implement the early assessment of patients and give them appropriate intervention through electronic game training.The scales applied to the assessment are: Mini Mental State Examination(MMSE),Ability of Daily Life(ADL)and Hachinski Inchemic Score(HIS).MMSE determines whether the patient is suffering from memory impairment.Usually the normal aging scores more than 29,while the MCI patients' score ranges from 23 to 27.The ability of ADL is to estimate the ability of self-care in daily life.Hachinski Inchemic Score(HIS)mainly determines the causes of dementia,excluding ischemia dementia.In conclusion,the subjectivity of neural psychology scales test is great,namely the measurement result is greatly impacted by subjective consciousness of the patients.According to EEG evaluation,the nonlinear features reflecting EEG signal ' s nonlinearity and chaos,such as complexity,symbolic transfer entropy,sample entropy,permutation entropy and multifractal exponents,are extracted and fused.The permutation entropy algorithm here is improved,increasing the entropy by about 0.28.It shows that the improved algorithm is more suitable for the extraction of nonlinearity characteristic of EEG signal.The classification of the normal aging and MCI patients based on the feature fusion method above has reached an average accuracy of 94.74%.EEG evaluation is less affected by subjectivity than scale assessment.The early intervention system provides appropriate intervention to MCI patients,delaying the occurence and deterioration of dementia.This system consists of music relaxation,video memory,number contrast and poetry recitation game,which makes itsimple.The EEG signal features before and after the games is classfied through Support Vector Machine(SVM),with the highest accuracy 64.89% and the average accuracy58.10%,proving that appropriate game training is beneficial to MCI patient intervention.Under the Visual Studio 2010 compiler environment,the system in this paper is designed by hybrid programming of Matlab 2009 a and C#.It is divided into three modules:the neurological scale test,the EEG signal analysis and the game training,which achieves dementia assessment and appropriate intervention.