MIT brain game leads to neural circuit, scientific paper

Image by Alex Norton, SebastianSeung Lab, MIT

Retinal Ganglion neurons mapped by gamers in EyeWire.

By Carolyn Y. Johnson Globe Staff 

A team of researchers from the Massachusetts Institute of Technology and more than 2,000 coauthors known only by online handles such as “crazyman4865,” “magret,” and “Riley_Light” on Sunday reported a new map of a neural circuit in the eye that helps detect movement.

The paper published in the journal Nature is the first report from EyeWire, a citizen science computer game created by Sebastian Seung, a neuroscientist who recently moved from MIT to Princeton University.


The goal of the game is simple: players (called “EyeWirers”) use their computer mouse to trace brain cells in black-and-white images taken of slices of a mouse retina. Their goal is to trace the path of a neuron.

EyeWire is just the latest example of a broader attempt to harness the talent, interest, and idle moments of regular people with online games that push science forward. A game that challenges players to make accurate protein structures, called Foldit, has helped contribute to a new computer algorithm to predict protein structures that outperformed previous methods. In a game called EteRNA, players create folding molecules of RNA, a multipurpose molecule that regulates DNA, with very real-world stakes: the most promising molecules are chosen and synthesized in the laboratory to be studied in real life.

It can be competitive and exciting. For example, a user called “dainesl”, chatting with other players in EyeWire, wrote, “I never thought neurology would be this intense. I thought wrong apparently.”

EyeWire’s first scientific output is a detailed map of a neural circuit in the eye that shows precisely how a network of cells is connected. That knowledge, combined with information about how rapidly those cells respond to visual objects, led the researchers to put forward a possible explanation, in which a choreographed set of neural signals occurs, allowing the eye to detect motion in particular directions.

Usually, the part of a research paper that describes how the experiment was done -- called the methods section -- is rather dry. The description of the methods is essential; in fact, it is the guts of the paper, where all the details of the computer programs, microscopy techniques, genetic markers, and statistical methods are laid out for other scientists to follow. But it can read like a jargon-filled acronym scramble for those who aren’t experts in the field.


That can change, however, when the methods describe corralling the collective efforts of thousands of users in a productive way. That’s because not only are researchers harnessing the game-playing prowess of lots of strangers whom they’ve never met; they have to make sure the research is good.

In order to do that in EyeWire, researchers utilized an expert “GrimReaper,” who could overrule the other players when they may have made a mistake. There are also elite players called scouts and scythes, who check the work and flag problem areas.

The tricky thing about creating these games is to make them “sticky” -- to make sure that people don’t just sign up once and abandon the game because it’s confusing or unrewarding. That’s a challenge that all games grapple with, but that Seung and colleagues hope they’ve solved, since 120,000 players have now registered. The players compete with one another and are ranked, and can chat with one another as they play.

David Condon, the editorial director of NOVA Labs, which creates online educational resources and games, said that challenge is why they have been making efforts to partner with scientists. The hope is NOVA Labs can use their storytelling and communication skills to help make real scientific problems even more fun and game-like.

Ancient DNA swaths inspire free spirit of experimentation

At hundreds of spots in our DNA, there are ancient swaths that have remained puzzlingly unchanged over hundreds of millions of years of evolution. No one knows exactly what to make of these regions of DNA, called ultraconserved elements -- they don’t appear to serve essential functions, so why are they preserved?

“They are considered one of the most mysterious aspects of the genome,” Ting Wu, a Harvard Medical School geneticist, said at a talk at a genomics conference Wednesday in Cambridge.


But Wu has a provocative idea about these ultraconserved elements: Perhaps they are a natural defense system against harmful changes to our DNA. And perhaps there could be a way to harness this mechanism as a therapy, triggering it to cull cells that carry harmful genome rearrangements, before there is enough of a problem that a disease is even diagnosed.

Wu presented her talk, which she cautioned was speculative and a bit fanciful, at the “Explorers” session of the Genomes Environments Traits Conference. The annual meeting brings together volunteers in the Personal Genome Project, an experiment run by Harvard geneticist George Church in which participants have their genomes sequenced and share the data with the world to advance science.

At the meeting, a packed room of 100 participants learned about the latest in genome science and could take part in studies. For example, a box outside the session was designated for participants to deposit armpit swabs as part of one study.

Wu’s playful, daring, idea-generating presentation fed off the experimental mood of the crowd, providing a glimpse of a side of science that rarely is visible when results are presented. Scientific talks are usually designed as a straightforward presentation of evidence in support of one idea or another. The occasional joke is woven in, but the science often seems “done” by the time the presentation is made. Wu’s talk presented ideas still being formed and she freely admitted that the evidence is being accrued and may well turn out to debunk her own idea.

The idea that ultraconserved elements play a role in protecting the body against harm is based on the fact that people inherit two strands of DNA, one from each parent. When those strands are lined up next to each other, it would be obvious if there are aberrations in the ultraconserved areas -- a buckle would form in the strand. Wu thinks that when ultra conserved elements are working, the pairing may be a protective mechanism -- if the DNA goes awry at those spots, the cell dies.

So far, the idea that this is a process important for health is supported mainly by circumstantial evidence, but it may also explain what goes wrong in disease. For example, Wu said that her research has found that some types of cancer appear to stem from genome rearrangements found near ultra conserved elements, suggesting that preserving these stretches intact is important for health. Now, Wu’s lab is working on various approaches to understand and rigorously test the ideas.

She isn’t holding her breath to see her ideas confirmed: “We are fully prepared to be incorrect,” Wu said. Should that happen, the science may lead to something even more interesting.

Carolyn Y Johnson can be reached at
Follow her on Twitter @carolynyjohnson.