Actually, the mental illness/intelligence correlation is probably just observer bias. There’s an article in that latest Skeptic about it.
We don’t really know enough about what genes control intelligence to answer this, but testing over years of public school students in the US have demonstrated that boys are over-represented of both ends of the bell curve of intelligence. Lots more smart men than women, but lots more stupid men too, and this held true when you removed people who were retarded with known etiologies.
I am not a geneticist, but I read an article by someone who was, who speculated that there are probably genes on the X-chromosome that contribute to structures involved in cognitive skills, that, when they go wrong, men don’t have a second chance to get the correct information from a second X-chromosome, and that usually when things are off, there are problems, but that there may be occasions when exceptional genes on the X-chromosome produce advantages, and in women, they get “corrected” by the normal X (except in the rare occasion where a woman might get two Xs with the same advantageous mutations).
That means there’s an implication that there’s an"ideal" intelligence that isn’t as high as possible, but closer to the mean, which may make sense, since the brain uses a lot of energy, and is subject to a lot of things going wrong in utero, so there’s probably a balance between “simple enough to make in nine months,” and “complex enough to function.”
But anyway, if the primary symptom of something were increased intelligence, it wouldn’t be a problem, which is the real issue with the question. High intelligence isn’t studied as a pathology, or even really a symptom. It’s not usually something you can check on the list of symptoms. If people with a particular heart condition or arthritic condition tended to score high on intelligence tests, it would likely go unnoticed, unless it were every single patient, and astronomically high. People with pathologies are watched for co-morbidities, in case a syndrome is present, but their strengths are not tracked. It’s just the way both research and treatment work.
It took years for someone to notice that left-handed people with learning disabilities tend to write with a “hook,” while those without LDs tended to write as a mirror image of right-handed people, in spite of the fact that people had know for decades that left-handers were slightly over-represented among the learning disabled (most learning disabled people are still right handed, though), and for a couple generations, people were saying “Isn’t there some way to predict as soon as kids start writing which lefties we should keep an eye on?” NOTE: not 100%; it is certainly possibly for a mirror-writer to have an LD, and there are probably hook writers without them, especially if some right-handed teacher ever did some bad hand-over-handing with them when they were little, and caused the hook. But it was only about 10 or 12 years ago that someone actually demonstrated this, in spite of the fact that resource rooms have been around since the late 1960s, and by law since 1974, and people have had a chance to observe this for 30 years.
So if it takes that long to make such an obvious connection, think how much longer people studying pathologies and deficits are going to take to make a connection between a pathology and a great ability.