Scientific American discusses our work on constraints on efficiency of neural circuits