Re:what about the other 10% (Score 1)

Well, I tracked down the actual article and the accuracy is actually quite overstated. One of the problems in creating diagnostic tests is setting all of the dials to optimize performance on the particular dataset. The 90% rate is an overall rate. Note that if you did something stupid like saying no one had PTSD then you would be right in their dataset 250/324 = 77% of the time. Their optimized performance (which would probably never be repeated in another dataset) produces a lot of false positives (31/103 = 30%). While PTSD is a very important area and this study demonstrates some differences in brain activity between those with PTSD and those without which demonstrates it is a real diagnosis, it is not a fool-proof diagnostic technique.