Cool. Let's see the methods paper they used to produce this model, then let's see the independent validation papers which would show us that the model is accurate in compared to previously established information. Then let's see an independently authored paper on a unique speculative use of this model that has positive results.
This model is in its infancy. It still needs a lot of experimental work (5 years? 10 years?) and in general the more complex models take a longer amount of time to be validated and widely trusted. A great example is the humanized mouse model, which is super complex (though not as complex as a brain) and took somewhere between 3 and 10 years depending on who you ask to go from the point of the OP model to mainstream accepted usage for some purposes. Maybe 5 years down the line we'll see something come out of the new brain model.
Brain science isn't my expertise, but there's a couple of other tidbits that caught my attention about this model. First, its vascular structure is missing entirely. I'm guessing the level of oxygenation of the fetal brain is important, but there may be ways of controlling for this gap. Second, gene expression isn't 100% equivalent to an actual human fetal brain at this level of development. There are probably ways to control for this also, but it's still a flaw. We don't expect models to be 100% true to reality (after all, it's a model), so I expect that these holes aren't going to prevent this model from being successful and useful.
Brain science is my expertise, and you're right that the lack of vasculature will limit the viability of this. It will never grow beyond a few mm, and will have a dead core most likely. Plus, at "5 weeks" there will be no astrocytes or oligodendrocytes. You need astrocytes to make proper synapses and oligos to get proper signalling speeds. Others have shown more advanced brain organoids, incl astrocytes, but this tech is still very far from a conscious piece of tissue in a dish.
Increasing probability that you are in fact a brain in a vat, part of an undergrad science experiment in the outerworld.
I mean, more seriously I am pro-choice, but do we draw a line somewhere? "The main thing missing in this model is a vascular system. What is there – a spinal cord, all major regions of the brain, multiple cell types, signaling circuitry and even a retina – has the potential to dramatically accelerate the pace of neuroscience research"
Of a 5 week old fetus? Is it ok to make 10,000 of these for random experiments? What about 10 weeks, or 15 or 19 or 40 weeks maturity?
I'm not saying I believe this in and of itself is unethical, but it seems extremely dangerous to do this research and publish and promote it without even beginning a discussion of, where would we draw a line? Does anyone really believe the line will be here, and not an inch farther?
Meh, better we use these models than entire living mice or rats, which would be the alternative. I highly doubt there is any kind of sensation or consciousness inside of the brain models, seeing as how they have no sensory inputs or organic energy inputs (especially oxygen) whatsoever. On top of that, even 5 week old fetuses (which these model brains are definitively not) aren't remotely viable or sensate anyway, and don't even have an associated behavior model AFAIK.
The ethical boundaries for experimenting on living mice are, in my opinion, a bit perverse at the extremes. There's a general consensus that mice are sensate and can experience in some form basic emotions and learning behaviors. Despite this, it's acceptable to do experiments which deplete or wildly manipulate various neurotransmitters or neural circuits inside of the mouse's brain far beyond normal physiology.
Imagine having 100% of the "feel good" chemicals removed from your brain and replaced with 10,000% the normal concentration of "feel bad" chemicals. Obviously this is a huge simplification, but it's pretty clear that it would be hellish-- and that experiment has actually been done using mice. Sure, it's not as bad as if it were done with a real person, but it'd be much better if we could spare the mice and use a model brain in a vat instead.
I gave an answer on Reddit a week or two ago to a what might be considered the current limiting case in clinical practice. In another comment in the same thread I tried to point out that there are a lot of edge cases, even where it seems to be a bright line.
It seems reasonable to me that we should monitor this train of research closely, and I'm pretty confident people will.
We also need to caution ourselves at avoiding things because they are scary. At some point, deciding not to move on from a tended field was probably terrifying. Obviously it turned out all right for their descendants.
Meanwhile, I would argue this is much less scary than moving to a monodiet in terms of genetic variety in what we eat, having a 9-5 job during peak sunlight, surrounding ourselves with artificial light, and having no idea what any of my neighbors' names are. Frankly, I'm not entirely sure I would object to literally growing fetus brains so long as we understood what we were scared of. But not flipping on a neural network because it "might" be conscious, thus ensuring ethical questions, strikes me as.... premature moralizing. We already know this model doesn't describe anything that could actually form into a brain, just see other comments on the vascular structure. It's like saying "hey, we discovered radioactivity, so let's put a moratorium on it so nobody makes a bomb".
