Remove items from 3D Matrix with 2D mask for each timestep

Hi,

I would like to mask a 3D matrix (A) using a 2D Mask-Matrix containing several NaN's (B). The mask should be applied for each 3rd Dimension step (here: timestep) of A. I have a loop solution based on a 2D approach, but would rather like to use a logical approach.

Can someone recommend a logical approach?

Matrices are for example:

% data
A(:,:,1) = [1 2
            2 4]
A(:,:,2) = [5 2
            2 3]
B_mask(:,:) = [1 2
               NaN 1]
% the loop solution based on 2D
for i=1:2
    AA=A(:,:,i)
    AA(isnan(B_mask)) = nan; 
    A(:,:,i)=AA;
end
% solution
A(:,:,1) = [1 2
            NaN 4]
A(:,:,2) = [5 2
            NaN 3]

Answer 1

[KSSV]

% data
A(:,:,1) = [1 2
            2 4] ;
A(:,:,2) = [5 2
            2 3] ;
B = [1 2
               NaN 1] ; 
B_mask = repmat(isnan(B),1,2) ;            

A(B_mask) = NaN 

A = 
A(:,:,1) =

     1     2
   NaN     4


A(:,:,2) =

     5     2
   NaN     3

[HL_2009]

Thanks, that works.