Wait, is that literally true? Is it true in general that, when you think about a shape S, an actual S-shaped region of your brain lights up?
No. Even in the lowest visual region (where there literally is one incoming neuron for each cone in the retina) the image is very badly distorted, as though by a fisheye lens. And when you respond to a high level instruction like “think of a square” I would bet against there being any activation at all in this region, it is all abstract representations where the patterns of activation are unrelated to the shape of a square.
No. Even in the lowest visual region (where there literally is one incoming neuron for each cone in the retina) the image is very badly distorted, as though by a fisheye lens. And when you respond to a high level instruction like “think of a square” I would bet against there being any activation at all in this region, it is all abstract representations where the patterns of activation are unrelated to the shape of a square.