The Brain Activity Map, an academic consortium that aims to measure and model all the connections in living brains, has garnered mass media attention and comparisons to the Human Genome Project. However, unlike mapping the relatively well-defined human genome, visualizing the brain's complete wiring faces severe technical challenges. Also, it is unclear what the outcome of such a project would be and whether it would get new treatments to patients faster than more focused research.

The project's leaders have proposed a timeline of about 20 years for developing the technology to collect and interpret data from human brains. Popular press coverage of the Brain Activity Map (BAM) has focused on the idea of being able to image the activity of the entire brain in real time but has largely overlooked the project's massive technical challenges.

Debate about BAM within the neuroscience community has focused on the wisdom of channeling research funding for non-hypothesis-driven big science.

BAM consortium member Rafael Yuste, professor of biology at Columbia University, told SciBX that the project's goals coalesced from a series of meetings in 2011 and 2012 organized by The Kavli Foundation, the Allen Institute for Brain Science and The Gatsby Charitable Foundation. All three are private philanthropic organizations focused on basic neuroscience and nanotechnology research.

Over the course of these meetings, BAM evolved from a basic research technology wish list into an overarching plan for a great leap forward in brain imaging technology.

The project's framework was described in a conference white paper1, a review in Neuron last summer2 and in a perspective piece in Science this week.3

President Barack Obama mentioned the project in his State of the Union address as a grand scientific challenge akin to the Human Genome Project. Thereafter, comments on Twitter by NIH director Francis Collins about BAM sparked speculation about the NIH's involvement in the financing and administration of the project.

No official announcement by the White House or NIH about BAM has yet appeared, and there is no timeline for one.

Modeling the brain

Although proponents of the project are likening it to the Human Genome Project, both the scale and scope of the projects are different.

In the case of the Human Genome Project, the task was finite, with the genome initially thought to consist of about 25,000 genes. In addition, the technology to sequence those genes existed, even if it needed to be scaled up.

There also was broad consensus in the research and clinical communities that sequencing the genome would have broad and immediate utility. In practice, the human genome's organization and function proved far more complex than expected, but the output of the Human Genome Project proved a useful starting point for subsequent genomic studies.

In contrast, the connection map of the brain proposed by the BAM project may not be consistent from one brain to another or even from one moment to the next.

Whether or not the analogy to the Human Genome Project is apt, BAM's goals and the technology to achieve them are not yet clearly defined.

BAM aims to build models of the brain in action from data gathered with a range of optical and electrophysiological technologies that do not yet exist. Because existing tools to study the brain capture only a fraction of its activity, BAM's first objective is to develop new technologies that can capture the full range of chemical and electrical activity of every neuron.

The consortium also wants to develop optogenetic and nanoelectronic methods to experimentally manipulate the activity of single neurons in living brains. The goal of this approach is to uncover the functional significance of every synaptic connection.

A third goal of BAM is to develop computational methods that model and eventually emulate the workings of the human brain.

"We're proposing a very massive project that could take a decade or two, on the scale of the Human Genome Project," said Yuste.

R. Clay Reid, senior investigator of neural coding at the Allen Institute, said the groundwork exists for the imaging technology but requires considerable scaling up.

"We've gone from being able to record from a handful of neurons to being able to routinely record from a thousand at a time. Now we're discussing going to 10,000 or 100,000 neurons," Reid said.

Yuste said the project will first test these imaging and recording technologies in small animals such as Caenorhabditis elegans and will work its way up to progressively larger and more complex organisms.

"We're proposing that in five years we might be able to reconstruct the nervous system of the worm and a significant portion of the nervous system of the fly. Maybe in 10 years you can do this on the scale of a million neurons, on the scale of the zebrafish brain or mouse retina," he said.

Yuste said it may eventually be possible to image and record the activity of the entire human brain and to externally manipulate complex neural circuits that influence behavior and perception.

He said a complete understanding of the brain's workings will help clinical researchers identify brain activity signatures of complex neurological diseases such as epilepsy and schizophrenia.

