基于极限学习机的“模拟阅读”脑-机接口异步化研究  被引量:1

Research of Asynchronous Imitating-Reading BCI Based on Extreme Learning Machine

在线阅读下载全文

作  者:李洋 谢国栋 官金安 LI Yang;XIE Guodong;GUAN Jinan(Key Laboratory of Cognitive Science,State Ethnic Affairs Commission,Wuhan 430074;Hubei Key Laboratory of Medical Information Analysis and Tumor Diagnosis & Treatment,Wuhan 430074)

机构地区:[1]认知科学国家民委重点实验室,武汉430074 [2]医学信息分析及肿瘤诊疗湖北省重点实验室,武汉430074

出  处:《计算机与数字工程》2018年第3期479-484,共6页Computer & Digital Engineering

基  金:国家自然科学基金项目(编号:91120017);2017中央专项业务费(编号:CZP17033)资助

摘  要:"模拟阅读"脑-机接口工作在同步状态,而异步脑-机接口能够让使用者自主控制系统,更具实用价值。为了实现异步化,需要对工作状态与空闲状态有效识别。论文利用基于典型的单隐含层前馈神经网络的极限学习机分类方法,讨论了在"模拟阅读"模式下工作状态的特征选择及模式分类。将靶刺激或非靶刺激作为工作状态特征,与闭眼、自由思想、文本阅读等三种空闲状态及其混合形式进行分类。结果表明,以靶刺激时间段作为工作状态特征能够有效地与空闲状态进行区分,而以非靶刺激时间段作为特征则不能有效区分;三种空闲状态的混合形式作为空闲状态有着稳定的识别正确率,表现出良好的普适性。The imitating-reading brain computer interface works on synchronous mode,while the asynchronous brain comput?er interface enables the user to control the system autonomously and has more practical value.In order to achieve asynchrony,working state and the idle state need to be identified effectively.In this paper,using the extreme learning machine classification method based on the typical single-hidden layer feed-forward neural network to discuss the characteristics selection and mode classification of working state in imitating-reading mode.The target stimulus or non-target stimulus is used as working state feature,classified with three idle states,such as eyes closed,freedom thought,text reading and its hybrid form.The results show that the target stimulus time period as the working state feature can be distinguished effectively from the idle state,and the non-target stimulus time period as the working state feature can not be distinguished effectively.The hybrid form of three idle states has stable recognition accuracy and good universality.

关 键 词:模拟阅读 异步脑-机接口 空闲状态 极限学习机 

分 类 号:TP391.4[自动化与计算机技术—计算机应用技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象