Improvement Of Analysis Method Of Neuron Spike Signal

The birth and development of neural recording electrodes provide a good recording condition for the neuron spike signals of postgraduates,and the detection and classification of the recorded signals is a problem that needs to be solved emphatically.In order to understand the first condition of how the brain processes information,the study of spike potentials has always been a hot issue.In this paper,a spike detection method is proposed,which combines RMS quadratic double threshold with peak-to-peak value and Teager nonlinear energy operator detection method to detect the spike of collected signals.PCA and K-means analysis are used to classify the detected spike preliminarily,and then the preliminary results are measured by the index of intra-class and extra-class spacing.The optimal results are analyzed as the initial cluster number and initial cluster center of Fuzzy-C analysis,and the final classification results are obtained.After the spike signals are classified,this paper uses the methods of neuron discharge interval analysis,autocorrelation and cross-correlation analysis and signal time-frequency analysis to analyze the discharge characteristics of each kind of neuron signals.The main results are as follows:(1)Compared with threshold detection method,the spike detection method proposed in this paper has improved detection rate and positive detection rate,especially when there are many spikes.(2)The spike classification method proposed in this paper can classify spikes accurately,which is more accurate than the traditional feature extraction detection method.(3)It is found that before and after cerebral ischemia,the firing frequencies of the three types of neurons vary.Autocorrelation analysis shows that the autocorrelation of the three types of neurons is very weak,showing different changes before and after cerebral ischemia.At the same time,the correlation among the three types of neurons is not obvious,and the phenomenon of synchronous discharge response is not found.In a word,the spike detection and classification algorithm proposed in this paper provides a method of data analysis for neurophysiology research,and has certain guiding significance and application prospect for related electrophysiological research experiments.