Data distribution randomness test in Matlab

I have a dataset describing the point distribution in a 2D space. The row and column index of a data point in the matrix corresponds to its coordinate in the 2D space.

After plotting out these data, I can see four general clusters, one up, one down, one left, one right. I'm wondering how I can use chi-square to distinguish such data cluster with random distribution? I heard about permutation method, how can I combine it with chi-square?

Many thanks for any suggestion

Kyle