iirc, "G" is the fundamental "intellectual capacity pool" from which all measurable characteristics (creativity, analysis, etc...) are derived.
I have yet to see a definition of the neurostructural basis of "G" that could be used to compare neurostructural differences between individuals and accurately predict differences in raw intellectual capacity - with the sole possible partial exception of density of syanaptic interconnection within the neocortex.
I suspect that "G" is likely to be found to be bunk as neuroscience trundles along on its current trend of exploding 19th Century notions concerning the psyche and intellect.
Thank you.