We're entering the era of big neuroscience. In just the past year, multi-million dollar brain research initiatives have been launched in the United States, Europe and Japan. Together, they will involve more than a thousand neuroscientists and will span more than a decade.
Their ambitions are bold: These marquee projects aim to map the brain in unprecedented detail, creating a picture of its different cell types, their activity and their interconnections. But their priorities differ: The U.S. Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative is aimed initially at developing the suite of tools needed to make such a map; Europe’s Human Brain Project (HBP) is creating computational models of the brain’s circuitry; and Japan’s Brain Mapping by Integrated Neurotechnologies for Disease Studies (Brain/MINDS) is largely focused on defining the circuitry of the non-human primate brain.
What has propelled neuroscience to the vanguard? What impact will these initiatives have on the field? How are the goals of these projects aligned? Will they work together? And how will they change our understanding of the brain?
The Kavli Foundation brought together leaders from these three initiatives—Bill Newsome, Sean Hill and Hideyuki Okano—to answer these questions.
Our participants are:
- WILLIAM T. NEWSOME - is the Harman Family Provostial Professor and Professor of Neurobiology at the Stanford University School of Medicine and Director, Stanford Neurosciences Institute. Dr. Newsome co-chaired the National Institutes of Health (NIH) working group that developed the long-term scientific vision for the BRAIN Initiative that was released last June.
- HIDEYUKI OKANO - is a professor of physiology and the dean of the Graduate School of Medicine at Keio University and a member of the RIKEN Brain Science Institute. Dr. Okano is one of two project leaders for Japan’s new national brain-mapping project, Brain/MINDS.
- SEAN HILL - is Titular Professor at the Brain Mind Institute at the École Polytechnique Fédérale de Lausanne (EPFL), co-director of the Blue Brain Project and co-director of neuroinformatics in the Human Brain Project (HBP). Dr. Hill also serves as the scientific director of the International Neuroinformatics Coordinating Facility (INCF) at the Karolinska Institutet in Stockholm, Sweden.
The following is an edited transcript of their roundtable discussion. The participants have been provided the opportunity to amend or edit their remarks.THE KAVLI FOUNDATION: What has propelled the brain to the forefront in terms of government science funding in the U.S., Europe and elsewhere?
BILL NEWSOME: There is a revolution occurring in experimental neuroscience. There’s the potential for incredibly rapid progress because of new tools that have been invented in the last five to 10 years that are enabling neuroscientists to make measurements of the nervous system that were simply unimaginable 10 years ago. We could think of them in science fiction but we couldn’t think about them in science reality. Yet now these things are becoming reality because of the new technologies.
SEAN HILL: This is the big-data era. The BRAIN Initiative is really about developing new technologies to acquire large-scale data because we have new ways to manage those data, analyze them and build models from them.
And as Bill said, we also have new neuroscience technologies that make this is a pretty unique moment not only to observe the brain with clarity but also to manipulate it. All of that combined makes it a really opportune moment.
HIDEYUKI OKANO: In Japan, we are focusing on mapping the nonhuman primate brain. It’s a useful model to fill the gap between the rodent and human brains, which are so different. Japanese groups have pioneered the genetic engineering of the common marmoset, a New World monkey, which puts us in a unique position to contribute to the international brain-mapping movement. We believe research on nonhuman primates is essential for understand the human brain and for developing knowledge-based strategy for diagnosis and treatment of psychiatric and neurological disorders.
TKF: Professor Okano, how long has it taken to lay the foundation for Brain/MINDS with basic research into the marmoset?
OKANO: Development of the transgenic technique started in 2006. We published a Nature paper about it in 2009. Since then we have tried to generate various types of neurological or psychiatric disease models such as a transgenic marmoset model of Alzheimer’s disease. Recently, due to the development of genome editing methods, we’ve been able to make knockout marmosets. That is a marmoset in which a gene has been deliberately disrupted. A knock-in marmoset, in which we insert a gene into the genome, is still in progress. Both of these techniques help us to explore the function of genes in the nervous system, but until recently it hasn’t been possible to use them in nonhuman primates.
TKF: The BRAIN Initiative has been called “America’s next moonshot.” Why is it important for governments to envision and engage in these of signature science projects?
HILL: It’s a way to attract attention, interest, funding and define a strategic focus. We’ve seen this in other domains. If you look at big physics projects such as CERN’s Large Hadron Collider or astronomy projects such as the Square Kilometer Array, they’ve gotten organized, they’ve gotten large-scale funding and they’ve been able to sustain that over many years.
