Classification And Recognition Of Severe Disorders Of Consciousness Based On Dual Mode Induction

Severe disturbance of consciousness refers to a state in which the brain function of the patient is inhibited after suffering a severe brain injury.The clinical manifestations are continuous coma,lack of self-awareness and the ability to perceive the surrounding environment.The severe consciousness disorder state is divided into the minimum minimally conscious state and the vegetative state.How to accurately identify the patient’s state of consciousness and improve the patient’s treatment rate is a hot spot in clinical medical research.At present,the commonly used method is the behavioral awareness scale,which is simple to operate but easily affected by subjective factors and has a high misdiagnosis rate.Medical imaging technologies,such as functional Magnetic Resonance Imaging,have high spatial resolution,but need to perform out-of-bed detection on patients,which has a high risk.Electroencephalogram detection technology has the advantages of high time resolution,bedside detection,lower cost,and repeated use.Therefore,based on the analysis method of EEG signal,this paper studies the recognition of the patient’s state of consciousness.In this paper,a dual-mode evoked experiment is designed to compare and analyze the changes in the characteristic values of the brain electrical signals of patients under different stimulation modes(auditory stimulation and somatosensory pain stimulation),so as to realize the judgment of the patient’s state of consciousness.First,EEG signals collected in the experiment were preprocessed,and then select the time-frequency domain features and nonlinear features under auditory stimulation,and extract the frequency domain features and nonlinear features under somatosensory pain stimulation,and finally through linear combination.Construct a set of multi-domain feature samples.Experimental results show that compared with traditional single-domain feature extraction methods,multi-domain feature fusion can improve the robustness and classification performance of the classification model.In the field of recognition of severe consciousness disorders,although support vector machine has a good performance in solving two-class classification problems,the classification results are highly dependent on the selected kernel function type and related parameter settings.Therefore,this paper proposes a support vector machine model based on multi-core learning.According to the combination of multi-domain features and solving convex programming problems,8 kernel functions are selected,including 3 polynomial kernel functions and 5 Gaussian radial basis kernel functions.The experimental results show that the average recognition accuracy of auditory stimulation experiments is 7.3% higher than that of somatosensory pain stimulation experiments under different stimulation modes,and the recognition accuracy obtained by the combination of multi-domain feature fusion and multi-core learning SVM algorithm is higher than that of single domain The single-core model and the single-domain multi-core model are higher,with an average recognition accuracy rate of 88.3%.Finally,it proves that the classification method proposed in this paper can effectively identify the type of consciousness disorder of patients and has certain clinical application value.