Find the mean difference and 95% confidence interval, between two 3D matrices

Hello guys,

I want to find the mean signal intensity difference between two 3D medical images, the p value and the 95% CI.

The image matrices are:

A=[256,768,160] and B=[256,768,160]

Each of these, is made up by horizontal concatenation of 3 matrices, with size: [256,256,160]. I can get the mean value of A and B like so:

mean_A = mean(A(:));
mean_B = mean(B(:));

The difference then is:

mean_diff = mean_A - mean_B

Ultimately, I would like to determine whether the mean intensities of A and B are significantly different, by using a statistical test. After trying

[h,p,ci,stats] = ttest2(A,B)

I get the following matrices for p value and 95% CI

p=[1,768,160];CI=[2,768,160];

Do I have to convert A and B into 1D vectors, before using ttest2, to get a single answer? And how can I achieve the latter? As a final note, I believe that the Wilcoxon rank sum test, is more appropriate in my case as the signal intensities are not normally distributed. However, I do not understand how to pass A and B into the Wilcoxon matlab function in order to get one p value and one 95% CI.

Many thanks for your suggestions.