NEWSOME: I don’t view the BRAIN Initiative as a big science project on the scale of a national laboratory, for example. It’s trying to promote a level of cooperation that will bring an interdisciplinary perspective to fundamental problems in brain science by funding smallish groups of investigators. The NIH-funded groups, for example, are on the order of two to six individual faculty members, bringing toolmakers together with tool-users, and theorists together with experimentalists.
Within the BRAIN Initiative working group at the NIH, we debated the value of big science projects. We believed that we might come to a point five years from now where we need larger, regional laboratories where scientists come to use equipment that simply can’t be afforded by individual universities. But we don’t believe we’re there yet. There’s still quite a bit of ferment with these new technologies and we need many smaller groups doing experiments and sort of having a little Darwinian party where ultimately the strongest approaches win but where lots of approaches are being tried first.
TKF: But there is an overarching national vision and a dedicated budget for the BRAIN Initiative...
NEWSOME: That’s true. But I would distinguish the national vision from big science.
Professor Okano, would you classify the Japan effort on the marmoset as big science in the sense of what the Allen Institute for Brain Science in Seattle is doing, where they’re assembling a team of more than 100 scientists to work on one project? Or would you say Brain/MINDS is more decentralized like the BRAIN Initiative is, in the sense that priorities will be stated and individual groups of investigators will have to apply for grants.
OKANO: Both ways. It’s centralized at the RIKEN Brain Science Institute, where there are strong teams working on marmoset brain mapping and behavioral analyses. But other institutes and universities across Japan are also participating in Brain/MINDS.
TKF: The BRAIN Initiative, HBP and Brain/MINDS are based on different models. Why was each of these models chosen and what do you think is gained or lost because of it?
OKANO: At Brain/MINDS, we felt we needed strategic collaboration between clinicians and basic researchers. For example, using a transgenic marmoset model of Alzheimer’s disease, clinicians and basic researchers are working together to identify the changes in the brain's circuitry during mild cognitive impairment and very early stages of Alzheimer's. Such collaborations aim to develop preemptive treatments for this disease before the onset of cognitive impairment. That's an example of the centralization in Brain/MINDS.
But we also think bottom-up development is important especially for new technologies for brain mapping. That’s why we also adapted this decentralization mechanism. Our strategy is to use centralization and decentralization in parallel.
NEWSOME: In the U.S., our planning committee was unanimous that we did not want to dictate solutions to problems. Instead, we wanted to identify the most important problems that were really blocking progress and let individual labs and teams of labs come up with the creative solutions to them. That’s why we’ve taken this more decentralized model.
The effort in Europe is quite different, with a much more centralized plan. That approach has virtues in that you can make things happen faster in principle. It also has some risks, which we’ve seen play out in Europe over the last year. Sean can probably comment on it.HILL: With the HBP, we saw an opportunity to bring funding from a European technology program to neuroscience. The Future Emerging Technologies Flagship program was seeking large-scale, innovative and risky projects to drive technological innovation for the benefit of society. Many scientific disciplines are focused on applying big-data techniques and high-performance computing to accelerate their respective domains. Not everyone agrees that this is the right time to take such an approach in neuroscience—but we do, especially given the tremendous amount of neuroscience data being generated from new techniques and the brain initiatives worldwide.
The key difference of the HBP is that it’s really focused on data integration rather than data generation. We want to be able to provide the platforms where data that have been generated can converge, where they can all be put into an integrated view of the brain.
TKF: That sounds like an ideal foil for the BRAIN Initiative.
HILL: Yes. One of the most common things I hear in meetings with key people from the BRAIN Initiative is, “You couldn’t have designed two more complementary projects.” The effort in the U.S. to develop new technologies and acquire multiple levels of brain data, and then our effort to integrate it. It wasn’t planned that way but it really worked out beautifully.
TKF: What was the biggest challenge to getting the BRAIN Initiative, HBP and Brain/MINDS off the ground?
NEWSOME: I think the biggest challenge is still in front of us and that is getting funding for the science. The U.S. BRAIN Initiative started very modestly in fiscal year 2014 with a $100-million budget. The working group recommended a linear increase in funding up to about $500 million dollars a year by 2019 and then a plateau at $500 million per year through 2025. But it remains to be seen whether the funding will be allocated by Congress.
