脑机接口中CSP最多被用于运动想象,效果很好,今天来了解下这个算法 数据准备: 右手动EEG数据R{c3,c4} 假设每个channel500个数据点 左手动EEG数据L{c3,c4} 假设每个channel500个数据点 取每个分类下同一时间点两个channle的值,分别产生两个情况下的序列点 画出来如下: x1 x3代表c3 c4 可以看到左右手在时序数据上是很难分开的 CSP做了一件事就是线性转换,从两边压缩变换后: 就变成了这个样子,这样就很好区分: s1来看: 蓝色方差小, 红色方差大 s2来看: 蓝色方差大,红色方差小
立个FLAG: 进军MNE
由于tensorflow使用python, 为了更好的兼容机器学习方面需求,今天立个FLAG:开始进军MNE。
Automatic Anatomy Labeling Categories
copy from http://www.pmod.com/files/download/v36/doc/pneuro/6750.htm but central is wrong, I had corrected them. AL Single-Subject Atlas The AAL-VOIs atlas is the automatic anatomical labeling result [5] of the spatially normalized, single subject, high resolution T1 MRI data set provided by the Montreal Neurological Institute (MNI)[6].
Determine the anatomical label of a source in Fieldtrip
为了加深理解,我还是用中文描述 1. 第一步要对标准MRI(Colin27)进行分层, EEG一般使用BEM{ Brain,Skull,Scalp}, 此为 “Headmodel” 2. 第二步需要将Channel Align到scalp表面 此处可以需要手动调节
Logging and Monitoring Basics with Tensorflow
Tensorflow has provided a method to record all the parameter during the training such as recall accuacry percision and so on. how to write down these data is very improtant , beasuse you need this data to decsrible whether you