I thought I was already excited by the amount of science-talk in Obama’s State of the Union Address this year, and then I picked up the New York Times this morning. The Times reports that President Obama’s federal budget proposal in March will likely include billions of dollars for a massive ten-year initiative, which aims to revolutionize systems neuroscience by developing technology to map the activity of every neuron in the brain in real time. Such a windfall may seem like too much to hope for from a budget-sensitive Congress, but don’t forget that in a fit of austerity the European Union recently promised a billion euros to Henry Markram’s Blue Brain project to merely simulate a complete neural circuit in silico. It may turn out to be a good year for Big Neuroscience.
For the moment, the details of the new initiative are still under wraps, but it is reported to involve a collaboration between NIH, NSF, DARPA, HHMI, and the Allen Institute for Brain Science, where Ricardo Dolmetch, Cristoph Koch and Clay Reid are already working on a rather similar-sounding projects.
The new NIH plan seems to be based on an article called “The Brain Activity Map Project and the Challenge of Functional Connectomics,” published in Neuron last June by A. Paul Alivisatos, Miyoung Chun, George Church, Ralph Greenspan, Michael Roukes, and Rafael Yuste.
In this manifesto, the authors argued that the scientific and technological time is right for large scale neuroscience – for researchers to tackle the emergent properties of vast interacting neuronal networks. Systems neuroscientists have begun to suspect that these larger networks, in which individual cells can participate on an ad hoc basis, are where the real information about the world is encoded, rather than the single-cells that have been a staple of neural physiology for so long. Emergent network properties are also likely to be the key to understanding complex neurological and psychiatric disorders such as autism and schizophrenia.
The authors proposed the Brain Activity Map (BAM) Project, upon which the current Obama initiative appears to be based, as a grand public effort akin to the Human Genome Project, with a goal of scaling up technology that could “record every action potential from every neuron in a circuit.” Their plan would begin at relatively low resolution using well-established 2-photon calcium imaging techniques, followed by direct imaging of electrical fluctuations in individual cells using advanced voltage-sensitive dyes or nanoparticles such as quantum dots or the much more exotic-sounding “nanodiamonds.”
Because of the limits of imaging through turbid neural tissue, these first two steps would be limited to measuring activity in the most superficial regions of the brain. To record from the murky depths of brain, the BAM Project would depend on the development of new high-tech methods. These might include nanoprobes – miniaturized electrodes with hundreds of thousands of recording sites – or wireless nanocircuits – nanobot-like devices which the authors visualize releasing “untethered in living brains.”
The Neuron article goes even further into the bleeding edge of future technology, proposing that it might be possible to use genetically engineered cells to transiently attach to neurons to stimulate activity or to record spikes in a ticker-tape of DNA. They write, “In principle, a 5-μm-diameter synthetic cell could hold at least 6 billion base pairs of DNA, which could encode 7 days of spiking data at 100 Hz with 100-fold redundancy.”
The paper estimates that within five years, such full-scale recordings could be done at the scale of the C. elegans nervous system, the medulla of Drosophila, or retina of the mouse. Within ten years it might be possible to record complete activity maps of of the whole Drosophila or zebrafish brain or a mouse cortical area, then the entire neocortex of an awake mouse within 15 years, after which it would be time to proceed towards primates.
If you are reeling with conflicting excitement and suspicion of such grandiose plans and the prospect that the government might actually throw billions of dollars at them, you are not alone.
I asked systems neuroscience guru Nick Steinmetz for his take on the BAM Project, and he echoed my own concern with the authors’ over-optimistic faith in Moore’s Law to bring us to the technological promised land. “It's telling, to me, that Yuste cites a 1991 paper that he was on as his reference for calcium imaging of neural activity, which only in the last five years has been able to be used to measure large numbers of neurons in vivo. It took 15 years to go from having this technology to using it, and now you're proposing less than 10 years to develop totally new technologies and use them?”
Despite the obligation to drop his “quanta of GABA” on the enthusiasm for the project, Nick also pointed out that a big funding push for systems neuroscience would almost certainly produce much-needed data. I tend to agree that whatever ends up happening to the details of the timeline, this project could move neuroscience forward into a new data-intensive paradigm. Nick and I both loved the idea proposed in the Neuron proposal for a set of neuroscience National Observatories to process the overwhelming amounts of data likely to be pursued by the initiative.
All in all, it seems like a renewed Federal interest in massive funding for systematic neuroscience research can’t be a bad thing, even if it takes a few years longer to fill your brain with nanobots than the NIH would like.