The U.S. has a really good plan. We have galvanized a lot of people into action. We have the attention of the young people—graduate students and undergraduates who will be tomorrow’s primary investigators. The question is, are we going to follow through and fund this?
HILL: The greatest challenge is establishing a strong culture of team science, collaboration and data sharing in neuroscience. So far, we have seen tremendous excitement, energy and engagement among the teams working within HBP. We spent many years and went through numerous reviews of the project prior to launch that helped form a very solid 10-year plan. It’s not all smooth sailing and some scientists would rather have the funding redistributed to fund individual research projects. However, we think that in the long run a focused effort on data integration will help demonstrate the value of collaboration in neuroscience.
OKANO: Grants of course. Brain/MINDS is supposed to receive $30 million a year for 10 years. Very fortunately, a new biomedical funding system called AMED (Japan Medical Research and Development Agency) is coming in the next fiscal year, 2015. That’s quite a different system for Japan: it’s an NIH-like organization. I would like to see AMED increase the budget of our brain-mapping project but we are discussing about it.
There are also the challenges that are coming. We have to think about how can we share the big data among the brain-mapping projects in the U.S., Europe and Japan. This is one of the big challenges we face.
TKF: Right. It’s going to be important to be able to share the data that result from these projects. Professor Hill, I know you’re involved with Neurodata Without Borders, an effort to develop a common format for neurophysiology data. But as co-director of neuroinformatics at HBP, are you talking concretely about how to share data between these brain projects?
HILL: Absolutely. I’m deeply involved in those discussions. For example, we’ve been talking a lot with the NIH about how to make sure that what’s being developed within the BRAIN Initiative will actually be able to complement and integrate within HBP. The challenge we’re still facing is getting the neuroscience community and the funders to ensure that data is shared and reusable.
We’re finding other groups like Dr. Okano’s team in Japan, as well as in China and Australia, that are really eager to work on making the data organized, available and integrated. So, these conversations are happening and we are developing concrete agreements. We need to do more but the willingness is there.
TKF: Can we expect that these projects will make neuroscience research more interdisciplinary? And what other consequences might they have?
HILL: Neuroscience research is already an immensely multidisciplinary field. What we need is to find the key links and bridging principles between the various levels of organization of the brain. I hope that what they do is promote sharing, collaboration and reusability of data to turn neuroscience into a cooperative endeavor. I hope we see the value in combining data from different groups to find the relationships between work done in different subdomains of neuroscience, such as genetics, physiology, systems neuroscience and cognition and behavior, and different levels of organization in the brain.
It’s going to take a world to understand the brain, so what I really hope we come out with is a new spirit of working together as a global community.
NEWSOME: We need a long-term culture change to make interdisciplinary research a steady part of neuroscience. I totally believe that as President Obama has said, it’s an all hands on deck effort: just doing neuroscience in the traditional way will not allow us to move forward as quickly as if we bring new talent and new disciplinary perspectives to the research.
OKANO: I agree with that point.
Another consequence is that the technical innovation involved in these brain-mapping projects will greatly contribute to neuroscience in general. For example, a new technique that hasn’t been mentioned is the ability to make the brain translucent. Tissue-clearing protocols are causing a revolution in human brain pathology. They allow you to see the architecture of amyloid-beta deposits, a hallmark of Alzheimer’s disease, in the translucent human brain. That has never happened in the history of neuropathology. So technical innovations for brain mapping will contribute both to understanding the normal function of the healthy brain and also disease states.
For this purpose, we need more advanced super computers, so these projects will also have an influence on computer science.
NEWSOME: I agree. I think that the advances in raw computer power and in the techniques of machine learning and data mining are fundamental to this new revolution in neuroscience.
Another game-changer is the technology that Dr. Okano mentioned a few minutes ago, direct gene editing. He should speak to this because he’s the expert here…
OKANO: Sure. Feng Zhang at MIT and the Broad Institute is developing in vivo genetic editing using the mouse brain. If this could be used in nonhuman primates and combined with behavioral analyses, we may make real progress toward understanding psychiatric disorders or autism-related disorders.
NEWSOME: Let me get that straight. You said direct gene editing in vivo. You mean in an adult animal?
OKANO: An adult animal. Yes.
NEWSOME: That's amazing.
