EEG-based Sudden Pain Recognition Method And Experimental Study

Pain,as a common symptom of many diseases,tends to cause some suffering to the patient.Researches have shown that pain can be detected by objective methods,making pain judgments especially for aphasia and paralyzed patients who cannot express pain information,and these are able to find out where the pain is located and to take appropriate measures to facilitate early treatment of the diseases,getting rid of pain.Some methods have shortcomings: the mechanism is unknown;the signal is weak;establishing the model is difficult and so on.Therefore,in view of the above defects,this paper focuses on the study of a sudden pain recognition method based on measurement of EEG signals,building a complete set of pain experimental platforms and continues improving,Ultimately this paper can identify the presence of pain and distinguish the location of the resulting pain,bringing great significance in medicine and social aspects.This paper first describes the nature of pain,the current status of research and the significance of detecting pain.Based on the EEG signal,a sudden pain identification scheme is designed based on the relationship between pain and EEG signals,and the overall block diagram of the identification method is given.In order to preserve useful information that is submerged by noise in the EEG signal,median filtering,band-pass filtering,band-rejection filtering and Hilbert-Huang transform preprocessing algorithms are used for preprocessing,which largely removes noise while retaining useful information;In order to explore the differences in signal characteristics relative to relaxation and pain,power spectral density,short-time Fourier transform and multifractal detrended fluctuation analysis are used to extract the characteristic values of relaxation and pain period and the difference is obvious;In order to identify the specific pain information,the extracted feature values are sent to the classifier to study and classify.Finally,whether the subject is painful or not and the location of the pain can be determined.At the same time,according to the factors of individual differences,in the process of signal processing,the original EEG signal s,the preprocessed EEG signals and the extracted feature values have been selected step by step to ensure the validity of the data and then are processed.In order to ensure the validity of the acquisition of EEG signals,this paper sets up a complete pain experimental platform,introduces experimental devices,specified specific pain experimental procedures,and controls other variables that have nothing to do with pain,in order to study the effect in EEG signals made by pain stimulation only.The experiment collects a large amount of data,enriches the database and provides data support for the above signal processing analysis algorithm.After setting up a complete platform and collecting valid data,the analysis find that the characteristic values of pain are significantly different.The classifier can identify whether the subject was painful,and the accuracy rate is 83.3%.There is also a certain effect on the difference of pain position,and the accuracy rate can reach 79.2%.The experimental data and the corresponding results prove the effectiveness of the design of the pain recognition scheme and the feasibility of all signal processing analysis algorithms.