Once upon a time, Dr. Wilder Penfield happened to cut open a patient's head and poked a part of his brain, which in turn caused the fellow to mutter something about seeing his grandmother (long dead) coming towards him with a freshly baked pie. It's worth noting that Penfield was a well respected neurosurgeon and the man having visions of nana and her baked goods was on the receiving end of a surgical procedure. Then again, when your name is "Wilder," it's pretty much a guarantee that your life will be interesting. I mention these things because Penfield went on to replicate and write about this experience, providing arguably the first real evidence for the engram, a physical manifestation of memory in the human brain.
Penfield went on to revise his hypothesis, and modern psychologists have tossed most of it away entirely, but sixty-some years ago it was a compelling case for the idea that your memories of grandma could be contained in a single cell in your cortex. Stimulate that cell and you call forth the memory, pie and all. Poke a different cell, you get a different memory.
It's a fun idea, but it falls apart quickly enough as you balance the estimated 84 billion neurons in the human brain against just how many and different kinds of memories we have, how they're related to one another (or not), and what facilitates access of some but not others. But credit where it's due, Penfield's work is likely responsible for much of what followed, even if that subsequent research was motivated by a desire to demonstrate just how wrong he was.
Nowadays, the idea of grandma cells has been replaced by patterns of activation. I'll leave it to the mathematicians among you to work out the possible combinations you can achieve with as few as 100 neurons out of an available 84 billion (and that 100 is a totally arbitrary number). The point is, you can make a much more compelling case for a pattern of firing among a specific group of neurons holding your memory of grandma (and her freshly baked pie).
Speaking as a cognitive psychologist, I quite like this model. It suggests that everything we know, everything we've experienced, pondered, dreamed, and imagined can be described as patterns of neural activity, that at the end of the day the thing that is each of us is a unique, individualized and highly organized collection of information. But speaking as a science fiction author, I'm not happy with this at all. Implicit in this idea is the limitation that because that collection of information has its basis in your wetware, it dies with you.
That grates against my aesthetic as an author. It seems... wasteful, even pointless. And I've been arguing against it from the very first story I sold. It was a little thing about the last survivor of Atlantis, a chronicler, who kept people alive by writing about them in an enormous book. That story was turned down by a lot of editors, including one very recognizable name (no, I won't tell you whom, it's not relevant) who replied that the story was about futility, and people didn't want to read futile SF.
But that was kind of my point -- even if I didn't have a clear grasp on it in my own mind at the time -- that the unique collection of information that defines each of us should not cease just because we've worn out the meat we're born into.
I've revisited this theme in subsequent fiction. It shows up several times in the collection of stories and novels that make up my Amazing Conroy universe, where Conroy (he's a stage hypnotist performing for aliens) encounters a member of the Svenkali, an ancient race that can invoke the personality and memories of any of their dead simply by speaking their unique name. The Svenkali consider this a kind of immortality, one that is part of their birthright, secure in the knowledge they'll be brought back as guests, visitors, teachers, and mentors for successive generations.
More recently I took this to the next level. In my latest novel, Barsk: The Elephants' Graveyard (Buy in hardcover/Kindle ebook)), I offer up a more "scientific" explanation, positing a previously unknown subatomic particle (let's call it a "nefshon") and inventing a new branch of study that was part physics, part psychology.
Consider that the process of laying down new memories and experiences involves coding the information associated with those events on these nefshon particles, that they are in fact the material correlate of the neuronal network of activation created for each thing in memory. Being particles, space isn't really an issue, and like the old Dorrito's slogan, you can always make more.
More importantly, we're talking about a particular kind of memory here, the one that psychologists like to call "episodic memory," the store of events that you were a part of -- as distinct from say, "semantic memory" which is more encyclopedic, things you could look up on Google with your smartphone. So, knowing that Charlie has a blog is semantic memory, it's a fact and out there in the world. But that you read this particular post on the blog, that's episodic; it happened to you, and encoded along with that event is your reaction to the post, where you were when it happened, who else was around, maybe what you were wearing or eating or listening to in the background. Episodic memories represent your full sensorium of an event, both in terms of the bottom-up experience of the present stimuli as well as the top-down associations the event triggers at that moment in time.
And I'm claiming that all of that gets encoded into your nefshons.
Now here's the fun part. The information encoded in nefshons exist independently of the source of that information. Consider starlight. The twinkling bits you're seeing in the night sky have been traveling long and far to reach you; in some cases the source of their light, their information, has ceased, been snuffed out or gone nova or collapsed in upon itself or used as fuel by the First Order's Starkiller base, take your pick. The point is the particles don't care. They're busily traveling, minding their own business, and if somewhere along the way there's someone with the right instrument or sensory apparatus to recognize and decode and perceive their information, then Bob's your uncle and you get to see that that bit of reflected light from halfway across the galaxy. Pretty sweet.
Now let's go back to nefshons, particles of memory (and by extension, personality). At the end of the day, specifically the end of your last day, even though you may be destined to be worm food your nefshons continue on without you. You're no long holding on to them and they begin to defuse, fast or slow, throughout space.
In Barsk, I conveniently provide a drug that temporarily grants the ability to perceive and manipulate nefshons, allowing users (whom I call Speakers) to pull together a sufficient quantity of one person's particles to recreate a simulacrum of the deceased and actually converse with them. The idea here is the same as that starlight we were just talking about. The organization of information that made you unique, that was a description of who you are (or were) as a human being, still exists and is just waiting for someone with the tools to interact with it. Maybe it happens (as it does in my novel), maybe it doesn't (as in that vast warehouse of aging VHS cassette tapes that lack a player). It's the tree falling in the forest, the mechanism allows for immortality, whether or not there's someone there to summon you and have a chat long after you're gone.
It's a different way of looking at memory, which was part of what I wanted, and of course it begs one critical question: if a Speaker conjures you up post mortem and has a chat, will you remember that conversation the next time it happens?