最大误差压缩下的脑电微状态模板差异研究

Difference of EEG microstate template under maximum-error compression

本文针对大体量脑电数据难以高效分析的问题,以脑电微状态分析为例,探讨了FShift保质压缩算法误差界对于微状态模版的影响。实验选用60例临床脑电数据,分别在原始数据(O)、FShift保质压缩后再解压数据(F)以及压缩概要(R)上计算AD/MCI两类人群的微状态模版。使用Pearson相关系数度量模板间的差异,F和O上得到的模板相似度为1,R和O上得到的模板相似度最小值为0.9808,证明了最大误差压缩下的脑电数据可以满足实际计算精度要求。

In order to cope with the efficiently analyzing large volumes of EEG data,EEG microstate analysis is taken as an example to explore the impact of the error bounds of the FShift quality-preserving compression algorithm on microstate templates.60 cases of clinical EEG data are selected for the experiment,and the microstate templates for the two AD/MCI populations are computed on the original data(O),the FShift compressed and decompressed data(F),and the compressed synopsis calculated by FShift(R)respectively.Pearson correlation coefficient is used to measure the difference between templates.The similarity of templates on F and O is 1,and the minimum similarity of templates on R and O is 0.9808,which proves that the EEG data under the maximum error compression can meet the requirements of actual calculation accuracy.