Problem with voxel integration

Hi I want to integrate 4 matrices (4 volumes) of 128x128x128 int16 aquired in 4 time point. So, a voxel (k,l,m) of each matrice is a Y(k,l,m) of bi-exponencial time function. Thats the raw data that I have in hands.

The thing is that the funcions, that I know, intregrate the pages of a multidimensional array resulting in a 128x128 matrix. For my problem, that would not work.

One way that I thougt, is to build 128 matrices of 128x128x4 realocating all the voxels, do the trapz integration, resulting in 128 matrices of 128x128 and finally build all the pages together in a single multidimensional array. But that solution seems like too much trouble.

Is it possible to integrate the 4 matrices % voxel-Timepoint-1(k,l,m) to voxel-Timepoint-n(k,l,m) % resulting in a single matrix of 128x128x128 int16 using a vector of X values (time point) Ex: tp = [4 24 48 72]? If so, how?

Thanks for your attention.

[Andre Leitao]

Solved Had to "double" and CONCATENATE cat(4,t1,t2,t3,t4) all matrices into a hypercube (128,128,4,128) double Than easily integrate with trapezoid function, using a time vetor: trapz(x,y,4)