The cerebral cortex exhibits a compression gradient in neuronal architecture, with smaller neurons representing low-level features in primary sensory areas and larger neurons with more connections in the frontal cortex representing higher-level features.

“if you take the cortex off the surface uh the cortical sheet off the that wavy you know cortical sheet you dig it off the rest of the brain the subcortical parts and you stretch it out like a napkin you can see there's a compression gradient there in the architecture of the neurons so at the primary sensory areas there are these tiny little paramal neurons that are representing these little these very low-level features and they feed into bigger neurons which feed into bigger neurons which feed into more bigger neurons so what's happening is you've got this very detailed array being impressed in its dimensionality until you get to the middle of the brain at the front where there are many fewer neurons but they're bigger and they have many more connections so it's a dimensionality reduction that's happening”