Visualizing Images From NIfTI File

Read an NIfTI file

This is example to read a NIfTI file and display an image , by using this example code you must be install package oro.nifti) first.

Download a NIfTI file from Neurohacking_data repository

library(oro.nifti)

url <- "https://raw.githubusercontent.com/muschellij2/Neurohacking_data/master/BRAINIX/NIfTI/Output_3D_File.nii.gz"
destfile <- "Output_3D_File.nii.gz"
name <- file.path(getwd(), destfile)
download.file(url, destfile,mode="wb") # NIfTI is binaryfile format
nii_T1 <- readNIfTI(destfile)

Show meta data from the file

print (nii_T1)

## NIfTI-1 format
##   Type            : nifti
##   Data Type       : 4 (INT16)
##   Bits per Pixel  : 16
##   Slice Code      : 0 (Unknown)
##   Intent Code     : 0 (None)
##   Qform Code      : 2 (Aligned_Anat)
##   Sform Code      : 2 (Aligned_Anat)
##   Dimension       : 512 x 512 x 22
##   Pixel Dimension : 0.47 x 0.47 x 5
##   Voxel Units     : mm
##   Time Units      : sec

As you see this file contain 22 images 512 x 512 pixels, one pixel using 16 bits

Visial an Slice an NIfTI file

image(nii_T1,z=11,plot.type="single") 

Visualizing All Slices

image(nii_T1)

All Planes: Coronal, Sagittal, Axial

orthographic(nii_T1,xyz=c(200,220,11)) 

Backmapping One Slice

is_btw_300_400<- ((nii_T1>300) &
(nii_T1<400))
nii_T1_mask<-nii_T1
nii_T1_mask[!is_btw_300_400]=NA
overlay(nii_T1,nii_T1_mask,z=11,plot.type="
single") 

This example not show how to change or write NIfTI file.

Case Study About F-Test, F-Statistic Application

F-Test

This is the case study about f-test, f-statistic application.

Question:

Prices of  three brand fashion: Snapzi, Irisa and LoloMoons is "similar" or not?

We have the null hypothesis :

Note. So, null hypothesis in one-way ANOVA or f-statistic usually is "all of them is similar". And alternative hypothesis would be that at least 2 "guy" different.

This data is collected:

As the previous article we can fully use the t-test to answer questions. However the downside of the t-test is the only test for 2 groups, with the largest group of t-test was performed will be great, it will be by shaving of convolution of N 2 groups. So in this case we'll use the f-test.
Step 1 : Calculate the average price of each group, calculated the average price of the entire data, which calculated based on a formula SSbetween below.

SSbetween also called Between-Group variables Between-Group Variability or Variance of the group means or Sum of square Between-Group.

Step 2: Calculate SSwithin based on fomular:

Step 3: Calculate Degrees of freedom:

With dfbetween equal total of groups - 1 = 3 - 1 = 2

With dfwithin equal total of sample - total of groups = 12 - 3 = 9

Step 4 : Calculate Mean Squares,:

MSbetween =  SSbetween / dfbetween

MSwithin  = SSwithin  / dfwithin

Step 5:  Calculate f-statistic = MSbetween / MSwithin

To be able to conclude accept null hypothesis or not, we calculate f-statistic (with alpha = 0.05)

Find in table f-table(0.05) with dfbetween (df1) = 2 and dfwithin(df2) = 9

http://www.socr.ucla.edu/applets.dir/f_table.html#FTable0.05

We have:

We have f-statistic = 4.2565

Because f-statistic > f-critical-value, we can concluded that not accept the null hypothesis. That mean is: the price of three brand fashion: Snapzi, Irisa, LoloMoon is not the same.