“more closely resembled that of non-transgender boys and girls of their desired gender”
I've looked into lots of studies like this. And when you dig into the details, the picture painted is like this:
Draw a line representing whatever particular sex difference in the brain they are attempting to measure. At one end, the male brain, at the other the female brain like this:
100__0
With male at 100 and female at 0
The male study subjects who identify as trans are positioned at 90. They've moved in the direction of the female end of the line. (Usually with the help of opposite sex hormones and/or by blocking their own.)
One can now claim that those positioned at 90 more closely resemble the result of the opposite sex than those positioned at 100.
Well, yes they do. What is not accurate is the claim that those at 90 are more similar to those at 0 than they are to those at 100. However, that's the claim often made. It's hard to spot, because the differences tend to be minute (the ends of the line aren't really 100 and 0 but at something like 56.7823 and 56.7896 with the subjects who identify as trans situated .0005 or so away from their own sex average).
Add to that the lack of controlling for sexuality, the tiny number of study subjects, and the rather hyperbolic way of presenting artificially produced minute shifts towards the opposite sex in people identifying as trans as evidence of well, anything, really, makes this an even more ludicrous claim than it already is.