使用神经网络从非特定人群中对创伤性脑损伤和中风进行EEG分类
Michael Caiola1, Avaneesh Babu1, Meijun Ye1
1Division of Biomedical Physics, Office of Science and Engineering Laboratories, Center for Devices and Radiological Health, FDA, Silver Spring, Maryland, United States of America.
PLOS digital health
|July 6, 2023
概括
机器学习模型有效地区分正常,创伤性脑损伤 (TBI) 和使用电脑电图 (EEG) 的中风. 基于特征的模型达到0.85AUC,而没有特征的模型达到0.84AUC,为神经疾病检测提供更快的部署.
科学领域:
- 神经学 神经学
- 生物医学工程 生物医学工程
- 机器学习 机器学习
背景情况:
- 创伤性脑损伤 (TBI) 和中风是具有诊断挑战的关键神经疾病.
- 目前的检测方法通常依赖于先进的成像或医院接入,限制了可访问性.
- 以往在脑电图 (EEG) 上的机器学习在分类这些疾病方面取得了中等成功 (准确率为0.71).
研究的目的:
- 评估无特征和深度学习模型在正常,TBI和中风EEG分类中的性能.
- 将基于特征的机器学习与无特征的深度学习方法进行比较,以提高诊断准确度.
- 探索梯度加权类激活映射 (Grad-CAM) 对EEG分类解释性的有用性.
主要方法:
- 利用全面的数据提取来扩展用于EEG分析的训练数据集.
- 基于特征的模型 (线性差别分析,ReliefF) 与没有特征的深度学习模型进行了比较.
- 用人接收机运行特征 (ROC) 曲线分析以评估模型性能,测量曲线下的面积 (AUC).
主要成果:
- 基于特征的模型实现了0.85.8的AUC.
- 没有特征的深度学习模型达到0.84.8的可比AUC.
- 通过突出分类关键EEG段,Grad-CAM提供了针对患者的具体见解.
结论:
- 机器学习和EEG的深度学习是TBI和中风检测和分类的有希望的工具.
- 无特征模型为基于特征的方法提供了一个可行的,更快,更具成本效益的替代方案,而不会牺牲显著的性能.
- 这些人工智能驱动的方法可能会提高诊断神经紧急情况的可访问性和效率.
更多相关视频
09:16Use of a Wireless Video-EEG System to Monitor Epileptiform Discharges Following Lateral Fluid-Percussion Induced Traumatic Brain Injury
Published on: June 21, 2019
25.7K
09:42Author Spotlight: Using Motor Imagery Brain-Computer Interface to Improve Motor and Cognitive Function in Stroke Patients
Published on: September 1, 2023
1.3K
相关概念视频
01:13Seizures: Classification
436
Epilepsy is primarily characterized by unpredictable seizures, either provoked by an identifiable factor, such as injury or illness, or unprovoked, occurring spontaneously without apparent cause.
Seizures are typically classified into two main categories: focal and generalized seizures.
Focal Seizures
Focal seizures originate from specific regions of the brain. These seizures are further sub-classified into two types:
Seizures are typically classified into two main categories: focal and generalized seizures.
Focal Seizures
Focal seizures originate from specific regions of the brain. These seizures are further sub-classified into two types:
436
01:23Brain Waves
1.6K
Brain waves are electrical signals generated by the neurons in the brain, which are regularly monitored to measure mental activities. Brain waves and their frequency ranges can be measured using an electroencephalogram or EEG. There are four main types of brain waves, each with distinct characteristics:
1.6K