The V3D software developed by computer scientist Hanchuan Peng produced this visualization of networks of neurons in the fruit fly brain.
Deducing Blueprints From Design
Neuroscientists who fill their labs with worms and flies confront bigger challenges than the animals’ relative simplicity belies. It’s true that these creatures’ nervous systems have only a fraction of the human brain’s daunting complexity, but mapping how their neurons connect to one another—which could reveal some fundamental principles about how neural networks function in all animals—requires not just patience and determination but also some astonishingly sophisticated tools.
A microscope image that reveals the connections between individual neurons can capture only a fraction of even a fruit fly’s tiny brain, which is a mere third of a millimeter in diameter. To achieve that level of detail in a picture of the whole structure, researchers must take many snapshots and then fit them together to construct a panoramic view.
It’s tempting to let a computer handle the tedious task of reassembling the images, but image interpretation doesn’t come as easily to a computer as it does to a human, and what’s more, a single high-resolution image from a fluorescent microscope can contain a processor-clogging 3,000 megabytes of data.
A new software package developed by Janelia Farm scientists is now being used worldwide to accelerate the assembly and manipulation of these detailed images. Computer scientist Hanchuan Peng led the development of the freely distributed package, called V3D, which can assemble three-dimensional images of neurons 17 times more reliably than commercial software. A collection of associated tools lets users easily analyze images and retrieve information of various types, such as the number and lengths of branches along a particular segment of a neuron.
The roundworm Caenorhabtidis elegans is even tinier and simpler than a fruit fly; it is about the size of a human eyelash and has just 302 neurons. Scientists already know exactly how those nerve cells connect to one another. In fact, each cell in the worm’s body has a name and has been mapped in published “worm atlases.” But under a microscope, a real worm never looks quite like a drawing in a book. Even experienced scientists can spend days matching up the cells they see in an individual worm with those on the official C. elegans map.
Janelia Farm group leader Eugene Myers, together with Peng and their Janelia colleague Fuhui Long, have now saved scientists the trouble. The team has created an interactive “digital atlas” that can interpret an image and identify most of a worm’s cells automatically. Researchers can prepare a worm for microscopy, snap a digital image, and in a few hours retrieve a detailed anatomical map—which dramatically speeds experiments.
Already, the Janelia team has collaborated with researchers at Stanford to use the atlas to study how genetics determines the fate of developing cells. In the future, the atlas might be used to guide lasers that destroy or activate cells to examine their function.
