It's not going to happen in ten years, and I sort of know what I'm talking about (I've been working on neural models of language processing for quite some time now). First, the brain is huge, really huge. The number of "connections" isn't even known, let alone how everything interconnects. Second, even at the level of a single neuron, there are quite a few problems to be solved. If we could figure out the exact structure of a human brain, but emulate it just using integrate-and-fire type neurons, I predict it would fail miserably. Third, brains change. We don't know how we start out, but our brain changes from learning (if you want to be bewildered: development of cat vision is fascinating). For our brain to reach a stage in which it could be subject to a mental disorder takes some 15 years of normal life, i.e. inside a body that functions in a society. So even if someone could build a baby brain in 10 years, and incorporate it into a very human looking body, it would still take 15 years to develop and then might just not have developed a mental disorder.

Fourth, we don't have a clue as to where to look for mental disorders. Of course, if you have a virtual brain, you could tweak parameters, but not only are these parameters at this moment unknown, it would also be unethical (for it would hurt a person, see also other threads), and still this would not offer more insight into the emergent property that mental disorder is as comparing people with these disorders at this moment.

Concluding, this is a seriously flawed project.