TECHNOLOGY FEATURE BRAIN MAPPING IN HIGH RESOLUTION Tools that make it possible to chart every neuron and its connections are helping neuroscientists to realize their dream of whole-brain maps. KHENG GUAN TOH/SHUTTERSTOCK KHENG GUAN BY VIVIEN MARX and Tracts (CONNECT), funded by the Euro- of neurons — the axons and dendrites — are pean Commission in Brussels, are producing 20 micrometres or less in diameter, but can esearchers wanting to understand the magnetic resonance imaging (MRI) maps that extend for several millimetres. Mapping the workings of the brain need maps on follow neuronal connections across the entire thicket of 86 billion neurons and their connec- many different scales — to link struc- brain, on a scale of tens of centimetres. tions in the human brain at this scale is still Rture to function, to parse the complexities of Some research teams are embarking on even a distant dream, but the goal of mapping the memory loss or learning disability, or to find more ambitious projects to create maps that 75 million neurons in the mouse brain could out what is different about the brains of people reveal structure and connections at the scale be a little closer. And the second might pave with neurodegenerative diseases. For instance, of individual neurons. The slender processes the way for the first. programmes such as the Human Connectome To achieve this, however, researchers need Project, funded by the US National Institutes NEW ANGLES ON THE BRAIN new tools, starting with refinements to elec- of Health (NIH) in Bethesda, Maryland, and A Nature special issue tron microscopy. In conventional electron- the Consortium of Neuroimagers for the Non- www.nature.com/neuroscience203 microscopy approaches, scientists have to invasive Exploration of Brain Connectivity make ultrathin slices of brain tissue and 7 NOVEMBER 2013 | VOL 503 | NATURE | 147 © 2013 Macmillan Publishers Limited. All rights reserved NEUROBIOLOGY TECHNOLOGY laboriously image them slice by slice to SHARP TOOLS build up a three-dimensional (3D) picture. One manufacturer is developing wide diamond knives that can be used In 1986, the complete nervous system of to shave o samples containing a slice of the whole mouse brain. one model organism was mapped using this SOURCE: DIATOME approach — the 302 neurons of the nematode KnivesText in here current please! Caenorhabditis elegans1 — but the process is use are usually too slow and cumbersome to scale up. Now, around 1 mm wide. electron microscopes at many universities are “being mothballed”, says neurobiologist Jeff Lichtman at Harvard University in Cambridge, Massachusetts, because they are deemed a throwback to an era when “all you could do was stain and look”. Nevertheless, high-resolution, 3D neuronal mapping by electron microscopy is gain- ing momentum. “It’s not just the old neuro­ Wide diamond knife has to be anatomy gussied up with new machines,” made from a perfect crystal. says Lichtman, who studies the neuronal architecture of the mouse brain. “It’s giving us three-dimensional information about structure at the super-resolution of nano­ Cutting edge is only metres — and that is invaluable.” These capa- 4 nanometres wide bilities are spurring the development of new when polished. ways to prepare, image and analyse brain tis- sue from model organisms such as mice, and Brain tissue embedded in resin from people who have donated their brains to science on their death. gains could make it feasible to map all the in the electron microscope, then cutting off an Until now, when neuroscientists published neurons in a mammalian brain. Other micro- ultrathin slice using an automated microtome cell-level images and analysis, they were look- scope manufacturers, such as FEI in Eind- within the instrument. The newly exposed ing at a tiny region of brain; for example, a few hoven, the Netherlands, are also developing surface of the sliced block is rescanned, and so hundred neurons in the mouse or fruitfly2,3. tools to enable detailed brain mapping. on until a stack of images has been obtained. Efforts to map all the neurons and circuits in The advantage of imaging the unsliced tissue, the mouse brain are just beginning in several PREPARING TO SEE Denk says, is that it does not matter if the slice labs. Creating structural maps is a high priority Before brain tissue can be imaged using elec- itself crumples. of the NIH-funded Brain Research Through tron microscopy, samples must be prepared Until now, Denk and his team have cut only Advancing Innovative Neurotechnologies by slicing and staining. These steps are impor- from pieces of tissue that measured much less (BRAIN) Initiative, which US President Barack tant for the subsequent process of using the than 1 mm across. But Denk’s ultimate aim is to Obama launched in April4. Another large-scale two-dimensional images generated by the image and cut slices from a whole mouse brain, effort is the Human Brain Project, funded by electron microscope to reconstruct the three which will mean dealing with tissue blocks that the European Union, which is focused on a dif- dimensions of the brain and its neuronal con- are some 10 mm across. Denk is now building ferent kind of map — a computational model nections. Scientists are exploring new ways to a whole-brain microtome to incorporate into of the brain. prepare samples so that they can obtain higher the microscope. This will require a diamond Neurobiologist and microscopist Winfried contrast for electron microscopy and view knife that is 8–10 mm wide instead of the Denk at the Max Planck Institute for Medi- larger pieces of brain tissue. 