A mantis shrimp’s eye is only a few millimetres across, but its intricate mix of hard and soft tissue makes it unusually difficult to image. Using a building-sized synchrotron in Sweden, researchers captured the entire eye in 3D while still resolving structures measured in micrometres – and uncovered a vessel-like network in the retina that had not been described in mantis shrimps before.
In 1895, Wilhelm Conrad Röntgen exposed his wife’s hand to X-rays long enough to produce a ghostly silhouette of her bones and wedding ring. In the 131 years since, scientists have progressed light-years in imaging, using technologies from magnetic resonance to electron microscopy to peer inside everything from living human bodies to hard-won samples from space.
One technique pushing imaging further is synchrotron-based phase-contrast X-ray computed tomography. In principle, it is a lot like computed tomography or X-ray—with one enormous difference. Instead of a hospital X-ray source, it relies on a building-sized synchrotron in which electrons travel at almost the speed of light. This means that researchers can image an intact object several millimetres across and still zoom in on structures measured in micrometres.
Researchers at Aarhus University in Denmark put a synchrotron imaging set-up in Sweden through its paces with an impressive test case: a mantis shrimp’s eye, a devilishly complicated biological structure about the size of a sesame seed that enables these ocean-dwelling creatures to see both ultraviolet and polarised light. Their results, published in the Journal of Structural Biology, show how the technique can bridge a scale that is difficult to capture with conventional imaging.
“We see the entire eye, measuring five by six millimetres, but we can also zoom in and see something the size of a micrometre,” explains lead author Anne Marie Møller Faaborg, a materials chemist and doctoral researcher at Aarhus University. “This is one of the only techniques that allow this jump in scale.”
Meet the mantis shrimp
The research group Faaborg is part of first encountered the mantis shrimp while studying the remarkable material properties of its claw. “This all started because of this hammer that they have,” says project leader and co-author professor Henrik Birkedal, a materials chemist at Aarhus University.
Although it is only 10 to 18 cm long, the peacock mantis shrimp packs an astonishing punch. It uses concussive force to break open mollusc shells, and its dactyl club strikes so fast that it creates vapour bubbles in the surrounding water, which collapse with enormous force and can emit tiny flashes of light. Its eyes, which move independently on stalks, are also extraordinary.
“They can see polarised and ultraviolet light, and the structures that allow them to do this are quite different from the human eye,” Faaborg explains. Mantis shrimp have compound eyes built from thousands of units called ommatidia. Each contains a lens, followed by a crystalline cone and then light-sensitive cells arranged around a central structure called a rhabdom – thousands of each, all packed into a space about the size of a sesame seed. “There is a lot to look at inside this eye, both in soft tissue and in hard tissue,” she says.
Since the 1700s, researchers have painstakingly pieced together the structure of the mantis shrimp’s eyeball through careful dissection, staining and the best microscopy of the age. These methods revealed many of its individual components but did not preserve their full 3D context. To bring this trajectory to the 21st century, Birkedal, Faaborg and their team took the right eye of a peacock mantis shrimp to an imaging centre in Sweden.
Meet the synchrotron
Objects the size of the mantis shrimp’s eyeball fall in an awkward no-man’s-land when it comes to imaging.
Conventional X-rays and computed tomography work mainly by measuring how strongly different materials absorb X-rays. Bone, soft tissue and other materials absorb different fractions of the beam, creating an image like a cast shadow. The problem is that soft tissues with similar densities can cast almost the same shadow – and become difficult to tell apart.
“In a normal X-ray source, you shoot electrons onto a piece of metal, and that generates light,” Birkedal says. “And if the conditions are right, some of that light will be X-ray light.” A synchrotron instead produces an exceptionally bright and coherent X-ray beam – properties that make it possible to use tiny shifts in the X-ray wave to distinguish soft tissues that would otherwise look almost alike.
“If we blast too much with the X-rays, we can induce damage,” Faaborg adds. Researchers can also make soft tissues easier to distinguish with stains or contrast agents, but these alter the sample. The mantis shrimp’s eye was instead imaged using its natural X-ray contrast, without added labels or staining.
