June 2012

Indipendent Component Analysis and non stationarity

Hi all, I'm just registered at this forum.
I'm studying Indipendent Component Analysis. I noted that it works only with time stationary processes, and fail in presence of non time stationary ones (exponential ramp, linear ramp, etc.). Are there other non-parametric techniques of blind source separation that work in in these cases? I found only parametric methods until now.

Forums: 

multi-label images 能在一个MASK里定义的ROI有上限吧?

‘我今天用 AAL .nii  作为 REST 中“ multi-label images in one mask'" 的 mask 文件,以提取AAL脑区的时间序列
但是出错了,似乎是由于 定义的ROI 数超过了 软件的限制,具体出错信息附后。麻烦指教,多谢!

REST 1.7, Matlab 2011b, WINDOWS XP 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

AAL_116或90 功能连接

经常看到基于AAL-116或90模板分区做全脑功能连接的文献,相关操作问题想请教下老师们。
 
1、 首先要算出这些区域两两之间的相关值,在DPARSF中做是否合适?
2、 DPARSF得出那么多z值后,进行单样本及双样本t检验,可能可以在SPSS下做,但是否在Matlab下做是否更简单呢,相关程序是否可以教一下呢,或者还有其他办法吗?