You should be able to use the U_offset and V_offset inputs on the brick node to essentially take samples using a different coordinate system. Set up the trigonometry to convert the normal U/V coords into a polar coordinate system, i.e. angle and radius from your center point, i.e. 0.5,0.5. Then use those polar coordinates to drive the u/v offsets into the brick node. That should get you something to start with, but you'll probably have the immediate artifact that the bricks near th edge will be wider than the bricks near the center. You could probably adjust that by using your calculated radius coord to drive the Brick_Width input so that the farther out you went (higher brick-V) the narrower the bricks (in polar coords). When projected onto the plane, it should even out. It's an interesting problem, and I'm tempted to fiddle with it myself, but I'm on another deadline right now. Maybe later.