Maybe these papers can bring some clarity into the matter:

http://www.sciencedirect.com/science/article/pii/S1053811911012286

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=7322909&tag=1

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6696237

Dear Brückner,

I appreciate for your quick response.

I could find the definition of cross-correlation coefficient in the very first review article. So the cross-correlation coefficient technically has the same meaning with Pearson's correlation coefficient. Also I found how to get cross-correlation image as well.

Thanks again, and have a good day!

Mr. Kim,

You will find the following review by Hyde and Jesmanowicz more informative in your analysis of cross-correlation coefficients in fMRI.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3305840/

The authors are originated the technique in the 1993 publication - "Processing strategies for time-course data sets in functional MRI of the human brain.", Mag. Reason. Med. 1993 Aug; 30(2)16-173.

http://www.ncbi.nlm.nih.gov/pubmed/8366797

Dear Mr. Morris,

Thank you for your valuable advice. I finally have the cross-correlation coefficients in fMRI. Also I became aware of how to calculate cross-correlation image, which is simply obtained by the dot product between time course data and boxcar function. It was exciting that cross-correlation image after some trimming process gives quite similar activation map to the fMRI map obtained from T-test!

At this point, a new question arises. The resulted calculation of cross-correlation image shows the intensity difference between image with and without stimulation. But the map is not quantitatively described. Is there any way to quantitatively show the fMRI cross-correlation image? Or is the method of simple difference commonly accepted by the fMRI field?

After studying about magnetization transfer, this new field inspires me and makes me curious about many things.

Thank you for your help.