# MICCAI小结

MICCAI 2013已经结束快两周了，做个小结继续上路。

1. 提高效率，这是最关键的；（效率）

# OCTOPRESS TEST

Welcome

alert("Welcome here!");


Congratulation!

# endophenotype一词的使用

Wiki上关于endophenotype的定义，“Endophenotype is a genetic epidemiology term which is used to parse behavioral symptoms into more stable phenotypes with a clear genetic connection.”

# 使用tksurfer可视化左右脑颠倒问题

Tksurfer是freesurfer中可视化surface结果的很方便实用的工具，命令举例如下：

tksurfer fsaverage lh inflated –overlay./fwhm10lh/c1.contrast/sig.mgh

# CentOS上非root安装git

CentOS 6之前yum源中没有git，只能自己编译安装。下面是简单的记录：

# Checking b-vector alignment

DTI数据处理中bvecs是很关键的一个信息，关于bvecs的确定，说简单也简单，但是执意去搞个清晰，绝对是一个坑，一个很深的坑，因为不同的核磁生产商以及不同的DTI处理软件都有自己的一套标准。

# 七年(Seven years forecast)

In seven years there will be no essential difference between comments on articles and peer-reviews,

In seven years there will be semantic means of definition of plagiarism and, as a consequence, a significant percentage of today’s articles will qualify as recycled crap,

In seven years there will be popularity contests and evaluations based on the popularity of the authors as measured by their impact on the web,

In seven years the best universities will gamify the teaching process,

In seven years all successful changes of the process of dissemination of knowledge will turn out to be among those born from private initiatives,

In seven years large research collaborations of mathematicians will be regarded as normal,

In seven years most of the articles which are now under the lock of the copyright belonging to the publisher will be seen as vanity publication and their most important use will be as data for programs of massively extraction of semantic content.

# 虽然苦了点，至少还活着

(下面是一次主题讨论上的讲稿，不喜可以绕行了微笑)

# Seeing is Believing, Binned Scatter Plot

“Seeing is believing” is an idiom meaning “only physical or concrete evidence is convincing”. According to the Wikipedia, it leads to a sophistry that “seen evidence” can be easily and correctly interpreted, when in fact, interpretation may be difficult. It is in the same way when drawing scatter plots with your correlation data.

# Correlation, p-value, CI, and Sampe Size

Lately in my research, I have been focusing on correlations between behavioral data from a large dataset. As is known, correlation is an expression of how well the linear relationship between two sets of data, that is to say, how they are related under the simple linear model. To give an example, for researchers, number of papers and their salary are well correlated. Thus, you can use a researcher’ number of papers to predict his/her salary. It is important to note that ‘correlation does NOT imply causation‘.

# Movement Matters, neural and psychological correlation of head motion during MRI scanning

In MRI experiments, we usually ask our participants to keep still in the scanner because head movement may affect the data quality. Meanwhile, we always do the motion correction before fMRI/rfMRI data preprocessing, aiming to reduce the potential influence of this unwanted noise. However, recent studies showed that there were motion effects even after the motion correction with the standard algorithms. For example, the artifact of participant motion affects the functional connectivity analysis of rfMRI data. In addition, head motion can also result in a bias for the calculation of both FA and MD in DTI data analysis.