Having the comparable structure to the brain of a 5 week old fetus is not the same as using the brain of a 5 week old fetus. This seems to be the same sort of technology as the lab grown meat stunt a few months back, only applied to neuroscience. Still some ethical issues here, but I'm much more comfortable with this sort of thing as opposed to actual human brains.
Oh I agree - but I don't think that's likely. At 5 weeks a human fetus isn't viable in any way shape or form; this 'model' has even less of a case for, well, being considered alive seeing as it lacks a vascular system.
Reading the article again, this model seems less of a 'hey we have a brain in a jar lets poke it' and more of a platform for studying neurological diseases/ethically generating brain tissue structure, as opposed to cells. This seems more like raw material, rather than harvesting life.
I would not be comfortable with an alien of superior intelligence prodding my brain on the grounds that it had the functioning capacity of one of their 5 week old foetuses.
Neither me, but you and I have agency and free will. This model has neither of those things.
I think the crucial question is does this 'model' have the capacity to develop those things. If it does, that creates a serious problem for continuing research along these lines. If not, then I don't really see them problem - in that case I would compare this to growing and running test on kidney or liver.
Morality we know is encoded in human brains anyway. We may not like it, but that doesn't mean it's not moral from alien's POV. And yes, in such case, tough luck for us.
If there were actually a reasonable case for why fuqismatk or missildrigg were necessary rather than arbitrary criteria for the thing to be bad, they might have a point. If I found out an alien had been doing something to me that is imperceptible and harmless from my perspective because I lack fuqismatk or missildrigg, I wouldn't feel any ill will toward that alien.
There's a difference between prodding your brain or prodding a brain modeled after your brain, which doesn't have a way to sense things prodding it.
let me help you take your example further:
"These humans, they might have agency and free will, but they don't have fuqismatk and missildrigg, so us gakimiking and monikiking their brain should be ok."
Recommended reading: Permutation City by Greg Egan. The storyline meanders a bit but there are so many ideas packed into that book that I still think about.
Diffusion. Oxygen levels in tissue are about 3%, compared to 20% in the atmosphere. In fact, that was quite a finding back in ... the late 80s I think: they could coax fibroblasts to divide many more generations simply by incubating them in a low oxygen environment. Free radicals are brutal.
Right, but then one has to ask the question if this is pain that we (humanity as a whole) care about or not.
We cause pain and suffering in hundreds of thousands of lab animals a year, not to mention our livestock, etc. While acknowledging that pain, we've determined it doesn't matter (or more precisely, that an animals pain has less weight in our moral/ethical considerations than equivalent human pain - if such a thing could be said to exist).
We've determined that any pain a fetus may or may not experience in an abortion is less meaningful than any social or economic pain the mother might experience (not trying to start a pro-life/pro-choice debate, here, but it's kind of hard to talk about this without looking at how we as a society handle abortion). What makes this different from using aborted fetal tissue? What meaning should we attach to potential pain in an organism created specifically for scientific study (assuming this was a living organism, which it doesn't seem to be, but I could be misreading the article).
Sorry for the rant, I'm still working out all the questions this raises and some of that spilled into this comment.
The lab-grown brain...contains 99 percent of the genes present in the human fetal brain.
"We are thrilled to have finally created a living human brain in a vat. This is an exciting experiment and one that has absolutely no possible ethical ramifications of any kind whatsoever, now or wherever it may lead in the future," the not-evil scientist Anand added enthusiastically and without visibly fighting the urge to twirl his totally not-evil mustache.
"It's simply completely obvious that this is a great direction to go in and one that will never give anyone any pause. More biscuits?"
So yeah, it sucks that when you read a headline, you don't know if it's talking about an animal model, an in vitro model, a computational model, or something in between. But that's how it is. They're all models.
This model is in its infancy. It still needs a lot of experimental work (5 years? 10 years?) and in general the more complex models take a longer amount of time to be validated and widely trusted. A great example is the humanized mouse model, which is super complex (though not as complex as a brain) and took somewhere between 3 and 10 years depending on who you ask to go from the point of the OP model to mainstream accepted usage for some purposes. Maybe 5 years down the line we'll see something come out of the new brain model.
Brain science isn't my expertise, but there's a couple of other tidbits that caught my attention about this model. First, its vascular structure is missing entirely. I'm guessing the level of oxygenation of the fetal brain is important, but there may be ways of controlling for this gap. Second, gene expression isn't 100% equivalent to an actual human fetal brain at this level of development. There are probably ways to control for this also, but it's still a flaw. We don't expect models to be 100% true to reality (after all, it's a model), so I expect that these holes aren't going to prevent this model from being successful and useful.