This is a huge donation for the Broad and probably equals all their previous fundraising. I think a lot will ask if it's too much to one institute or too little to solve the problem, but overall I'm very excited. Biology is changing very fast right now, and it's probably the right time for a big new push. Psychiatric disease and brain research is the next big frontier and private funding like this allows places to take risks they otherwise couldn't on grants. The Broad Institute is the best genomics institute by far, albeit a little light on the clinical side. The money will not be wasted on anything hokey. All that said though, there is a large chance that it still won't be enough to solve a very tough set of problems.
I really appreciate that the Broad keeps a lot of attention on basic (or at least, pre-clinical) research. That kind of work is critical to making big discoveries but it's harder to defend for funding in a world where we want to put drugs on shelves. I still want to see drugs on shelves, but starting at the clinic isn't the best way to make discoveries that lead to real therapeutics.
I'm happy that they're putting money into developing better diagnostic methods for mental disorders, because, frankly, the existing tools are terrible.
After 9 years of misdiagnoses by several different doctors, earlier this year one astute doctor was able to uncover that I had bipolar 2 and not unipolar depression or SAD. I had mixed feelings about it (ironic) but I was put on the correct medications and they have made a world of difference. In fact, I read that it can often take close to a decade, on average, to diagnose types of bipolar properly, because doctors and patients simply don't communicate well. Any tools that shorten that time gap would benefit individuals and society as a whole.
None of your examples entail building a mind. Also there's a kind of irony in my example, since as a branch of computer science it is independently funded and seemingly unrelated to mental health.
You have a model of the brain. You do not have a model of the mind. You assume that by simulating a brain with significant details, a simulated mind will emerge. I find that's a big pill to swallow.
Nah, I don't assume it. Hence 'may' and 'attempts'. But the explanation is simple enough. If you can discover what it takes to code a mind then you're also going to learn some of the systemic faults that minds, qua minds, can develop. Of course, I could be wrong about this. It could be that mental diseases are all purely down to hardware issues (brain health). But it seems like a bad bet. For example: addiction, or at least specific types of addiction, depend on people's culture and choices.
Sure, it isn't ethical to experiment w/o consent. But in order to program an AGI, you need first to conjecture an explanation of how the mind works. One may be able to deduce from that explanation some of the ways in which minds can go wrong. Also, assuming consent is given, it will likely be easier (and less physically invasive) to observe the internal state of an artificial person.
I was initially worried when the headline said "psychiatric" research, but was pleasantly surprised when I read the article and found out that what they really meant was neuro-biology and its genetic origins, not study of patient behavior. The "top-down" approach has failed psychiatry (or more accurately, psychology) for half a century. I'm glad to see the bottom-up approach (i.e. starting with genes and cells and then working your way up to thoughts/feelings/behaviors) taking the lead.
It's at least worth noting that it's only within the last decade that such an approach has become viable -- we're still entirely unable to establish any relationship between genetics and psychiatric illness via theory, so massive, inexpensive genetic analysis was a prerequisite to this sort of study (which still relies on correct diagnosis in order to find genetic correlates).
It is a wonderful donation, but I do wish that philanthropists would spend money on lobbying congress and The President to increase science budgets by the billions of dollars per year. Money could also be donated to those running in congress who are favorable to increasing the science budget of NIH, NSF.
In fact, I would say your model has it backwards -- as a philanthropist, rather than devote my money to a cause, I should coerce/influence politicians with money, so as to compel through force of law other peoples money to flow to that cause.
Research subsidies capture the positive externalities to society that result from conducting that research. In other words: because much or all of the value of discovering--for instance--a new, accurate model of mental illness is leaky and impossible to capture, it's underprovisioned by the market. Government funding correct the inefficiency that would otherwise result.
That's the obvious reasoning, and I'm sure you knew this when you asked the question. Whether a specific topic has those particular externalities or not is always up for debate.
One could argue that increased government funding for science is more likely to fund basic science and other "non-sexy" but still important and likely to pay dividends research.
In theory (and hopefully) but all of the non-sexy stuff ends up as political wedge issues. And when that happens we have people like Palin decrying research on fruit flies in Paris, France.
