6. Statistical Test 

If you click on  , the following options appear :

The statistical tool box is divided into three steps :

1. Create Groups : Compute mean and var images of a group of individual

2. Ttest Computation : Compute t-test between two groups

3. Ttest Visualisation : To visiualize t-test between two groups

1. Create Groups

The fisrt step to compare groups of brain is to compute mean and var images of a group. To do so, use the following option :

First create a new directory for the mean and variance images as for exemple : Ttest/AC-OLP and give a name to the group, for example AC

Then load images using the button  then press the compute button :

Four images will be created : AC-MeanImage.mhd + AC-MeanImage.raw and AC-VarImage.mhd + AC-VarImage.raw

A file which contains the number of individual in the groups is also created : AC-NumberOfIndi.txt

2. Ttest

Then once you have computed the mean and var images of two groups you can compare them using the option ttest :

First load the Mean image of group 1, for example AC-MeanImage.mhd

Then load the Mean image of group 2, for example D1-MeanImage.mhd

press the compute button  at the end of the computation, you will be asked for the path and directory to save images. For example, create a new directory :

Ttest/AC-OLP_vs_D1-OLP and called the image AC-OLP_vs_D1-OLP.mhd

Three image are created :

1. Mean image of both groups : AC-OLP_vs_D1-OLP-MeanIm.mhd

2. Positive Result of ttest which means AC vs D1 : AC-OLP_vs_D1-OLP-Pos.mhd

3. Negative Result of ttest which means D1 vs AC : AC-OLP_vs_D1-OLP-Neg.mhd

A file with the total number of individual is also created to determine the degree of freedom for the t-test visualisation.

3. Ttest visualisation

If you want to visualize results of groups comparison, use the following option :

Click on the load button, first load the mean image of both group. For example : AC-OLP_vs_D1-OLP-MeanIm.mhd then load the positive or negative ttest test for example AC-OLP_vs_D1-OLP-Neg.mhd.

The following window appear :

To change the p value use the p value window. For example instead of 0.05, use 0.02, press the ok button, , then you get :

It the same principle with the cluster size option, instead of 200 use 1000, then press the ok button, , then you get :