One of the early goals of the BRAIN Initiative is a “parts list” for the brain. While we have a pretty good idea of the various cell types in most organs, for the brain we still lack a comprehensive list of the kinds of cells and how many of each kind of cell should be present. This is a surprisingly hard problem, partly because of the sheer numbers and diversity of cell types in the human brain. In mid-May, Evan Macosko and his colleagues at Harvard and MIT reported on Drop-seq, a new technique for single cell analysis.1 Drop-seq is a high throughput system for identifying cell types by the genes they express using bar-coded beads in microscopic droplets. The system is fast (processing 10,000 cells in 12 hours) and inexpensive (less than 7 cents per cell). As a test case, the team studied over 44,000 cells in the mouse retina and identified 39 cell types, consistent with previous reports using far more labor- intensive methods.
Another goal of the BRAIN Initiative is to move from correlational studies to causal or mechanistic studies of brain function. By delivering genetically altered receptors or channels into the brain, cells and circuits can be turned on or off with great precision, allowing a better test of causal mechanisms. Optogenetics is one version of this approach, using a pulse of light from an implanted laser to control the activity of individual cells deep in the brain. Another version, sometimes called chemogenetics, uses the Designer Receptor Exclusively Activated by Designer Drugs or DREADD technology which targets a genetically engineered receptor to a specific cell type. Because this receptor can only be activated by a designer drug that activates no other receptor, specific cell types can be influenced more precisely than is possible with natural receptors and current drugs. Last month Bryan Roth and his colleagues at the University of North Carolina reported on a new generation of DREADDs, based on the kappa-opioid receptor but activated by a pharmacologically inert drug. This new designer receptor reduces neuronal activity, meaning that DREADDS can be used for bidirectional control of neuronal circuits.2
Next generation imaging has been a major effort in the first year of the BRAIN Initiative. Early last month, Lihong Wang and his colleagues at Washington University and Texas A&M University reported on a new form of non-invasive imaging called photoacoustic microscopy.3 With this novel technique, single capillaries and the oxygenation of blood can be visualized in the brain through the intact skull of a mouse. This method achieves much higher spatial and temporal resolution than any current method for in vivo brain imaging through the skull. While still in development, photoacoustic imaging could be a new modality for human brain imaging in the future.
In addition to imaging, direct recording of brain signals is ripe for a revolution. Developing new ways of recording brain signals will require engineering, materials science, nanotechnology, and a lot of luck. Charles Lieber, an NIH Pioneer Awardee, and his team at Harvard just described syringe-injectable electronics.4 This technique involves injecting a polymer net that is a soft, conductive electrode array covering the cortex for direct recording. Their first experiments in mice show that this approach can be used to record from and stimulate thousands of cells. Presumably a similar approach could someday be used to record from human cortex without major neurosurgery.
Will the BRAIN Initiative allow paralyzed people to walk? If we could understand brain signals well enough to “hack” them, could we use these signals to drive an interface that could allow someone with a spinal cord injury to grasp or walk? If this seems like science fiction, check out this video of work from Andy Schwartz at the University of Pittsburgh.
Perhaps even more amazing was a report last month from Richard Andersen and his colleagues at CalTech, entitled “Decoding motor imagery from the posterior parietal cortex of a tetraplegic human.”5 In contrast to previous studies demonstrating that signals from the motor cortex can drive a robotic arm, this new report uses a posterior part of the brain associated with attention and intention. Electrodes implanted in this region were capable of driving a robotic arm to reach and grasp, including signals that controlled both left and right sides. The goal, of course, is to use brain signals to drive the person’s own limbs, not robotic limbs, but these early successes with robotic arms suggest that the codes for controlled movements can be channeled perhaps from multiple areas of the cortex and that getting people out of wheelchairs could be one of the achievements of the BRAIN Initiative.
This is still the very beginning of what President Obama called “the next great American project.” None of us can quite imagine where the BRAIN Initiative will lead, but the first year has already revealed more progress than any of us could have expected. Freeman Dyson, speaking of astrophysics, noted that “New directions in science are launched by new tools much more often than by new concepts.” 6Judging from this first year, it appears this insight will hold true for neuroscience as well.
1 Macosko EZ et al. Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets. Cell. 2015 May 21;161(5):1202-14. doi: 10.1016/j.cell.2015.05.002.
2Vardy E et al, A New DREADD…, Neuron, May 2015
3 Yao J et al. High-speed label-free functional photoacoustic microscopy of mouse brain in action. Nat Methods. 2015 May;12(5):407-10. doi: 10.1038/nmeth.3336. Epub 2015 Mar 30. 4 Liu J et al. Syringe-injectable electronics. Nat Nanotchnol. 2015 Jun 8. doi: 10.1038/nnano.2015.115.
6 Dyson F. Imagined Worlds. Cambridge, MA: Harvard University Press, 1997.