Could training AIs on human brain activity be key to AGI?

So wouldn't an AGI be able to mimic human intelligence if it learned accurately enough which brain regions show what activity in different settings?