Let's assume that 20 minutes into the future we finally get the hang of restoring the human brain following extreme head trauma. How much of that person's old memories would the newly regenerated brain retain? Some of them? All of them? What if the hippocampus specifically is damaged or destroyed?
I don't think the mechanism of memory storage is well understood. People here are saying that memories are stored in the connections between neurons. While this might be true in the sense that destroying those connections destroys memories it's probably quite a bit more complicated than that.
Imagine you meet me at a party. You remember my name and occupation while we're talking. That's short term memory. Now imagine I made an impression on you, because of course, I'm funny and charming. I might make it into your long term memory. It's known that sleep is a critical part of this process (because if you don't sleep enough your ability to form long term memories suffers) and that the hyppocampus is necessary (because if you cut it out like they did in the case of HM, you can no longer create long term memories) But we have no idea what that brain structure does or how it does it. We literally have no idea precisely where or how long term memories are stored. Any assertion here will be speculation. If we did know we'd be half way to reading storing and even creating and implanting memories which would be hugely powerful and terrifying. Once we unlock the storage mechanisms of the brain we'll be able to read a whole mind and theoretically store it, run it on a computer, write it to another mind, modify it along the way, implant knowledge and skills, create fake memories, run minds in simulations, everything. Sometimes I wonder if the reason we're not able to do cool stuff is that we'd inevitably cock it up. Maybe in all the possible worlds where this is possible the world ends in fire and pain and the reason the world still exists is that this earth hasn't discovered the secret of godlike powers.
Memories are stored in the pattern through which neurons connect. Those patterns are formed during the experiences that the brain records. If those patterns are lost, the memories are gone forever. If the synapsis in those connections just aren't firing, though, then the memories can be restored. The neurons just have to start firing again.
A best case scenario would be one where you can have brain implants that make an interface between brain and machine. See Johny Mneumonic, and the Neuromancer trilogy for ways in which memory may be backed up electronically, and Ghost In The Shell for a plot twist where memories - even fake ones - can be moved from machines to biological brains.
How much of that person's old memories would the newly regenerated brain retain? Some of them? All of them? What if the hippocampus specifically is damaged or destroyed?
Memory storage is poorly understood, but we do know that different areas of the brain store different information, and that recall happens by association - memories are stored, at least in part, "holographically", and sometimes distributed in both emispheres (as a contrived example: the shape of a house on the left, its colour and smells on the right).
Depending on the damage, the victim could:
- experience selective (field) amnesia, where all knowledge about certain topics, or details, but not others, is either lost or unretrievable
- experience procedural amnesia, in which some skills are lost (e.g. speech, or the knowledge of one or more languages)
- experience episodic and/or declarative amnesia (the victim no longer remembers his home, or the text of the Declaration of Independence).
Some even more complex kinds of amnesia might ensue: for example the inability of remembering most things unless they're put into rhyme or sung (whilst they all seem to be words, the brain treats speech, poetry and song differently; this is most easily seen with people that may stutter when speaking normally, yet being able to sing or recite unimpaired).
With a whole half of the brain gone and regenerated blank, chances are that the victim would experience all of the above in some measure. For procedural amnesia, which side of the brain gets damaged matters: if the dominant emisphere (usually the left) is damaged, Broca's aphasia will arise - the victim is perpetually "lost for words", and possibly aural agnosia as well - the noncomprehension of speech. The victim will have to re-learn his mother tongue, which begs the question - is the newly made brain tissue still plastic enough? Otherwise he'll speak his own language as a foreign language.
For the same reasons, expect significant changes of mood and character.
A part of the damage will not be to the memories themselves, but to the knowledge of having them - the capability to recall them at will. This may lead to someone ignoring everything about cars (or starships) to the point of not recognizing one, and yet perform flawlessly once seated at the controls, so long as they act on instinct and not on the memory they no longer have. If, as we know it can happen, a memory is in the undamaged half of the brain while its recollecting associations were in the damaged half, the victim will feel some topic is "familiar" and not remember ever having studied it.
Damage to the hippocampus will in all likelihood not matter, if the hippocampus is replaced; otherwise you'll get partial anterograde declarative amnesia, i.e. the incapability of learning/remembering new things, while the capability of learning new skills will remain intact. If the nondominant emisphere hippocampus was damaged, you might even not notice anything except a slight slowness in learning.
I once knew a man who suffered a degree of diffuse brain damage after an event that would have been fatal without CPR. He made a full recovery in about a year both as far as he could tell and as far as friends could tell.
During his recovery he was repeatedly frustrated by "knowing he knew" something but not being able to access it. It could be as little as one word, intermediate like how to program in Pascal, or huge like most of his life as a teenager.
Mostly the memories returned, which suggests they were still there but no longer properly connected, like (crude analogy) a computer disk with a corrupt file index. Sleep is known to be required to lay down memories. It is also known to be involved in forgetting. The same processes may discover disconnected memories and reconnect them.
Obviously he could not tell how many memories were recovered internally and how many were patched back from external sources.
Not really an answer, but this may be a useful insight. The issue for replacement of a large chunk of brain tissue with new "blank" tissue is to what degree memory is distributed. We are evolved to deal with single brain cells dying every day, and to have a reasonable ability to survive individual small lesions and micro-strokes (often without noticing). My colleague lost some years-worth of random brain cells in a few minutes of oxygen deprivation. But there are brain functions that are not fully distributed, such as motor functions. Lose all the tissue containing such a vital localised function and the damage may not be reversible merely with replacement tissue.
There is a lot that we do not know about brains!