"You could have new types of diagnostics and therapeutics," said Yuste. "If you could map circuits in the brain with this level of detail, you could map the spread of epileptic seizures and observe which neurons are participating or not. If a brain region is involved in the spreading of seizures, you could stop that process" with external nanoelectronic manipulation.

Delivering insights

Even if BAM manages to build the technology needed to capture whole-brain data, not all neuroscientists are convinced the project will be able to deliver meaningful insights into the human brain's high-level functions, let alone reconstruct a nervous system.

Eve Marder, professor of biology at Brandeis University, said the proposed imaging technology could potentially be useful in studying larger numbers of neurons than is possible with current techniques but was uncertain whether the approach could scale up to the whole-brain level.

"Everyone agrees that there's value to getting multiple recordings from brains during behavior and that some problems will not be solved by recording only from single neurons, one at a time," said Marder.

However, she suspects the scale of activity in the whole human brain will defy interpretation. She said it is already difficult to interpret data from recordings of thousands of neurons, so piling on more data may not lead to greater insights.

 "We already have the ability to collect more data than we know how to visualize or think about," she added.

Marder said it was not clear what aspects of whole-brain activity are functionally relevant or pertain to specific behaviors, so large data sets of overall brain activity will likely be hard to interpret.

"If your goal is to understand the brain, we don't have any consensus on what that even means," said Marder. She said the theoretical framework for relating the brain's electrophysiological activity to behavior is ill defined.

To address that concern, Yuste and Reid said BAM will use "big data" computational methods to identify hidden patterns of brain activity and correlate them with behavior. This will require developing new analytical methods and building massive computational facilities.

Even if BAM succeeds in capturing and analyzing the totality of brain activity, it is not clear whether this information would yield insights into high-level brain functions such as cognition and memory.

Partha Mitra, professor of neuroscience at Cold Spring Harbor Laboratory, argued in a Scientific American article4 that the complex functions of the brain are unlikely to arise merely from the sum of every neuron's activity but rather are influenced by external stimuli and prior history.

If so, BAM's proposed high-resolution snapshot of the brain's activity would not necessarily make it possible to reconstruct brain activity in silico.

Money woes

Another concern is how to fund the project during a period of shrinking basic research budgets. The New York Times estimated that the total cost of BAM will be in the range of $3 billion.5 By comparison, the National Institute of Neurological Disease and Stroke's budget for 2013 is $1.6 billion.

Yuste said he expects the NIH, the White House's Office of Science and Technology Policy and the U.S. Department of Defense's Defense Advanced Research Projects Agency will announce how BAM will be funded this month.

Reid said private institutions like the Allen Institute and the Howard Hughes Medical Institute would provide some financial support, but the bulk of the funding would come from government agencies.

Cornelia Bargmann, professor and laboratory head at The Rockefeller University and a Howard Hughes Medical Institute investigator, is worried that concentrating neuroscience research money into BAM would reduce small grants to academic researchers.

"These are hard times for federal funding of research grants. I would want to be very, very sure that this was new funding and not resources that are taken away from a system that is already deeply stressed. An end around the normal scientific review process would be a mistake," Bargmann told SciBX.

The BAM team wrote in a Science perspective that they want the project to be funded by new commitments from the government, not from existing neuroscience funds.

Even if BAM is ultimately paid for by new money, Bargmann is skeptical that funding large research centers is the best way to advance neuroscience.

"Are we talking about central planning inside the Beltway and the genome model of large 'centers'? I would not be a fan even if that were new money; that's not what the field needs," said Bargmann.

Marder also was skeptical about launching another big science project.

"Neuroscience is at a moment where we could make tremendous advances in the next 10-15 years. This will require a mixture of larger- and smaller-scale projects, but there are a lot of good people who are being pushed out of the field because there isn't enough money," said Marder. "Diverting the meager resources that we have right now for bigger projects would be a mistake."

Reid expects the funding will be distributed to a combination of large regional academic centers and smaller laboratories. He said it will take some time to build up the technology and imaging facilities ultimately envisioned by BAM, so for now the funding will go to smaller labs.

