That's really interesting. I'd heard the meme that they'd shown CBT led to changes in the brain--I think Wessely made that point somewhere. But I had no idea that they did the same thing essentially that Wessely and Chalder did in their most recent study--they measured changes before and after CBT and attributed the changes to the CBT. The comparison in the brain study was a group of healthy controls, measured at the same two time points but with no CBT. It makes no sense. You need a control group of patients who don't get CBT to say anything about whether CBT did anything.
In reading the abstract, I see where the fallacy might come in. The rightness of CBT is a given--they're not testing CBT. It's already been proven to work. They're testing whether the known benefits from CBT are also reflected in brain changes. If you know the CBT already works, then any brain changes that you can claim are linked to the specific cognitive problems facing patients were obviously also induced by the CBT.
