Obama’s BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Initiative, announced today, concerns what is arguably the most important scientific project we humans are doing today: figuring out how higher biological intelligence works, by exploring and mapping it in living and preserved brains all relevant resolutions. Neuroscientists have developed powerful new mapping tools and software in two main categories. Functional connectomics (also called Brain Activity Maps) is the process of mapping synaptic connectivity and neural activity to biological function, including memory, in living brains. To make these maps we have new tools for monitoring neural action in vivo at molecular, cellular and circuit levels, like optogenetics, calcium imaging, nanoparticle sensors, and other clever advances. Structural connectomics (also called just Connectomics Maps) is the process of mapping synaptic, cellular, and nuclear (epigenetic) information in chemically preserved, nonliving brains (worms, flies, snails, zebrafish, mice, monkeys, humans, etc.), as a path to figuring out function. There are also new tools and software for the automated slicing, scanning, and mapping of synaptic connections. It was observation of the rapid advances in these areas that led led Ken Hayworth and I to co-found the Brain Preservation Foundation in 2010.
The new idea is that combining these two forms of brain mapping may finally allow us to uncover the neural coding system, the ways networks of neurons store short and long term information in their association patterns and strengths. The paper that launched the Brain Activity Map proposal is The Brain Activity Map Project and the Challenge of Functional Connectomics, Alvisatros et.al., Neuron 74, June 21, 2012 (5 pp). It’s a great intro to the exciting promise of this field, and a call to action. Wikipedia has no page yet on functional connectomics (perhaps a neuroscientist will start one) but they do have a page now on the BRAIN Initiative.
There are many potential benefits to functional and structural connectomics for science and medicine, but their greatest promise, in my opinion, is that they will accelerate our ability to build intelligence in our much faster and eventually far more capable electronic systems. Some of the brain’s circuit structure and function will turn out to be highly similar from brain to brain (developmental) and some will be unpredictably different (sometimes called “evolutionary” or “Darwinian” differences). Understanding the developmental parts of the brain, and how they constrain and enable the evolutionary parts, will get us much farther down the road of building self-improving artificial intelligences. Activity and connectomics maps, and a few other new tools for monitoring neural activity at molecular scale will of course provide many medical and neuroscientific benefits, and these can be sold most easily to the general public, but the intelligence benefit for science and society, via advances in computational neuroscience and machine learning may quickly become the most important for us.
Obama hinted in his State of the Union address in February that he wanted to see America’s brain-mapping and related neuroscience efforts “reach level of research and development not seen since the height of the Space Race.” Science writer John Markoff, in a great NYT article Feb 17th, summarized the views of the founding scientists behind the Brain Activity Map proposal, that funding on the order of $3B, or $300M/year, should be publicly committed to this project. That would make it less than the $3.8B we spent on the Human Genome Project from 1998-2003, an investment which returned, according to a 2011 Battelle report, $796B in new economic activity between 1998 and 2010. A return on investment of greater than 200, one of those rare ROIs you see when opening up an entirely new field.
Functional and Structural Connectomics promises to have that same kind of fundamental impact, opening up neuroscience and bringing all the benefits of understanding natural intelligent systems to the technology world. In addition, understanding how the brain uses connectionist features like redundancy and neuroplasticity to protect its critical functions would be huge advances for medical science and therapy. I recommend reading Norman Doidge, in The Brain that Changes Itself, 2007, for fantastic and motivating examples of how resilient our brains can be to memory loss and damage.
Unfortunately, in his announcement today President Obama has committed just $100M to the project for its first year budget. And the money committed so far is a hodge-podge that is not project or map focused. Consider that Europe’s Human Brain Project just got $1.3B committed from the EU for the next ten years, even though that project is doing far more theoretical, lower-resolution simulation work that will be highly likely to have a much poorer payoff, in a world where we haven’t yet cracked the static and dynamic neural coding algorithms. Yes, the Human Genome Project started with the same small seed funding of around $100M the first year. But that was when genomics was untested, proteomics a dream, and understanding and mapping the brain still largely unreachable. We’re way beyond those early days now. We know how important maps are, and that we have tools available to make them, and the data sciences folks and hardware to analyze all the new public domain data that will result. It’s time to match real funds with the rhetoric.
