feedforwardnet with matrix input

Hi, i need to create a feedforward neural network with 8 input nodes and 5 output nodes.

I need to recognize a pattern that is composed by 10 samples, so :

Pattern 1 is composed by 8 channel (8 sensors) and the single pattern is represented by 10 elements, is a matrix 8x10 for a single pattern. (and the same is for the others 4)

I have registered 5 differents signals for each pattern so i have 25 inputs where each one is composed by a 8x25 matrix. In output i have only 5 nodes because i want to recognize 5 different patterns.

How can i format the data for the network training please? i'm getting crazy with this porblem