Mind Uploading for Dummies

By Wil Forbis
June 1, 2015

I ask my readers a question: are any of you considering uploading your mind into a computer? If you are, I think you should be aware of some potential problems.

Mind uploading might sound crazy but the possibility is oft-discussed by futuristic scientists and psychologists who think it could arrive in mere decades. They argue that as we continue to study the functionality of the human brain (and, by extension, the human mind) we will find ways to replicate these functions in computer software.

Uploading one's mind is, in some ways, a desirable goal. An uploaded computer mind would lack many of the limitations of our biological "meat" brains. No more aging, no more physical degradation, no more death. The limitations of the physical universe could be "programmed away" in our new software environments. But mind uploading would also challenge our notions of what it means to be alive.

How would such a thing even be possible? To understand mind uploading we first need to understand what the mind is.

The Connectome
Science is in broad agreement that what we refer to as our minds - our thoughts, emotions and subjective experiences - arise out of our physical brains. This notion rather brusquely sweeps away any idea of a non-material soul or ectoplasmic entity inhabiting our bodies. Readers may take offense at this but the soul versus non-soul debate is lengthy and best fought outside the confines of this article.

How does our mind come from our brains? The brain is, as you may know, a complex network of around a 100 billion interconnected, wire-like cells called neurons. A single neuron can have multiple connections to its neighboring neurons; these connection points are called synapses. The total number of synapses in the brain is in the range of 100 trillion. Neurons become active with electrical signals and at each synapse an active neuron can pass its electrical charge to the connected neuron. Some of of these connections are excitatory in that they encourage neurons down the line to fire electrical signals, and some are inhibitory in that they discourage further electrical firing.

The point here is that the brain is seriously complex.

With this understanding in place, one could think of the brain as a grid of city streets governed by strange rules that direct traffic (comprised of not cars but electrical signals.) And how does one understand a city but with a map? A project is currently underway to map all the neurons and synaptic connections in the average brain; the name for this proposed map is the connectome. A connectome for the human brain has yet to be completed, though scientists have diagramed out the connectome of a worm's nervous system.

In and of itself, a connectome of the brain wouldn't tell us everything we need to know about a brain's functions. We need to know not just that neuron A shares a synaptic connection to Neuron B but also how likely A is to stimulate or inhibit B. This is known as the synaptic weight of a connection. We also need to factor in brain chemicals known as neuromodulators that float around in the brain and affect neuron firing.

If, using to-be-determined methods, we can create a connectome and gather associated information for a specific individual's brain, we will then (the theory goes) have a good understanding of how a that person's brain works. Could that information be used to upload that person's mind to a computer?

Building a Computer Brain
Readers unfamiliar with computer programming may be confused at how the information held in a connectome could result in an uploaded computer mind. Consider one of the most essential commands in computer programming: the If/Then statement. If/Then basically says, "If this condition is met, then do this action." One example might be, "If the user clicks the Go button on a web page, then pop up a particular image." Or, "If the user clicks the Launch button on an application screen then launch the United States' nuclear missiles towards China." For all the complexity of the human brain, similar rules pervade it. Brain logic might be, "If this collection of neurons that track the heat of my little finger are activated, then fire this different set of neurons responsible for pulling my finger away." Or, "If this collection of neurons responsible for recognizing my grandmother are activated, then activate this other collection of neurons that instigate hugging behavior.*" Like computer programs, your brain is essentially a series of receptors looking to fire at certain stimuli (the "If") and actions/behaviors that respond to received stimuli (the "Then.")

* Readers might balk at this description. "What about free will?" they ask. "Don't I chose to hug my grandmother, not simply obey the commands of my brain like a good little robot?" I confess that this author is quite dubious about free will, but keep in mind that the firing of neural networks can always be overruled by other neural networks. So, if you had some part of the brain that housed free will it could outvote a more reflexive part of your brain. The If/Then rules of the brain don't by themselves eliminate the possibility of free will.

From here we can construct a scenario in which one might upload a brain. First, scan a human brain for its synaptic connections and synaptic weights and construct a connectome map from that information*. Then use that connectome to replicate that brain as a computer software program. Have we not just uploaded a human mind?

