Log in

Previous Entry | Next Entry

The notion of "uploading"--analyzing a person's brain and then modeling it, neuron by neuron, in a computer, thereby forever preserving that person's knowledge and consciousness--is a fixture of transhumanist thought. In fact, self-described "futurists" like Ray Kurzweil will gladly expound at great length about how uploading and machine consciousness are right around the corner, and Any Day Now we will be able to live forever by copying ourselves into virtual worlds.

I've written extensively before about why I think that's overly optimistic, and why Ray Kurzweil pisses me off. Our understanding of the brain is still remarkably poor--for example, we're only just now learning how brain cells called "glial cells" are involved in the process of cognition--and even when we do understand the brain on a much deeper level, the tools for being able to map the connections between the cells in the brain are still a long way off.

In that particular post, I wrote that I still think brain modeling will happen; it's just a long way off.

Now, however, I'm not sure it will ever happen at all.

I love cats.

Many people love cats, but I really love cats. It's hard for me to see a cat when I'm out for a walk without wanting to make friends with it.

It's possible that some of my love of cats isn't an intrinsic part of my personality, in the sense that my personality may have been modified by a parasite commonly found in cats.

This is the parasite, in a color-enhanced scanning electron micrograph. Pretty, isn't it? It's called Toxoplasma gondii. It's a single-celled organism that lives its life in two stages, growing to maturity inside the bodies of rats, and reproducing in the bodies of cats.

When a rat is infected, usually by coming into contact with cat droppings, the parasite grows but doesn't reproduce. Its reproduction can only happen in a cat, which becomes infected when it eats an infected rat.

To help ensure its own survival, the parasite does something amazing. It controls the rat's mind, exerting subtle changes to make the rat unafraid of cats. Healthy rats are terrified of cats; if they smell any sign of a cat, even a cat's urine, they will leave an area and not come back. Infected rats lose that fear, which serves the parasite's needs by making it more likely the rat will be eaten by a cat.

Humans can be infected by Toxoplasma gondii, but we're a dead end for the parasite; it can't reproduce in us.

It can, however, still work its mind-controlling magic. Infected humans show a range of behavioral changes, including becoming more generous and less bound by social mores and customs. They also appear to develop an affinity for cats.

There is a strong likelihood that I am a Toxoplasma gondii carrier. My parents have always owned cats, including outdoor cats quite likely to have been exposed to infected rats. So it is quite likely that my love for cats, and other, more subtle aspects of my personality (bunny ears, anyone?), have been shaped by the parasite.

So, here's the first question: If some magical technology existed that could read the connections between all of my brain cells and copy them into a computer, would the resulting model act like me? If the model didn't include the effects of Toxoplasma gondii infection, how different would that model be from who I am? Could you model me without modeling my parasites?

It gets worse.

The brain models we've built to date are all constructed from generic building blocks. We model neurons as though they are variations on a common theme, responding pretty much the same way. These models assume that the neurons in Alex's head behave pretty much the same way as the neurons in Bill's head.

To some extent, that's true. But we're learning that there can be subtle genetic differences in the way that neurons respond to different neurotransmitters, and these subtle differences can have very large effects on personality and behavior.

Consider this protein. It's a model of a protein called AVPR-1a, which is used in brain cells as a receptor for the neurotransmitter called vasopressin.

Vasopressin serves a wide variety of different functions. In the body, it regulates water retention and blood pressure. In the brain, it regulates pair-bonding, stress, aggression, and social interaction.

A growing body of research shows that human beings naturally carry slightly different forms of the gene that produce this particular receptor, and that these tiny genetic differences result in tiny structural differences in the receptor which produce quite significant differences in behavior. For example, one subtle difference in the gene that produces this receptor changes the way that men bond to partners after sex; carriers of this particular genetic variation are less likely to experience intense pair-bonding, less likely to marry, and more likely to divorce if they do marry.

A different variation in this same gene produces a different AVPR-1a receptor that is strongly linked to altruistic behavior; people with that particular variant are far more likely to be generous and altruistic, and the amount of altruism varies directly with the number of copies of a particular nucleotide sequence within the gene.

So let's say that we model a brain, and the model we use is built around a statistical computation for brain activation based on the most common form of the AVPR-1a gene. If we model the brain of a person with a different form of this gene, will the model really represent her? Will it behave the way she does?