As I said, the scientists involved in the BRAIN initiative know we’ll need at least $3B to make major discoveries with activity maps alone, and this doesn’t even include connectomics maps, which deserve a few billion as well, if we really want to figure out the neural coding language in any complex animal (say, a fly, or perhaps an Etruscan shrew, a mammal with only 1 million neurons). $5B is not a lot of money for the incredible intellectual advances we can expect. To put this in perspective, we are presently spending $85 billion per month on QE3. Obama cobbled this $100M together by redirecting existing funds in NIH, DARPA, and NSF budgets, so it isn’t even new money, it’s just reclassified R&D. An NIH working group has been designated to develop a multi-year plan with cost estimates by June 2014, and Obama has fast tracked the group by asking for an interim report by fall 2013. But its still quite unclear what the goals of the project are, and whether connectomics maps will even play a role. If they pass on funding synapse-level connectomics maps, that will be a major failure of nerve.
Isn’t $100M a great start for Year 1? Not in my book. What would have been commendable, for a project with this magnitude of potential benefit, would have been starting with a level of finding that is ten times more, or at least a billion dollars up front, and a commitment to seek at least a billion a year for the next ten years. That’s enough to influence students to enter into this field, and would place this project in the light it deserves – one of the best science projects we could work on at this unique point in human history. We should and can demand a lot more from this second term president, particularly one who understands science and tech the way he does. Obama has committed to a commission to study the bioethical issues that might emerge (a concession to conservatives perhaps), but so far his “dream team” of 15 neuroscientists have not committed to connectomics maps, as far as I’ve read. Perhaps they will, but given the vagueness of today’s announcement, it’s quite possible we we’ll see something better in the future. But this isn’t the kind of start that inspires confidence.
Ultimately, as readers of this blog know, whether second-term American politicians have the courage to say it publicly or not yet, smarter machines, even more than adding more 20th century-style jobs, have become the primary wealth creator in the developed world, so that’s where our thoughts should go first, as we look for ways to improve our lot. I think it’s time we got serious as a species about realizing what kind of progress the universe has engaged us in. We are here to use our wits and works to become something greater than ourselves. Our highest role appears to be to take what the universe has done with us and make something even smarter, more ethical, more productive, and more resilient as our progeny. This is what civilization has been about, since the birth of technology, as I see it.
Want to let the Obama administration know your thoughts on making Brain Mapping, including connectomics maps, a top funding priority? You can send a brief email to the White House by using this form, as I have. Thanks.
A “clarified” mouse brain at right (lipids removed, all else stays in place). Transparent to optical microscopy, all proteins, receptors, RNAs can be repetitively interrogated with molecular probes. Amazing!
4/18/2013 Update: The Stanford press release on 4/10 announcing CLARITY, the Karl Deisseroth lab’s amazing new method for optically transparent brain mapping, just makes what I said above more appropriate and urgent, from my perspective. Deisseroth is one of the 15 experts on Obama’s neuroscience dream team, so I’m sure he advised the White House of its implications. The CLARITY paper was accepted for review at Nature in September 2012. The CLARITY method is like PCR, a multipurpose, revolutionary new research tool that will open up vast new imaging and molecular phenotyping research capabilities in any biological tissue, and in particular the brain. Salk’s Terry Sejnowski said: “It’s exactly the technique everyone’s been waiting for.” He told the Associate Press that it will speed up brain anatomy research by “10 to 100 times.”