Other imaging techniques, such as electron microscopy, which is designed to peer inside cells, are essentially too zoomed-in to give scientists a sense of how the structures are arranged. Doing so “would take forever, if it were at all possible,” says Birkedal.
The best of both worlds
Synchrotron-powered X-ray computed tomography can offer the best of both worlds. First constructed in the 1940s, synchrotrons are huge loops that use electric fields and magnets to accelerate particles to near-light speed. Particle accelerators also include machines such as the Large Hadron Collider in Switzerland, with scientists essentially playing bumper cars with subatomic particles to see what pieces pop off. A synchrotron used for imaging has a different purpose: steering rapidly moving electrons around the ring produces intense X-rays for experiments.
“In a synchrotron, the energy of the electrons is very high and they move almost at the speed of light,” Birkedal says. The crucial advantage for imaging is that the resulting X-rays can be made exceptionally intense, focused and predictable: “The X-rays come out basically right where you want.”
In 2016, MAX IV in Lund, Sweden became the first operational fourth-generation synchrotron. Its largest storage ring has a circumference of 528 metres and feeds X-rays into experimental stations called beamlines. Faaborg and her colleagues used one of these, DANMAX, to perform phase-contrast tomography.
“What is new about these synchrotron sources is that they have a specific property called spatial coherence,” Birkedal says. This enables researchers to detect not only how much X-ray light a tissue absorbs but also tiny shifts in the X-ray wave as it passes through materials with slightly different properties. The computer can then turn these shifts into visible contrast between structures that would otherwise look almost alike.
As a simple analogy, he points to what happens when visible light passes through water. “If you take a glass of water and you put a pencil in it, it looks like it is broken at the air–water interface,” Birkedal says. The effect is much more subtle when it is X-rays passing through shrimp eye tissue but enough for the detector and computer to pick up.
Seeing the whole eye – and what was hidden inside
Scanning the mantis shrimp’s eye at MAX IV took just under six hours across 40 positions, which pales in comparison to the “dozens of hours” other techniques could require, Birkedal says, “to the point you would not even do it”.
The result was a 3D floor plan of the mantis shrimp’s eyeball in remarkable detail, totalling 3 terabytes of data. For scale, that is enough to store the complete works of Shakespeare about 600,000 times over, or roughly 750 copies of Pixar’s Finding Nemo in high definition. The data files were so large that the researchers first had to reduce the resolution to view the entire eye on their laboratory computers. They could then return to selected areas in the original full-resolution data to inspect the smallest structures.
“We kind of knew how big the data would be, but we did not anticipate how challenging it would be to analyse,” Birkedal says. “In a way, we can see the forest and all the trees in the forest at the same time. We have everything, which makes finding something specific very difficult – say, if you are looking for just maple trees.”
However, Faaborg was rewarded for sifting through that mountain of information: she found a structure within the shrimp’s eye that had not been identified before, a network of vessel-like tubes weaving irregularly through the retina. They appeared in the midband, a specialised region involved in colour and polarised-light vision. What the tubes do is not yet known.
More than a picture
At the same time, the scans reproduced structures researchers already knew from decades of 2D microscopy, including the ommatidia and their coloured photo filters. In other words, structures they already knew were there appeared where they should.
“Looking at the full context of a system like that in 2D would be nearly impossible,” Faaborg says.
And the dataset did more than produce pictures. The researchers could measure the volumes of the eye’s photo filters, trace pores through its hard chitin cup and even pick out the repeating bands in individual muscle fibres. Those bands were about 8–11 micrometres apart, consistent with measurements previously reported for crustacean muscle.
The same trick – keeping the whole structure in view while moving down to microscopic detail – is already useful far beyond mantis shrimps. Synchrotron scans have, for example, been used to build AntScan, a library of full-body 3D images covering more than 2,000 ant species. Faaborg is also applying phase-contrast tomography to bone biopsies from people with X-linked hypophosphataemia, a rare form of rickets.
Both projects are examples of where synchrotron-based imaging really shines, the authors say: studying sensitive biological materials that scientists would rather not cut up, stain or otherwise alter.
“We have samples from living human beings who have donated a biopsy from their bones,” Faaborg explains. “That is a very precious sample.”