The evidence suggests that, no, it won't. Because subtle genetic variations can have significant behavioral consequences, it is not sufficient to upload a person using a generic model. We have to extend the model all the way down to the molecular level, modeling tiny variations in a person's receptor molecules, if we wish to truly upload a person into a computer.

And that leads rise to a whole new layer of thorny moral issues.

There is a growing body of evidence that suggests that autism spectrum disorders are the result in genetic differences in neuron receptors, too. The same PDF I linked to above cites several studies that show a strong connection between various autism-spectrum disorders and differences in receptors for another neurotransmitter, oxytocin.

Vasopressin and oxytocin work together in complex ways to regulate social behavior. Subtle changes in production, uptake, and response to either or both can produce large, high-level changes in behavior, and specifically in interpersonal behavior--arguably a significant part of what we call a person's "personality."

So let's assume a magic brain-scanning device able to read a person's brain state and a magic computer able to model a person's brain. Let's say that we put a person with Asperger's or full-blown autism under our magic scanner.

What do we do? Do we build the model with "normal" vasopressin and oxytocin receptors, thereby producing a model that doesn't exhibit autism-spectrum behavior? If we do that, have we actually modeled that person, or have we created an entirely new entity that is some facsimile of what that person might be like without autism? Is that the same person? Do we have a moral imperative to model a person being uploaded as closely as possible, or is it more moral to "cure" the autism in the model?

In the previous essay, I outlined why I think we're still a very long ways away from modeling a person in a computer--we lack the in-depth understanding of how the glial cells in the brain influence behavior and cognition, we lack the tools to be able to analyze and quantify the trillions of interconnections between neurons, and we lack the computational horsepower to be able to run such a simulation even if we could build it.

Those are technical objections. The issue of modeling a person all the way down to the level of genetic variation in neurotransmitter and receptor function, however, is something else.

Assuming we overcome the limitations of the first round of problems, we're still left with the fact that there's a lot more going on in the brain than generic, interchangeable neurons behaving in predictable ways. To actually copy a person, we need to be able to account for genetic differences in the structure of receptors in the brain...

...and even if we do that, we still haven't accounted for the fact that organisms like Toxoplasma gondii can and do change the behavior of the brain to suit their own ends. (I would argue that a model of me that was faithful clear down to the molecular level probably wouldn't be a very good copy if it didn't include the effects that the parasite have had on my personality--effects that we still have no way to quantify.)

Sorry, Mr. Kurzweil, we're not there yet, and we're not likely to be any time soon. Modeling a specific person in a brain is orders of magnitude harder than you think it is. At this point, I can't even say with certainty that I think it will ever happen.


( 24 comments — Leave a comment )
Oct. 23rd, 2012 10:29 pm (UTC)
This is all very fascinating. Mind if I share on the various social networks?
Oct. 23rd, 2012 10:35 pm (UTC)
By all means, feel free!
Oct. 23rd, 2012 10:56 pm (UTC)
It also still doesn't fix the Star Trek (transporter) effect. If we copy a person, even if it is a perfect copy, what happens to the old person? We can't upload our consciousness. I don't think that'll ever be possible. Making a copy of it...maybe. But transference is a fantasy.

Oct. 23rd, 2012 11:32 pm (UTC)
The other thing is, what I call the alcohol problem. Nobody could figure out why people stopped hunting and gathering and started agriculture, because it involved more labor and effort and was less efficient. Until they finally realized that you needed agriculture to produce potables. It took nearly 50 years for the idea of people wanting alcohol that bad to sink into the consciousness of anthropology and be accepted. We founded civilization merely because we wanted to get drunk. Without those biological outlets and impulses...evolution stops, except as we self-direct it. Or on a more simple level...if you upload my brain, what happens to my sex drive and it's outlet. I guarantee after 50+ years of no sex, I'm going to be criminally insane.

It's a little like the idea of heaven...it makes a better ideal, and somewhere to visit, than somewhere any one would actually want to live.