1-mm knife used at present, he says. cal Research in Heidelberg, Germany, thinks When brain tissue is well stained with chem- To make these knives, he is collaborating that the technology to icals, thinly sliced and viewed using electron with diamond-knife manufacturer Diatome image an entire mouse microscopy, researchers see slices that look in Biel, Switzerland. The longer blade width brain at cellular reso- like thin plates covered with tiny soap bubbles. is a challenge, says Diatome engineer Helmut lution is within grasp. Some bubbles represent the cross-section of a Gnaegi. At 10 mm, it is around ten times wider Late next year, he and neuron, and the soap-bubble boundary shows than a typical diamond knife, and the blade Lichtman are each where one neuron ends and another begins. must be made from a large diamond that is scheduled to receive a Contrast is crucial, because scientists need to free of crystalline imperfections (see ‘Sharp new multi-beam scan- discern neurons from a wealth of other cell tools’). ning electron micro- types and organelles in the brain tissue. They The entire 10-mm cutting edge must be scope (SEM) for their can then trace each neuron in each of the polished to perfection, Gnaegi says, because labs, developed by imaged slices, by hand and by eye. even a single knife mark would interfere with W. DENK/MAX PLANCK INST. FOR MED. RES. FOR W. DENK/MAX PLANCK INST. Carl Zeiss Microscopy But such imaging has its problems: stain- the smooth sample surface that Denk requires. in Oberkochen, ing often does not show enough structure, After polishing, the cutting edge must have a Germany. Now in the slices might be warped and images can be maximum thickness of only 4 nm. This is the prototype stage, Winfried Denk blurry. a “formidable challenge” that the company this instrument will estimates that Denk and his team previously developed hopes to overcome by the end of this year, says enable the imaging image capture of an approach called serial block-face electron Gnaegi. He and his team have designed special of brain slices to be a whole mouse microscopy (SBEM), which obviates the need polishing equipment for the task. accelerated hugely — brain could yield to prepare slices before imaging them. Instead, Gnaegi’s team is also addressing the build- possibly by as much 60 petabytes of successive images are obtained by scanning up of electrostatic charge that occurs when as 60 times. Such data. the face of an unsliced block of tissue placed the knife tries to cut through tissue that has 7 NOVEMBER 2013 | VOL 503 | NATURE | 149 © 2013 Macmillan Publishers Limited. All rights reserved TECHNOLOGY NEUROBIOLOGY been embedded in non-conducting resin. The reading’). It also dispenses with features such charge tends to make the sections stick to the as energy-dispersive X-ray spectroscopy detec- sample surface, which makes imaging difficult. tors or back-scattered electron detectors, both “An electrically conductive knife surface with of which are used in standard electron micro- optimized gliding properties leads to a sample scopes to map variations in a sample’s chemical surface free of cutting debris,” says Gnaegi. composition. CARL ZEISS MICROSCOPY The SBEM approach helps with precise ren- Dellemann explains that including several dering of tissue in three dimensions, but it is detector types would mean incorporating hard to make it high-throughput. Researchers three arrays of 61 detectors each, and would would have to set up many microscopes in a require a mechanism to guide the right signal dedicated facility. The alternative is an instru- to each one. Although possible, such a configu- ment that is “intrinsically a parallel imaging Dirk Zeidler (left) and Gregor Dellemann are ration would add cost and complexity to the machine”, Denk says. That is why his lab is so developing the new Zeiss microscope. instrument and slow it down. interested in the Zeiss multi-beam machine Some scientists had voiced doubts about that is due next year. Looking at images generated by the device the multiple-beam SEM when they first heard Lichtman and his team take a different has been “very exciting, because you realize about it, says Zeidler, who notes that they even approach to sample preparation.