Possible, but I'm guessing that the massive amount of simulator time, muscle memory, and practice are probably closer to the true explanation.

In a realistic simulation, which Gran Turismo aims to be, being more aggressive than in real life is not an advantage.

Apart from the cost of training, where the GT players have the advantage is that failure is not as penalizing as it would be in a real car. If you crash, you can be back on the tracks in seconds for no cost and no time in traction. If you mess up a corner you can restart.

Rather than desensitizing them to danger, it allows for exploratory driving, which, especially when combined with the amount of track time available, would allow skills to be learned much faster than in the real world.

That is to say, the learning processes for someone driving a simulator and someone driving a real car are not the same, but these approaches must converge to a common end point which is the best driving style in order to be the fastest car on the track.

That always stood out to me in Ender's Game. Especially when Ender had to learn the lesson to not be so flippant with the units under his command.

That doesn't translate. I spent considerable time playing racing simulations, and still my first time on a real track was an intimidating experience.

The issue isn't how you felt, it is what kind of behavior you learned in the simulation context which would transfer to what you did on the real track. Doubtless your intimidation affected your behavior, but if a simulation can improve real performance then it is unlikely that a little intimidation completely erases all the tendencies learned in the simulation. It can't be assumed that things which improve performance will transfer, but things which could be dangerous won't transfer.

In my case my hands were numb and in pain from the strong grip.

After a while (days? weeks?) you get used to the real car and the racing simulation muscle memory comes back.

One personal anecdote may well be true and valuable, but it still isn't usable data.

Normally this rejoinder is used by people who have provided data and received an anecdote in response. However, you haven't even provided an anecdote . . .

Well, no, of course not! If you look above, I asked a question. In case you are wondering, it wasn't a rhetorical question.

My estimated likelihood that you pose questions in order to make subliminal rhetorical points and so tend to readily project that behavior onto others would have increased, but one anecdote isn't usable data.

EDIT: Also, I see I forgot the question mark, which may have caused confusion.