Oct. 23rd, 2012 11:48 pm (UTC)
So what would happen to you if, tomorrow, something happened (paralysis, for instance) to render you unable to act on your sex drive?
Do you feel you wouldn't be able to go on without becoming insane?
Oct. 24th, 2012 12:57 am (UTC)
That's a good point and one I hadn't considered. But part of the problem was...more general than I made it sound. If you can't touch (e.g. "skin hunger" problems in male prison populations), if all input is...non-biological sensors...so much of who we are is tied up in our skin, which was sort of along the same lines of part of what Tacit was saying above. We have more than 5 senses, spatial awareness, temperature, etc and...though we rarely consciously acknowledge them, given complete deprivation from them...what is left and is it sane by human standards? The effects of total sensory deprivation are well documented, but...what about...if the only thing you could ever see was a TV. Or hear was a recorded voice. If you could never feel...the chair under you, or the wind on your skin or...that change in air pressure when a door opens behind you in the room, etc etc. And reproduction is a major major biological motivating force. How do two uploaded personalities make babies (Where do 'new' humans come from any more)? How do they burn the sin out so they can talk like rational adults?

We can't ethically upload personalities until we at least have a better grasp of what constrains and constructs a personality and...modern computer science thinks microsoft 'fences' and ergonomic keyboards are human-friendly interfaces. They're like Leo with his muscle power doing what we need a nuclear reactor to do.


Edited at 2012-10-24 01:01 am (UTC)
Oct. 26th, 2012 03:35 am (UTC)
Fair points. :-) And that varies from person to person, of course.

Then you have the ascetics' argument that it would increase & focus your intellectual powers... Imagine what Hawking could do if freed from his body, for instance.
Oct. 26th, 2012 07:11 pm (UTC)
*nods* I am not sure I agree with the ascetics argument, in this case, but it is another one I hadn't considered.:)

Oct. 24th, 2012 01:05 am (UTC)
The other thing is...biology will decrease my sex drive, over time. The flow of chemicals will stop/alter, and eventually I'll die. When we start talking functionally immortal timelines, though...and a brain stuck in X life-phase post-puberty, after 500 years? I'm pretty sure insanity would set in.

Oct. 24th, 2012 12:42 am (UTC)
If I can't dance/fuck, it's not my Rapture?
Oct. 24th, 2012 01:09 am (UTC)
I guess angels don't have primary or secondary gender/sexual organs, according to the Catholics. Lame party.:/

Oct. 23rd, 2012 11:40 pm (UTC)
Really interesting. Thank you for giving me something break the study up. If you dont mind would love to print it out for a discussion group.
Oct. 24th, 2012 01:29 am (UTC)
By all means! :)
Oct. 24th, 2012 12:49 am (UTC)
Well put indeed. You forgot one final bit, though, one that simply ends the delusion:

"When you die, you will be dead and will not return."

The dissonant denial of this simple fact seems to drive this silly pursuit of brain archiving. Count how many vitamins Mr. K. pops every day for evidence.
Oct. 24th, 2012 01:11 am (UTC)
Also, religion, and things like vampires and zombies and...a surprising number of humanity's other willful self-delusions.:)

Oct. 24th, 2012 01:29 am (UTC)
Fascinating ponderments. Thank you.
Oct. 25th, 2012 09:13 am (UTC)
1) on the toxoplasmosis:
As you said, to make an exact copy, one would have to get down to the molecular level to account for all the variable proteins, so that is a long way off, if indeed it could ever be done.
If it CAN be done though, then of course your parasite will be copied along with everything else. And it also wont do to just copy the brain- the rest of your body shapes your personality and behavior as well, especially your gut (http://www.newscientist.com/blogs/shortsharpscience/2011/08/-eating-a-strain-of.html, http://www.newscientist.com/article/mg21028136.500-do-early-gut-problems-set-the-brain-up-for-depression.html), and since your body consists of way more 'non-you' bits than 'you'(http://mpkb.org/home/pathogenesis/microbiota), all those little guys need to be mapped as well to give a functioning model that would behave just as you would.

2) the autism question:
a) it starts, as always, with consent. Has the person with Aspergers or Autism personally consented to their upload? If they did, they are also able to consent to having their receptors tweaked to stop expressing autism spectrum behavior. A backup can be stored in case they do not like the changes and wish to revert to their former state or they may choose to let two instances of themselves run, expressing each.
b) if the person has Autism that does not allow them to communicate consent, only the person with power of attorney may make that decision. Once uploaded, the model may be tweaked to the point where the uploaded person can in some way communicate their wishes, in which case a) applies, and the person can even choose to have their uploaded instance deleted.

