Research On Mental Workload Assessment Method For Mild Cognitive Impairment

As aging deepens,early detection of Mild Cognitive Impairment(MCI)is increasingly important to prevent Alzheimer Dementia(AD)and improve the quality of life of elders.Mental workload,derived from human factors,consists of cerebral and perceptual activities performed while performing multiple tasks,such as the cognitive functions of the brain.Therefore,it is highly important to design and implement a method that enables an accurate assessment of older adults’ mental workload and can provide a critical window for clinical therapeutic interventions in cognitively impaired populations.However,a large number of studies have focused on the analysis of abnormal brain cognitive functions of MCI by electroencephalography(EEG)signals,neglecting the quantitative assessment of MCI’s mental workload.Furthermore,although EEG-based machine learning and deep learning techniques are widely used for mental workload assessment of AD/MCI,most of them use balanced datasets.Few studies have been conducted to address the class imbalance problem,usually solved by converting class imbalance to balance,which may compromise overall representation learning.In summary,existing mental workload assessment methods can not cope well with the task of measuring mental workload of cognitive impairment patients.To address the above challenges,we have carried out a mental workload assessment method study for MCI in this paper,the main work and contributions are as follows.(1)This thesis proposes a linear discriminant cumulative estimate of mental workload,named EEG-Based Mental Workload Assessment Index of Mild Cognitive Impairment(EMCI).It estimates subjects’ EEG power spectra in α and β rhythms and can be used for the screening and diagnosing of MCI.The results show that the EMCI indexes are sensitive to changes of subjects’ mental workload,and that MCI’s indexes are significantly lower than HC(healthy control),which may be caused precisely by cognitive dysfunction.In addition,a matched prototype system is designed in this paper to validate the validity of EMCI.(2)This thesis proposes a class imbalance oriented AD/MCI mental workload detection method,named CIAM,which is a mutual learning framework using two Ghost attention-improved capsule networks to learn from each other,in order to enhance the student network’s detecting performance and facilitate balanced detection of mental workload.The results show that the method has a relatively well detecting performance and is generalized to both class imbalance and balance scenarios.(3)This thesis designs a mental workload assessment system for cognitive impairment patients,which streams the collected EEG signals to a remote server,uses EMCI indexes to calculate the mental workload changes of the subjects,and uses the CIAM mental workload detection method as an aid to detect and validate the EMCI indexes.Thus,it can be seen that the system provides a proven solution for mental workload detection.