| 摘 要: 癫痫属于神经系统疾病,反复发作和持久倾向将导致机体损伤,因此提前发现癫痫发作有助提升患者的生活质量。为了全面且深入地探究人工智能在预测癫痫发作方面的研究进展及趋势,首先介绍了目前常用的预测癫痫的脑电公开数据集、评价指标和预处理技术,其次将基于人工智能的癫痫发作预测研究划分为基于机器学习和基于深度学习两类,并分别进行分析。分析结果显示,基于深度学习的癫痫发作预测,准确率能达到95%以上。基于以上研究结果得出人工智能应用于癫痫发作预测具有良好的发展前景。 |
| 关键词: 癫痫发作预测;深度学习;机器学习;脑电图 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 国家重点研发计划(2020YFC2008700);国家自然科学基金(61971275,81830052,82072228)资助;上海市地方高校能力建设科学技术委员会(23010502700)资助 |
|
| A Review of Epilepsy Prediction Based on Artificial Intelligence |
|
WANG Wenjie1,2, YAO Xufeng2
|
(1.School o f Health Science and Engineering, University of Shanghai f or Science and Technology, Shanghai 200093, China;
2.College o f Medica Imaging, Shanghai University o f Medicine & Health Sciences, Shanghai 201318, China)
1176728277@qq.com; yao6636329@hotmail.com
|
| Abstract: Epilepsy is a neurological disorder, and recurrent seizures and persistent tendencies will lead to physical damage. Therefore, early detection of epilepsy seizures can help improve patients' quality of life. In order to comprehensively and deeply explore the research progress and trends of Artificial Intelligence (AI) in predicting epilepsy seizures, this paper first introduces commonly-used and publicly available EEG ( Electroencephalogram) datasets for predicting epilepsy, evaluation indicators, and preprocessing techniques. Then, it divides the research on AI-based epilepsy seizure prediction into two categories: machine learning and deep learning, and analyzes them separately. The analysis results show that in deep learning-based epilepsy seizure prediction, the accuracy can reach over 95% . Based on the above research results, it can be concluded that the application of AI in epilepsy seizure prediction has promising prospects. |
| Keywords: epilepsy seizure prediction; deep learning; machine learning; Electroencephalogram |
