December 2012

关于alphasim校正新的思考

最近又研究了一下alphasim校正,但是不清楚自己想的对不对,所以来这里和大家探讨下。

其实个人感觉alphasim并不是那么弱,只是很多研究者可能在错误的使用它。最近通过看一些文章发现:alphasim中的平滑核大小并不是我们进行预处理时smooth平滑核的大小。这个大小是需要第三方重新进行估计的。如: 3dFWHM。所以rest中给出的平滑核大小和需要的cluster size就失去了本身的意义。

比如:预处理平滑的时候用的8 8 8,最后会出来如下这个,那么我们alphasim校正时的应该是14.3 13.8 13.4 mm mm mm,我的理解对吗?但是我没有想通为什么这个平滑核和预处理时的平滑核相差会这么大呢?而且下面这三行该如何解读?resel代表了什么?

How to convert from voxel index to MNI coordinates and vice versa.

Please see the figure below (aal.nii opened by MRIcroN).
The MNI coordinates of the current position is -66 -92 -40, and its value is 0. (-66x-92x-40= 0 on the titile).
The voxel index is 25 34 32. (X 25 Y 34 Z 32). 

Read 3D EPI functional images -----out of memory

使用新的rest 和DPARSF
数据·处理全部采用 nifti 4D
在 Read 3D EPI functional images 步骤就齐刷刷的报告 out of memory
我的机器,DELL的入门级工作站,4G内存,应该不是玩fMRI的人里面最差的吧?
不知各位老大有没有低配置硬件的解决方案?时间是小事,别报错就成呀!!
多谢多谢!

DPARSFA 2.2 开始运行就报错 求助

Win 2003 server
Matlab 2011b
SPM 8
REST 1.8

出错信息如下,求助

Error using cfg_util (line 835)
Job execution failed. The full log of this run can be found in MATLAB command window, starting with the lines
(look for the line showing the exact #job as displayed in this error message)
------------------
Running job #2

rest计算degree centrality归一化问题

各位老师:

做完degree centrality会自动出来一个归一化的图(减去均值除以标准差),请问不归一化的话会有什么不好的吗? 如果在subject内部做归一化,个人感觉会破坏掉原来值的太小,从而导致组比较的时候结果不同(归一化与不归一化)。

有点困惑,望解答下

REST V1.8 and DPARSF V2.2 were released!

Dear all,
 
Merry Christmas and Happy New Year!
 
It’s said that December 21, 2012 is the end of the world, thus it marks the beginning of a new era. Here, we celebrate the Christmas and New Year in the new era, and are pleased to release the stable version of REST V1.8 and DPARSF V2.2. :)
 
To facilitate parallel computing and lots of new features, many functions were re-structured, especially for DPARSF advanced version. Thus, we released the pre-release of REST V1.8 and DPARSF V2.2 on September 5, 2012 and invited users to help us to refine them. Many thanks to our users, for testing, using and reporting bugs or problems encountered, we kept updating the pre-release and now believe REST V1.8 (REST_V1.8_121225) and DPARSF V2.2 (DPARSF_V2.2_121225) have reached a more stable stage (probably there still will be bugs, we will fix them soon and make a next stable release upon receiving reports).
 
We encourage the users to re-perform the analyses with the stable version if you used the earlier pre-releases of REST V1.8 and DPARSF V2.2. We expect most of the results will keep the same, but there are several changes in the stable version.
 
Of note, the default parameter in the pre-release is set to “Calculate in Original Space” which calculates the R-fMRI measures on the data before spatial normalization and resampling. This way was suggested to maintain the intrinsic fidelity of spontaneous fluctuations, and thus be beneficial in revealing changes in subtle small brain areas (Wu et al., 2011. Empirical Evaluations of Slice-Timing, Smoothing, and Normalization Effects in Seed-Based, Resting-State Functional Magnetic Resonance Imaging Analyses. Brain Connect 1, 401-410). We set the default parameter in the pre-release to promote the capability of calculating in original space with DPARSF. However, the same study (Wu et al., 2011) found no significance between strategies of calculating in original space or MNI space, and they suggested using the latter way when targeting large and robust functional networks by providing consistent spatial extent. In the stable release of DAPRSF (DPARSF_V2.2_121225), the default parameter is conservatively set back to “Calculate in MNI Space” for a better comparison with most of the previous studies. Another consideration is that if data of multiple sites were involved (e.g., the FCP 1000 data), ReHo and degree centrality need to be performed in MNI space to address the issue with difference in voxel size. Nonetheless, users could choose either way by selecting from “Template Parameters” based upon their objectives.
 
In addition, before VMHC calculation, DPARSF_V2.2_121225 now can create a symmetric mean T1 template from all the participants, and will normalize the T1 image for each subject to the symmetric T1 template (created in Step 2), and apply the transformations to the functional data. In addition, REST-GCA could handle multiple ROIs (other than 2) in ROI-wise GCA in REST V1.8 pre-release, but the outputs were flipped. We fixed this bug in the REST_V1.8_121225, now the outputs were put in according to the description in REST-GCA readme.
 
We hope REST_V1.8_121225 and DPARSF_V2.2_121225 can continually help our users in your research and applications and thus promote the resting-state fMRI studies.
 
On behalf of the REST Team, I wish you a happy, healthy and prosperous 2013!
 
Best,
 
Chao-Gan
 

运行DPARSF出现Undefined function 'list' for input arguments of type 'struct' 问题

在Matlab2012a中运行Dparsf出现如下问题:

Warning: Run spm_jobman('initcfg'); beforehand 
> In spm_jobman at 107
  In DPARSF_run at 364
  In DPARSF>pushbuttonRun_Callback at 976
  In gui_mainfcn at 96
  In DPARSF at 43 

Slice Timing出错

 大家好:
用DPARSF做预处理,slicetime总是过不去。
显示Error running job: Error using fileparts

用SPM自带的reslice,也报错,
显示:Error running job: Error using fileparts

Too many output arguments.
In file "D:\Program Files\MATLAB\R2012a\spm5\spm_slice_timing.m" (v671), function "spm_slice_timing" at line 183.