I am trying to insert a vector into a 3D matrix at specified
locations based on the values in a 2D matrix.  I am
currently  looping through the row column positions obtained
from the 2D matrix and reshaping the 1D matrix to assign it
to the 3D matrix.  

   My2D = round(10.*rand(10,10));
   My1D = rand(100,1);  
   My3D = zeros(10,10,100);

   [r c] = find(My2D == 5);
   d = length(My1D);

   for ix = 1:size(r)
      rx = r(ix);
      cx = c(ix);
      My3D(rx,cx,:)  = reshape(My1D,1,1,d);  
    end

This code works fine but I would like to know if there is a
way to speed up this assignment by using the single index
vector from the find statement to assign multiple 1D vectors
 to the 3D vector in a single step?

Thanks
Rob

bsxfun(@plus, ...
  (0:size(My1D,1)).*(size(My3D,1)*size(My3D,2)), ...
  sub2ind(size(My3D),r,c,repmat(1,10,1)))

to which you would assign

repmat(My1D,1,size(r,1)).'

You can do this in a single step without reshaping,

My3D(bsxfun(.....)) = repmat(....);
--
   "No one has the right to destroy another person's belief by
   demanding empirical evidence."                -- Ann Landers

My2D = round(10.*rand(10,10));
% change to a 3d array
My1D = rand([1 1 100]);  
My3D = zeros(10,10,100);

[r c] = find(My2D == 5);
My3D(r,c,:) = repmat(My1D,[length(r) length(c)]);

that does the assignment directly and doesn't use a for
loop.

hope it helps!