Questions about correlation analysis(statistical analysis) in DPABINET

Submitted by DHLee on

Hello, Prof. Yan~!

Previously, I performed correalation analysis using 'statistical analysis menu' in 'DPABI' to calculated R maps(in this case, I calculated voxelwise R maps between fALFF maps and clinical measures). 

Now I am trying to perform correlation analysis using 'statistical analysis menu' in 'DPABINet' to calculate R matrix(containing r-value matrix with clinical measures, ex. 160x160 matrix) associated with clinical measures. 

Therefore, this time, I try to do correlation analysis in edge level, not voxel-level(I try to perform NBS later). 

After the correlation analysis, the outcome files are two: R, R_Cohen_f2(both matlab data). 

but I felt strange about the contents of R(matlab data): NetworkMatrix(160x160), StatOpt(TestFlag:'T')

I thought that the NetworkMatrix should indicate 'Correlation values(Range: -1~+1)', but the results ranges from -3.XX ~ +3.XX(similar with t-stat). 

but I definitely performed 'correlation analysis'('Group images': patient's 160x160 matrix, 'Correlation seed series': clinical measures of the group, below image). 

In this case, I wonder if the results of the analysis is appropriate. and How I interpret the results of correlation analysis? 

I look forward to hearing from your opinion. Thank you. 

 

Attachment Size
캡처.PNG 43.38 KB