* Please note the technology to perform such a detailed scan does not yet exist.

Well, no. As many have pointed out, we've simply duplicated a human mind. The flesh and blood owner of the brain will still possess the original version of that mind, and will eventually grow old and die and flesh and blood creatures do. No immortality here. The computer version of the mind is simply a copy, a clone.

Proponents of mind uploading have some intriguing thought experiments to respond to this problem. One goes along these lines. What if we took a human brain and somehow surgically replaced a single neuron with a mechanical or electronic device or even software program that behaved exactly as the original neuron did? The synthetic neuron would fire in all the same instances as its biological counterpart. (To put it another way, it would follow the same If/Then rules.) Would the person who owned the brain notice any difference? Now use this same process to replace a thousand neurons, a million, a billion, until... you got it, we've replaced the entire brain*. What would happen to the person's consciousness? Would they shift into some non-conscious robotic state? Or would their consciousness "transfer over" to the new materials? Nobody really knows. For all the advances made by science, we are still in the dark as to how consciousness---our subjective state, our experience of seeing, hearing, thinking and being---even arises out of the brain. Is it merely an emergent property of very complex networks? Would a software program with the complexity of a human brain also develop consciousness? Again, nobody really knows.

*For a few more scenarios in which the transfer on mind from body to machine could occur, read the first several paragraphs here.

The Mind as the environment
Now let's consider a view of the mind that's different from the connectome theory. This view is advocated by philosopher David Chalmers (among others.) He argues that the mind extends beyond the realm of the brain and goes into the surrounding the physical world. To grasp this notion, take stock of your experience right now. You are seeing things, probably hearing things, maybe tasting and smelling things if you're reading this over lunch. Your experience, your mindstate, would be very different without this particular outside stimuli. So, in a sense this stimuli is part of your mind.

Here's another way to think about it. The more popular "your-connectome-is-all-you-are" theory I first mentioned says that your mind arises based on various electrical signals zipping through the circuitry of your brain. But what happens when I look at an apple? Photons from the sun bounce off the apple into my eyes which connect directly to my brain. This results in the firing of neurons which somehow results in my subjective experience of seeing the apple. Is not this pathway of photons going from the apple to my eye similar to a pathway of a firing neuron? So is not every outside component (the apple, photons bouncing off it, etc.) part of my mind*?

* We may, at this point, be stumbling onto a bigger problem: the mind does not exist. At least it does not exist as a object or thing which is how we tend to think of it. Rather the mind seems an ethereal, hard to grasp set of subjective experiences with vague borders that we don't fully understand. How can one hope to transfer such a delicate machine?

If Chalmers is on to something then we have a problem with mind uploading. If we upload only the brain part of your mind, not the external environment, we are only uploading part of the mind. But maybe we could create a "synthetic" environment for our mind? Consider that your sense of reality---its sights, sounds etc---arises out of the firings of neurons. If we scanned a person's brain and isolated which part of their brain fired at the sight of a cat, and we then scanned and uploaded that person's connectome, we should have no problem presenting this neo-mind with the sight of a cat, or any other sight or sensation. (Well, if we ignore all the various caveats in this article.) This presumably would be how the denizens of the world from The Matrix saw their false realities.

You have already been assimilated
It would seem there are a lot of bridges to be crossed before we can confidently upload minds to a computer. But there may be a simple way around them. What if you are already a synthetic mind? Cosmologist George Smoot argues in a TedX talk that the odds are likely that you are such a mind. He presumes that humanity will eventually get the technology to create computer minds and eventually will be able to create them on a whim with the ease that computer game developers create accessory characters for video games. (We're talking about those non-player characters who tell you the goals of your quest etc.) If creating these entities becomes so easy that they outnumber biological entities by millions to one then logic would dictate that the odds are that you are such an entity. You might say, "But I don't live in the era of synthetic mind production, I live in the dark ages of the era before." No my friend, you only think you do. Smoot argues that you're akin to a non-player character in game set in the Middle Ages. You think you're in the past but you are not.

So I hope, dear readers, you tread carefully in the realm of mind uploading. All may not be what it seems.