And yet Obama’s team still proposed just $100M in funding for brain mapping for the first year. That’s simply ridiculous. Please, America, wake up! It’s time to spend some real money on neuroscience and bust humanity out of its ignorance. Stop being scared of how much better things will soon be, once we’ve cracked the riddles of neural information processing. Someone also needs to give Deisseroth a serious prize or two. Optogenetics and CLARITY, both out of his lab, are each profoundly important biological sciences breakthroughs.
10/27/2013 Update: Ugh! Obama’s BRAIN initiative (April 2013) has barely started and it’s already been co-opted. Politics is not pretty. http://www.nytimes.com/2013/10/25/science/pentagon-agency-to-spend-70-million-on-brain-research.html?_r=0
DARPA will spend their chunk of the funds (half of the money we’ve committed so far to the initiative, $70M over five years, which is peanuts, as I’ve said before) on a very-low-yield clinical project (deep brain monitoring and stimulation) vs a multipronged effort to improve human brain structural and functional connectomics (circuit tracing, electrical activity mapping, optogenetics, nanosensors). The potential for brain mapping as the #1 focus of the initiative is gone, mere months after they announced it.
Apparently Obama got the wrong partners (DARPA, NSF, NIH) together for his BRAIN initiative. The 2012 Brain Activity Map proposal that Alvisatros, Church and others made to the White House was all about functional connectomics. This has now taken a back seat to deep brain stimulation and monitoring experiments. Drats! I like DARPA, but I’m sure that initiative mostly won’t work, without functional maps, and I’m not even a neuroscientist. But DARPA likes clinical work with near-term potential benefit (or at least the potential promise of it). It would have taken a firm hand to keep them focused on Brain Mapping, which is the real prize accessible to science, at this stage of our collective technical abilities (more accurately, ineptitude) when it comes to the brain. That leadership is missing today.
We need a lot more money, at least a billion dollars a year, devoted to funding the Basic Science of structural and functional Brain Maps, not these expensive clinical junkets. How else can we solve the memory code, and thereby understand how neural nets actually work, and thus make better AIs? Or, as my friend Steve Coles, MD, PhD says, what is the genetic reason why we humans have a Broca’s area and chimps don’t? What is the connectomics of higher intelligence? All most future-important questions about human, social, and machine intelligence are dependent on better brain maps. Ken Hayworth and I started the Brain Preservation Foundation (http://www.brainpreservation.org) in 2010 with the realization that these maps are coming, and will greatly improve our understanding of who we are, and what we can do with our memories and identities after biological death.
One fine day we can expect a real Human Brain Mapping initiative, one that really does greatly improve our collective understanding of the brain, for all humanity, for all time. Just like the Human Genome Project uncovered the epigenome and illuminated the proteome, and now we need Human Epigenome and Proteome Projects, which also haven’t materialized, because we are so broke and unmotivated to do Big Life Science.
The world needs Brain Maps, Epigenome Maps and Proteome Maps as the new science moonshots for the next five to ten years. These would be completed under budget and under time with more powerful computers than we expected, just like the HGP was.
In the meantime, we get this BRAIN initiative elephant, designed by committee. This is a major loss of vision and leadership here. #ObamaFail
Vote for scientific and technical leadership in 2016, irrespective of party. It’s high time we get some representatives who see, sooner or later, how extraordinary humanity’s future will be. Sooner would be nice, eh?
