roipoly creates a 2D mask, and 2D images are either black and white, grayscale, or pseudo-color. If the result you get is pseudocolor, you could adjust the color map so that the high values mapped to color, but then all the high values would map to color, including ones that were already high in outlineMat.
What you could do, if you do not need to do further grayscale processing, so something like,
roi = roipoly(zerosMat,x,y);
outlineMatC(:,:,1) = outlineMat + R * roi;
outlineMatC(:,:,2) = outlineMat + G * roi;
outlineMatC(:,:,3) = outlineMat + B * roi;
where R, G, and B are the appropriate color components of the color you want the ROI to become.
Yes, this could be shortened to one or two lines, but they are not going to be immediately understandable
roi = roipoly(zerosMat,x,y);
outlineMatC = repmat(outlineMat,1,1,3) + cat(3, R*roi, G*roi, B*roi);
Sometimes clarity is best.