"I believe that neuroscience is ready for some of it being big science, but the majority of the work will be done in smaller groups for the foreseeable future," said Reid.

Gone fishing

Pharmaceutical Research and Manufacturers of America (PhRMA) put out a general statement of support for BAM, but there are no authors from industry on either of the two papers the BAM team has published.

Douglas Williams, EVP of R&D at Biogen Idec Inc., said BAM appears to be "a fundamental science project that is going to generate insights, but it's hard to predict where and when the payoff will be."

He added, "In a world of unlimited dollars, I'd love to see this done, but I think there are very pressing public health threats in neurology that are more important. It's not that this isn't a fascinating project to undertake, but public health considerations should garner more basic research funding."

Although the project's ultimate goal is to image the human brain, the most useful aspect of the work may be imaging the brains of mice to assist in preclinical studies.

Yuste said it may be possible to monitor the effects of drug candidates on whole-brain activity to gain insights into how they affect brain functions that go awry in disease.

"This technology will provide a new type of assay from brain-related pathology that could be used for drug discovery," said Yuste. "You could image or record the effect of a drug on every neuron."

Williams said another potential application could be the development of high-resolution brain imaging tools for clinical biomarker studies. "I see this project as a generator of new technologies," he added.

Benjamin Matteo, cofounder and CEO of Eos Neuroscience Inc., said methods developed by BAM could make it easier to evaluate the effect of therapeutics that stimulate specific neuronal activity. Eos is developing optogenetic gene therapies to treat blindness.

Eos' most advanced candidate is EOS-13, an adeno-associated virus (AAV) vector that encodes channelrhodopsin-2 (ChR2), an algae-derived, light-activated ion channel protein. EOS-13 has completed preclinical testing for retinitis pigmentosa.

"We can now restore the ability of blind mice to navigate and can record populations of ganglion cells to show that these cells are firing normally in space and time," said Matteo. "Right now we're not at a stage [where] we can say what the animal is seeing, but we can tell that the mice are seeing something. More resolution in real time would give us more useful data" about the extent of vision restoration from EOS-13 therapy.

Regardless of the benefits of BAM's proposed tool building, Matteo thinks the project could benefit from a more clear focus on disease.

"Better tools for recording and understanding circuits is only a part of the answer. Other pieces will come from hypothesis-driven work," said Matteo. "If you ask me whether to spend $3 billion on this purely basic work, I would say no; there needs to be more of a translational component."

Osherovich, L. SciBX 6(9); doi:10.1038/scibx.2013.206 Published online March 7, 2013

REFERENCES

1.   Yuste, R. et al. The Whole Brain Activity Map: merging nanoscience and neuroscience for technology and health. (2011)

2.   Alivisatos, A.P. et al. Neuron 74, 970-974 (2012)

3.   Alivisatos, A.P. et al. Science; published online March 7, 2013; doi:10.1126/science.1236939

4.   Mitra, P. What's wrong with the brain activity map proposal. Scientific American (March 5, 2013)

5.   Markoff, J. Obama seeking to boost study of human brain. The New York Times (Feb. 17, 2013)

COMPANIES AND INSTITUTIONS MENTIONED

Allen Institute for Brain Science, Seattle, Wash.

Biogen Idec Inc. (NASDAQ:BIIB), Weston, Mass.

Brandeis University, Waltham, Mass.

Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

Columbia University, New York, N.Y.

Defense Advanced Research Projects Agency, Arlington, Va.

Eos Neuroscience Inc., San Francisco, Calif.

The Gatsby Charitable Foundation, London, U.K.

Howard Hughes Medical Institute, Chevy Chase, Md.

The Kavli Foundation, Oxnard, Calif.

National Institute of Neurological Disease and Stroke, Bethesda, Md.

National Institutes of Health, Bethesda, Md.

Office of Science and Technology Policy, Washington, D.C.

Pharmaceutical Research and Manufacturers of America, Washington, D.C.

The Rockefeller University, New York, N.Y.

U.S. Department of Defense, Washington, D.C.