IE 11 is not supported. For an optimal experience visit our site on another browser.

EyeWire Video Gamers Help Untangle the Retina's Space-Time Secrets

Thanks to thousands of video gamers, researchers have traced how cells in our retina track objects that move across our field of vision.
Get more newsLiveon

Researchers have traced how cells in our retina track objects that move across our field of vision, thanks in part to thousands of video gamers.

The findings, published online Sunday by the journal Nature, validate a concept that explains how some nerve cells are stimulated only by motion in a specific direction and not in other directions. They also validate the use of video games, crowdsourcing and the other tools of citizen science for making rigorous scientific discoveries.

"You no longer have to have a Ph.D. in neuroscience," said Amy Robinson, creative director for the EyeWire neuro-gaming venture, which contributed to the study. "You could be a high-school student, or a sculptor, a dental assistant or retiree. All you have to have is now is an Internet connection and an interest in gaming."

More than 120,000 EyeWire gamers ("EyeWirers") from more than 100 countries have signed up to play online video games in which they trace 3-D representations of neural wiring. They rack up points by tracing the circuitry quickly and accurately. It's not always easy: Expert users, known collectively as the Grim Reaper, occasionally have to go into the game to correct the neural maps.

The project to trace retinal neurons known as starburst amacrine cells was particularly challenging. "We weren't sure it would work, because starburst cells are extremely tricky to map," Robinson told NBC News.

EyeWire's team of experts set up a special "Starburst Challenge" that identified the most adept gamers for the job. More than 2,000 of them contributed to the mapping project and are listed as co-authors of the Nature paper.

Image: Retinal cells
Assortments of cells in the retina are traced in different colors in this EyeWire 3-D reconstruction.Alex Norton / EyeWire / MIT

Mapping the retina

The amateur EyeWirers and professional researchers collaborated to map the detailed connections between the wiring for the starburst cells and other types of cells in a mouse retina.

Those cells include photoreceptors, which are activated when light hits the retina; different kinds of bipolar cells, which register the electrical signals from the photoreceptors and relay them to the starburst cells; and ganglion cells, which take in signals from the starburst cells and send them along to the brain.

The starburst cells play a vital role in motion perception. They're called starbursts because tiny neural wires, or dendrites, burst out in many directions like spokes on a wheel. When an object moves in a direction in alignment with one of those dendrites, heading outward from the center of the starburst cell, a signal is passed along via the ganglion to the brain. But when the motion is in a different direction, there's no signal.

This phenomenon, known as direction selectivity, was discovered in 1964.

"Although half a century has passed since the original description of direction selectivity in the retina, a satisfactory explanation of this phenomenon has been difficult to achieve," Sebastian Seung, a Princeton neuroscientist and EyeWire's founder, said in a news release about the latest study.

Image: Starburst cell and bipolar cell
An EyeWire 3-D reconstruction shows a starburst amacrine cell in yellow-green and a single bipolar cell in blue.Alex Norton / EyeWire / MIT

How our motion detector works

The study lays out a coherent explanation for the retina's motion detector. Thanks to the detailed maps, the researchers could see a clear pattern in the arrangement of the various nerve cells. One type of bipolar cell, BC2, tended to be placed near the hub of a starburst cell, on each spoke-like dendrite. A different type, known as BC3a, was farther out on the dendrite.

From previous studies, the researchers knew that BC2 cells have a significant built-in time delay: They fire about 50 milliseconds after they're activated. The BC3a cells have a shorter time delay. A light stimulus moving across the retina would activate many bipolar cells, firing at different times. But the signal would be strongest only if the stimulus activated a BC2 cell first — and then activated a BC3a cell a fraction of a second later, farther out on the same dendrite.

On that dendrite, the two off-sync bipolar cells would end up firing a signal at the same time. That double-jolt signal would activate the relay to the ganglion cell, basically telling the brain that the object is moving in a direction determined by the orientation of the strongly firing dendrite. Single-jolt signals wouldn't register, however. That's how the retina's motion detector can send such a specific signal to the brain.

This isn't the only way our brain registers motion, but the system appears to have deep evolutionary roots. "Perhaps the retina evolved this capability because perceiving motion is so crucial for survival, as in the examples of a person jumping away from an oncoming car, or a frog swatting at a buzzing fly," Seung said.

Image: Wiring diagram
This schematic shows how different cells in the retina are "wired" together. Light stimulates the photoreceptor cells (PC), which transmit electrical signals to different types of bipolar cells (BC2 and BC3a). The BC2 cell has a built-in time delay for relaying a signal (indicated by the Greek letter "tau"). The BC3a cell doesn't have as much of a time delay. Both cells relay signals to the starburst amacrine cell (SAC). When the two types of cells fire at the same time, the signal is strong enough to set off a relay signal in the SAC that is picked up by a ganglion cell (GC). The ganglion cell then alerts the brain that motion has been detected in a specific direction.EyeWire

How the insights can help

Robinson said the insights into the retina's "space-time wiring" could give neuroscientists a better understanding of the process by which neural stimulation is turned into perception. Eventually, such research could also help engineers design better electronic systems for tracking motion.

On another level, the study demonstrates how games can combine the wisdom of crowds with the wisdom of experts to produce a whole that's greater than the sum of its parts.

"We believe that crowd wisdom requires amplifying the expert voices within the crowd, and also empowering individuals to become experts," the researchers wrote. "Fortunately, such goals are well-matched to the game format."

Robinson said EyeWire is already working on a more sophisticated game that would map the olfactory cortex's connections to other parts of the brain, with the aim of tracing the linkages between specific smells and the emotional responses produced by those smells. Such a project will require faster computers as well as better ways to organize citizen scientists, she said.

"We need order-of-magnitude improvements in the rate at which we reconstruct cells," Robinson said.

In addition to Seung and Robinson, the authors of "Space-Time Wiring Specificity Supports Direction Selectivity in the Retina" include Jinseop Kim, Matthew Greene, Aleksandar Zlateski, Kisuk Lee, Mark Richardson, Srinivas Turaga, Michael Purcaro, Matthew Balkam, Bardia Behabadi, Michael Campos, Winfried Denk and the EyeWirers.