This science research is very, very expensive costing many billions of $$$ on an ongoing basis. Regular charity is a "drop in the bucket" for this kind of research which should be paid for by governments anyway. For health, total spending was $3.8 trillion this past year. Spending more money on health research could have significant ROI in terms of reducing the growth rate of that bill.
And for those who didn't see the live presentation, they all made the point that the biggest story here was that they hoped this announcement would galvanize the community in order to drum up more attention, excitement and thus money.
As an example, many of the big pharma companies have already or currently are pulling out of the psych field. If this huge investment jump starts more researchers to go into this field instead of something else (e.g. cancer) it could create a positive feedback loop.
I don't prefer that model from a moral point of view, but it is worthwhile from a practical point of view. Lobbying the government has a higher ROI than research for most companies. As such many companies lobby heavily for loopholes, benefits, etc all in explicit name of jobs. (And private benefit for execs and owners)
Why concede that fight? If very long term basic science is of communal benefit and can't happen in the current market structure, why not use 10% of the money to fight against lobbyists asking for corporate handouts?
Because in so doing, you're just getting into a shouting match saying "My handouts aren't really handouts. They're important." Which is precisely what everyone who is looking for government money does.
Rather than taking part in the K Street circus of perverse incentives, why not just spend the money on research?
This article is a enlightening counterpart to the 'emerging neuroscience' chorus, who would have us believe that a full model of consciousness and the brain is just around the corner. That those trying treat schizophrenia haven't a clue what causes the onset, and why certain drugs are effective tells me that we really are at the stage of the absolutely primitive models of brain activity.
Anyone who says a full model of the brain is right around the corner is either a fool or a charlatan. The BRAIN initiative was a first start to _develop the technologies_ needed to take the measurements that would underly potentially explanatory brain models.
In Conneconomics (http://dx.doi.org/10.1101/001214) we try to estimate the cost and time needed to map the connectome alone (only one component of a holistic brain model).
I'm glad to see more money spent on these issues, hopefully we will have some breakthroughs soon.
While the current medications surely do help a lot of people, I've watched really smart friends devolve into zombies while on some of them. People I've known for 15 years who are definately not the same people they were when they started. They themselves don't notice the change, of course, because it happens slowly.
We really need better options with less severe personality change side-effects.
I applaud any efforts being made in the mental illness space. It's difficult because it can't be "seen" and anything we can do to make diagnosis as objective as possible would be great.
Behavior genetics research on schizophrenia has been going on since I was born, in other words since long before most Hacker News participants were born. The crucial issue in the research, mentioned in the article kindly submitted here, is that any one gene only has a small effect on risk for schizophrenia. "Dr. Lander cautioned that each variant accounts for only a tiny portion of the risk of developing schizophrenia. 'It shouldn’t be used for a risk predictor,' he said." Behavior geneticists, such as those I know locally,[1] and the pioneer of behavior genetic research on schizophrenia[2] hope that large-scale genomic studies of the kind that are now possible will help tease out how gene differences between one individual and another interact with environmental influences to increase or decrease risk for schizophrenia, which should lead eventually to better treatments and better prevention strategies, through deeper fundamental understanding of the disease. It is plain that environmental factors matter too, because sometimes identical ("monozygotic") twins are discordant for schizophrenia.[3] This research path will be slow and frustrating, precisely because it is well known already that no one gene has strong effect, and no particular gene acts without interaction with the environment, but this kind of fundamental research is indeed necessary to refine the diathesis-stress model[4] of the development of schizophrenia.
AFTER EDIT: I'm still learning the rules about informed, thoughtful discussion here after 2073 days of participation as a registered user on Hacker News, so I'd be glad to understand what someone considered objectionable about this comment. I was just surfing by this comment to add a link emailed to me by Irving Gottesman, the prominent schizophrenia researcher who was mentioned in the acknowledgements of the book A Beautiful Mind. That link discusses the study in the journal Nature published today.
This is a great donation, that seems to be going to exactly the right place. However, it's a little unfortunate that Mr. Stanley only started donating to this cause after a member of his own family was affected.
Agreed. And in this case, I think it's significant that the family member affected has controlled his illness long ago. The point of this donation is to help other people for whom such control is not yet possible.