I agree with you that it's highly unlikely to happen, but fun to ponder -I'm reading Stross's 'Accelerando' (http://en.wikipedia.org/wiki/Accelerando_%28novel%29) at the moment, so I had uploads on the brain anyway :)

Edited at 2012-10-25 09:16 am (UTC)
Oct. 26th, 2012 10:50 pm (UTC)
Do you know if there is a way to tell if someone has been affected by said parasite (Toxoplasma gondii)?

I too have always lived with cats, save for a few years where my brother's allergies were too bad, and these years felt very stressful. Life without cats doesn't seem quite complete to me, and I think I might be affected by that parasite, but I'm curious if there is a way to know if that's the case.
Oct. 27th, 2012 02:42 am (UTC)
Here's a nice article on T gondii

The takeaway I got from it is the effects on a person's personality are subtle; there's trends in behavior in the infected, but so much difference in people's personalities that many infected act "normal" and many uninfected have all the personality traits. So your love of cats isn't necessarily due to the parasite, even if you have it.

To answer your question, I'd assume there's a test for the parasite (or the research wouldn't be possible).
Oct. 27th, 2012 03:02 am (UTC)
Thanks for writing this. I just learned some new things about neurotransmitters.

While I'm skeptical of uploading too (the abovementioned transporter problem; the digitized copy of your brain is just that, a copy), I'm also skeptical of your argument here. Genetic differences in how neurotransmitters work between individuals requires modelling all the molecules in the brain? I don't follow.

What if genetic differences were just treated as global settings in a brain model, affecting how each neurotransmitter works (IE more sensitivity to vasopresin, reduced oxytocin production). Compared to complexity of modelling the network of neurons and glial cells, adding a few settings and sliders to that model's behavior would be trivial.

Of course, adding in the parasites gets more interesting, but my guess would be a large proportion of them are themselves simply tweaking how the neurotransmitters work. Though at this point, we've barely scratched the surface on how many parasites/symbiotes affect human behavior, so we're not going to know for sure how to account for them anytime soon.

Would be ironic though, if they were the lion's share of one's personality, and sometime in the future a simple blood test could look at one's genes and parasites and determine personality with high accuracy. Of course, that would make uploading a lot simpler...
Francisco Hertel Maiochi
Oct. 30th, 2012 06:41 pm (UTC)

I know you don't know me, but I read your blog since I found out about it in the PolyWeekly podcast, when I was researching for my graduation paper. Now I am a teacher and clinical psychologist.

So, my idea is that we'll reach this uploading technology when we come around to use the human brain as a computer, creating an good enough interface. then we can clone the persons brain, circulate our own blood through it, so we have the same bacteria, etc, and then somehow copy one brain's state to the other.
Is this a possible alternative?
Nov. 2nd, 2012 10:48 pm (UTC)
Martin J Potter
With havin so much written content do you ever run into any problems of plagorism or copyright infringement? My blog has a lot of exclusive content I’ve either authored myself or outsourced but it appears a lot of it is popping it up all over the web without my authorization. Do you know any methods to help reduce content from being stolen? I’d genuinely appreciate it.
Nov. 5th, 2012 08:34 pm (UTC)
I found your post very interesting. Adding the new knowledge (of not knowing much about!) epigenetics, people are constructed very differently, not just because of mutations/different genotypes, but because of environmental influence.

Knowing the DNA-sequence hardly says anything, it's the state of activation and inactivation that holds much of the key for genetic expression. And as a student, I feel like the knowledge is still rudimentary at best, especially for epigenetics and neuroscience.
Nov. 13th, 2012 07:13 pm (UTC)
Thought you might like to see this: http://www.slate.com/blogs/xx_factor/2012/07/03/toxoplasmosis_and_suicide_it_s_probably_not_the_cats_causing_danish_women_to_kill_themselves_.html

"A series of small studies that compared personality tests in carriers and noncarriers found that men with toxoplasmosis were more “expedient, suspicious, jealous, and dogmatic,”"

This doesn't sound like you at all. Makes me wonder if your extreme cat-loving behaviour really is a result of an infection, or if you're just an exception to this trend.
( 24 comments — Leave a comment )