Unpaired, dependent data comparison, or comparison of paired means

[nannamp]

Hi

I'm trying to find a statistical test that can help me compare the means of two dependent groups.

My data:
1: parameter values from white matter region in an in vivo rat brain
2: parameter values from white matter region in the same rat brain, bur after the tissue has been fixated.

I cannot align/register the two brains to each other, so the region I have outlined in vivo is not exactly the same as the region I have outlined in the fixated tissue.


My wish: I would like to compare the means of the two white matter regions.
My suggestion: A paired t-test, but as I know that each pixel value in data set 1 does not exactly correspond to the pixel value in data set 2, the paired t-test is not exactly right to do.

Is there a test of dependent but NOT paired data?

Thank you for your help!

Nanna

 

[Miner]

Are we talking about a single brain, or is each datum taken from a different brain? Paired data do not mean exactly the same point of measure. For example, the classic shoe wear example used in Statistics for Experimenters by Box, Hunter and Hunter deals with the wear of an entire shoe, not a particular spot on the shoe.