HILL: We also think that there’s a lot to learn from the brain to develop novel computing technologies. The brain is in many ways the most powerful supercomputer on the planet and yet it also an extremely low-power computing device. So we think there’s a tremendous amount to learn from it about how to achieve brain-like computation with very low power.
TKF: Professor Newsome, in September the BRAIN Initiative announced that it had attracted a number of new private partners in the form of foundations and industry. What role does the private sector currently play and how do you think it will evolve?
NEWSOME: One of the most obvious roles for the private sector is that the big data giants like Microsoft and Amazon and Google have data handling, data storage, data curating, data access, and data analysis techniques that are far in advance of what’s available in universities and in government laboratories. Some representatives of those companies are very interested in teaming with the BRAIN Initiative. Marrying their technology and know-how to brain research is a very intriguing possibility. But we have to be sure that we protect the public nature of those data. We have to ensure that such partnerships would serve a large, national purpose.
Another area where corporate partners could be important is in developing devices for getting data out of the nervous system and developing new therapies for disease conditions. A peculiar thing happens in universities. The engineers are justifiably interested in what’s called “research-grade” engineering—a proof of principle of what a solution would look like. That’s where the real intellectual rewards lie in engineering departments. So once they’ve got that proof of principle, they tend to move on to the next problem. But developing a real device—not just proof of principle—that can enhance experiments in neuroscience laboratories or treatment in patients requires a huge development pipeline. Corporate partners might be better for that sort of long-haul engineering than university engineers. An unresolved question in my mind is, what is the motivation for the companies? Does the potential market for these devices justify the effort that must be made by the companies?
TKF: How will each of you judge the success of the Brain/MINDS, HBP and the BRAIN Initiative in 10 years?
NEWSOME: The NIH working group was charged to state milestones and deliverables for the BRAIN Initiative. It was an interesting process because this project is not as simple as the Apollo program or the Human Genome Project. In fact, when NIH Director Francis Collins called me and asked me to serve as co-chair of this working group, one of the first things I said to him was, “You understand that this is a lot more complicated than the Human Genome Project. Neither you nor your successor is going to be able to stand here 10 years from now and say this project was completed on time and under budget.”
That’s because the goal of the BRAIN Initiative is much more complex. Our own brains are trying to understand themselves. That is a deep scientific voyage, not an engineering project. Where are we going to put a flag down and say “now we understand the human brain?” What I’m convinced that there are certain benchmarks that are prerequisite to understanding the human brain and we should be keeping our eyes on those benchmarks.
One benchmark is how many types of brain cells there are. We can’t understand the system without having a parts list. So 10 years from now, we should have a reasonably complete inventory of cell types in strategically important parts of the brain. And for select parts of the brain we should have independent genetic access to those cell types to be able to manipulate them.
We also need to have much better connectivity maps of a wide variety of experimental animals and we need to make a good start on human connectivity maps at the end of 10 years. After all, how will we understand a system without delineating basic wiring diagrams?
Another milestone that we need to evaluate is large-scale recordings. The Brain Activity Map—a BRAIN predecessor—called for recording every action potential from every neuron in the brain. I don’t think we need every action potential from every neuron, but we certainly need action potentials from a lot more neurons than we have right now. Exactly how many is necessary is a deep theoretical question that requires serious input from theorists and statisticians. They must help us answer this question in a principled way in the next five to ten years.
So those are a few concrete metrics of success. If we’re hitting those benchmarks at 10 years, we are progressing toward the ultimate goal of understanding the brain.
OKANO: Brain/MINDS is a very ambitious project so it will require a lot of effort to make it successful. Our ultimate goal is to strategically promote brain science research that will benefit society. Particularly, if we could contribute to conquering devastating brain disorders such as Alzheimer disease and schizophrenia through Brain/MINDS, then that would be a great success. Obviously, for this goal, we need the collaboration of people working outside of neuroscience and also efficient cooperation among the global brain projects.
HILL: I felt a little bit like I stepped into the future when I went to a meeting of the virtual observatories in astronomy. They’ve already done a lot of the things that we’re aiming to do in the sense that they’ve got observatories from all around the world that are sharing, through a common infrastructure, the data that they’re measuring. And they’re integrating it into a unified picture of the sky and our universe. These same data contribute to building large-scale simulations of the universe. I think that’s extremely inspiring. If we can have a common infrastructure that allows us to have a globally integrated view of the data being produced, and the tools to run large-scale simulations from the data, we will really have made progress in neuroscience.