{"title":"MFCC-CNN: A patient-independent seizure prediction model.","authors":"Fan Zhang, Boyan Zhang, Siyuan Guo, Xinhong Zhang","doi":"10.1007/s10072-024-07718-y","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Automatic prediction of seizures is a major goal in the field of epilepsy. However, the high variability of Electroencephalogram (EEG) signals in different patients limits the use of prediction models in clinical applications.</p><p><strong>Methods: </strong>This paper proposes a patient-independent seizure prediction model, named MFCC-CNN, to improve the generalization ability. MFCC-CNN model introduces Mel-Frequency Cepstrum Coefficients (MFCC) features and Linear Predictive Cepstral Coefficients (LPCC) features concentrated in the low frequency region, which contains more detailed information. Convolutional neural network (CNN) is used to construct a seizure prediction model.</p><p><strong>Results: </strong>Experimental results showed that the proposed model obtained accuracy of 96 <math><mo>%</mo></math> , sensitivity of 92 <math><mo>%</mo></math> , specificity of 84 <math><mo>%</mo></math> and F1-score of 85 <math><mo>%</mo></math> for 24 cases in CNHB-MIT dataset. The overall performance of MFCC-CNN model is better than the other models.</p><p><strong>Conclusion: </strong>MFCC-CNN model does not need to be specifically customized for different patients. As a patient-independent seizure prediction model, it has good generalization ability.</p>","PeriodicalId":19191,"journal":{"name":"Neurological Sciences","volume":" ","pages":"5897-5908"},"PeriodicalIF":2.7000,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Neurological Sciences","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1007/s10072-024-07718-y","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/8/9 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"CLINICAL NEUROLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Background: Automatic prediction of seizures is a major goal in the field of epilepsy. However, the high variability of Electroencephalogram (EEG) signals in different patients limits the use of prediction models in clinical applications.
Methods: This paper proposes a patient-independent seizure prediction model, named MFCC-CNN, to improve the generalization ability. MFCC-CNN model introduces Mel-Frequency Cepstrum Coefficients (MFCC) features and Linear Predictive Cepstral Coefficients (LPCC) features concentrated in the low frequency region, which contains more detailed information. Convolutional neural network (CNN) is used to construct a seizure prediction model.
Results: Experimental results showed that the proposed model obtained accuracy of 96 , sensitivity of 92 , specificity of 84 and F1-score of 85 for 24 cases in CNHB-MIT dataset. The overall performance of MFCC-CNN model is better than the other models.
Conclusion: MFCC-CNN model does not need to be specifically customized for different patients. As a patient-independent seizure prediction model, it has good generalization ability.
期刊介绍:
Neurological Sciences is intended to provide a medium for the communication of results and ideas in the field of neuroscience. The journal welcomes contributions in both the basic and clinical aspects of the neurosciences. The official language of the journal is English. Reports are published in the form of original articles, short communications, editorials, reviews and letters to the editor. Original articles present the results of experimental or clinical studies in the neurosciences, while short communications are succinct reports permitting the rapid publication of novel results. Original contributions may be submitted for the special sections History of Neurology, Health Care and Neurological Digressions - a forum for cultural topics related to the neurosciences. The journal also publishes correspondence book reviews, meeting reports and announcements.