Study On The Algorithm Of Intraoperative Eletrocorticography Functional Location Of The Human Brain Based On Sample Entropy

Epilepsy and brain tumor are the most frequent brain lesions in clinical.The sum number of patients suffering from epilepsy reaches ten million in China,two to three million of them must undergoing surgical treatment which is called medically intractable epilepsy.The key to success of neurosurgical operation is thorough excision the focal lesions while keeping important brain region from damaging,which is a problem of current clinical research in neurosurgical operation.Currently,for clinical application,electric cortical stimulation(ECS)is the most widely used methods for mapping functional areas of the cerebral cortex intraoperative,which can achieve intraoperative functional location of the human brain.But it is time-consuming and labor-intensive,if not used properly the method could lead to cerebral cortex injury and even trigger epileptic seizures.In order to overcome the drawbacks of the existing method,we propose a new method of intraoperative eletrocorticography(ECoG)functional location of the human brain based on sample entropy.First,acquisition and preprocessing of ECoG.Check each lead signal and abandon the lead which is with interference,remove baseline drift and divide the signal based on the mark file recorded by electroencephalograph operator;decompose the signals with Multiscale Wavelet Transform,and reconstructing it;set up the sample sets of the motor areas and the counting area and the naming area respectively.Second,study the algorithm of intraoperative functional location of the human brain based on sample entropy.Design of feature extraction algorithm based on sample entropy for pre-processed ECoG;Train the classification model of the feature samples of ECoG using support vector machine and apply 5-fold Cross Validation to refine the model.Build an optimal model of the motor areas and the counting area and the naming area for everyone.Last,he location algorithm of each functional area is tested and evaluated respectively.Using the optimal model to classify the characteristic samples of the test set ECoG,according to the recognition rate,the evaluation algorithm.The results show that for the object of 5 patients and 51 leads in the experiment,the predictions outcome of intraoperative ECoG functional location of the human brain based on sample entropy of each functional area leads: the correct recognition rate is 92.2% and the false detection rate is 7.8%,and the data processing time is only no more than 10 minutes.It shows that this method is of feasibility and promising application in intraoperative brain functional location precisely,quickly and safely.