These structures are essentially just extensions of images, but hyperspectral. You can view them as 2d images and use something (sum intensities and grey scale?) in the 3rd dimension to represent the spectral profile over the range (6 in your case). For example, in RGB images the dimensions are mxnx3 where the 3rd dimension hold the RGB values for each element in the mxn matrices.
I am not well clear with the question since you have quoted as ''their values''. But if I have understood your problem correctly from your example, then you need a representation of x(m,n,p) matrix.
Now representation means you have to show the variation of x(m,n,p) matrix either through a plot or an image, but with respect to what ? For example, if you need a plot representation, you have to decide which has to change with respect to other. Else in image plotting, you can use surf, mesh, patch to look-up your data as (m x n) with p-representing the RGB/Gray color.
If you need to show a data like x1 (m,n,p), x2(m,n,p), etc... with varying m, n and p sizes, then it will be better to use a time series plot rather than a image plot.
With a detailed question and data format, I will be able to answer to the point.