It's not a new idea, it's only fairly recently that we'd have the ability to even map out the complexity of a mouse brain.
Of course the entire thing will be slowed down by having to simulate connection speeds between neurons, when some simulated neurons are on the same chip, and others are on chip. (I'm going with the idea memory chips rather than something like a hard drive).
Of course, organic brains, even something as simple as a mouses brain, is operating in parallel that exceeds any computer the researches can potentially afford at this time. So to deal with that, they're going to have to keep things synchronized by doing a limited number of calculations at once, and go through the states of every neuron for a single tiny fragment of time, so the system will be slowed down to a great extent due to this.
I have no idea how much all this extra time will take, but it could as crazy as one second of mouse time could be hours, days, or even years of our time.
A mouse brain has about seventy million or more neurons.
Now we can't forget that neurons fire at different levels, which might change how the signaling in the brain works, that's something we're still trying to quantify, but it's not a binary thing of on/off.
We know that the synapses move and change, which is effectively the wiring of the brain, and we don't know the rules that control that.
Memory, which is used to influence a lot of actions, is still mostly a mystery to us, but we do know that it involves breaking/changing DNA, something we can't do with hardware, but will have to have yet another abstraction layer with it's own data in the simulation.
Even if we get all that stuff worked out, we still have a LOT to learn about the brain and what it's composed of. There are some ideas that certain structures in the neurons function as quantum processors, and there's not one or two per neuron, it's more. (An article I saw last year indicates more than a dozen in a human neuron, but that was last year that I read it.) If they actually do have quantum processing, then not only does that brain get's so much more complicated and would now require a quantum computer to process, and our current quantum computers are horrifyingly primitive and only handle a hand full of qbits.
Now despite all those drawbacks, their research certainly holds value, but people expecting to be able to make sci-fis brain dumps/ brain tapes/ mental clones/ etc. are not only flat out of luck, but looking wistfully for something still deep in the realms of fiction.