Ever Smarter World 
I’m happy to see this kind of research going ahead, especially if it has benefits for conditions like Alzheimer’s. One thing that concerns me is issues regarding intelligence in general and the related question of the nature of consciousness. The people doing this kind of research tend to have a very narrow understanding of both intelligence and consciousness, depicting them in narrow terms, constrained by paradigmatic and civilisational presuppositions. Intelligence as defined in modern terms tends to focus on utilitarian functions like memory capacity, processing speed, capacity for abstraction; and verbal-linguistic, visual-spatial and mathematical-logical acuity. These are all the kinds of cognitive processes which are highly valued in modern society and education. But will the people funding and doing this research be interested in enhancing other cognitive skills and aptitudes which are not valued and poorly understood in the modern university? It’s effectively a rhetorical question. What about the capacity to bring the mind into stillness? That actually means stopping thinking, in the narrowly-defined modern sense of the word “think”. And stopping thinking is anathema to the modern mind. I’m not going to go into the long discussion of what value there is in mindfulness. I’ll leave it for others to contemplate: if they are not too busy multi-tasking. 🙂
Personally, I think you will see a split developing in society between people who want to enhance intelligence in the modern sense of the word, and those who will choose another way: to enhance spiritual intelligence as a first priority. The latter intelligence can certainly be mediated via technology, but it isn’t necessary. In my experience, the kind of people who value it will also tend to be better at identifying some of the problems caused by an imbalanced development of the so-called “left-brained” cognitive abilities.
All these cognitive potentials are perfectly compatible. It’s just a question of balance, and at present modern education, academia and science are creating very imbalanced minds, IMHO. I would like to think those doing the research into the brain will have an appreciation for this. I suspect it won’t be the case, however, and the current imbalance will be exacerbated as a result of the application of the research, i.e. the technology that emerges.
Hi Marcus, Thank you for the thoughtful post. I know from past conversations that we look at these issues from different frames but I always enjoy and learn from your perspective too, and find much to agree with. I think most of the folks doing neuroscience work have narrow understandings of intelligence and consciousness because they are in the trenches, using reductionist methods that have worked very well to get us where we are today, so I don’t fault them for not having wider visions. We also need the synthesists and social scientists and philosophers and other political and social interpreters of all this work too of course, and we have more of those folks today than ever before.
What is exciting to me is to find that we are now getting reductionist models of things like consciousness and morality to go with our synthetic models. For years scientists and philosophers said consciousness would be an ultimate mystery because of its subjective nature, and that science would never have anything to say about morality, and that topic would lie always with religion. Now we are beginning to describe moral universals across all cultures, are learning to understand consciousness at the level of neural synchrony, and can even give it back to people in persistent vegetative states. http://www.guardian.co.uk/science/2006/sep/12/health.healthandwellbeing
In the future, I can see us coming up with tools for improving the daily quality of our consciousness. I think there’s lots of evidence in neuroscience and medicine now for the value of mindfulness and single-tasking, it just isn’t widely known yet. But it will be, as the web gets smarter and does more of our thinking for us. I do agree that society swings between splits and convergences as technical capacity grows, and can see your point that there will be more subcultures over time, including many who choose more simple ways, but in general, I think integration will only continue to accelerate, and technology will become more organic and “invisible” to use Mark Wieser’s term – permeating our environment and acting the way we’d like a natural biological system to act, yet at a new level of intelligence.
So while you suggest individual spiritual states may be the surest path to increased collective wisdom, I see increased integration with and intelligence in our electronic extensions as the surest path to greater levels of morality and responsibility for humans as a species. I’m sure both paths will continue to go forward, but I don’t see the big split or social tension emerging that some futurists think may result from this. I expect more slow cities, simplicity, mindfulness, and other movements (and I’m big on simplicity and mindfulness myself) yet we’ll also see “invisible” and “smart” technology integration movements that will appeal to both groups. We shall see, as they say.
Great analysis John! … What about a scenario in which advanced brain mapping research is being done behind closed doors with the blessing / urging of intelligence, defense and financial security elements? This *could* be one of those tech growth areas that some would like to slow, especially considering the impact the knowledge could have on software, search, A.I., etc. I can see a valid global security argument.
That’s an interesting speculation. I don’t see how a clandestine mapping project could be kept secret for very long however, given the number of people that would have to be involved. There’s still so much to be automated. And the blowback if uncovered would be huge. Unfortunately I think using mapping to improve our software is something that is going to take a decade or two. One of the big wins will be uncovering the neural code, the way networks of neurons store information, short and long term. That could come in just a few years or twenty, in my intuition, depending on how serious we are about funding all this research.