Eyewire


Eyewire is a game to map the brain from Sebastian Seung's Lab at Princeton University. This citizen science human-based computation game challenges players to map retinal neurons. Eyewire launched on December 10, 2012. Over five years, 250,000 people from 150 countries have signed up. The game utilizes data generated by the Max Planck Institute for Medical Research.
Eyewire gameplay advances neuroscience by enabling the reconstruction of morphological neuron data, which helps researchers model information processing circuits. Anyone with an internet connection can participate by solving 2D puzzles to fit together segmentation produced by artificial intelligence used to map the connectome. In Eyewire, players reconstruct 3D models of neurons from electron microscope images by solving puzzles. Playing requires no scientific background.

Gameplay

Eyewire challenges players, "Eyewirers", to map neurons in 3D. Upon registering, players are directed through a tutorial that explains the game. Supplementary are available on the Eyewire Blog.
In Eyewire, the player is given a cube with a partially reconstructed neuron branch stretching through it. On the right side of the screen is a grayscale image of the cross sections of neurons. The player learns to "color" inside a gray outline of a single neuron branch, which usually extends from one side of the cube to another. As a player colors, segmentations that were generated by AI are added to the 3D section on the left of the screen. Reconstructions are compared across players as each cube is submitted, yielding a consensus reconstruction that is later checked by expert players of rank Scout and Scythe. These players have the power to extend branches, remove erroneous segments, and flag cubes for further review. This end result is volumetric reconstructions of complete neurons.

Scoring

Each volume is presented to two to five different players. Generally, the trace chosen by the majority of accurate players is accepted. Players win points based on whether their tracing matches the majority of other players' tracings, time spent on the cube, and the new amount of neural volume found.

Goal

The goal of Eyewire is to identify and classify specific cell types as well as potentially expand the known broad classes of retinal cells. Eyewire aims to advance the use of artificial intelligence in neuronal reconstruction. The project aims to help determine how mammals see directional motion.
Over 1,000 neurons mapped by Eyewirers may be explored on the Eyewire Museum, a browser-based visualization tool that pairs anatomical and functional data.

Methods

The activity of each neuron in a 350×300×60 μm3 portion of a retina was determined by two-photon microscopy. Using serial block-face scanning electron microscopy, the same volume was stained to bring out the contrast of the plasma membranes, sliced into layers by a microtome, and imaged using an electron microscope.
A neuron is selected by the researchers. The program chooses a cubic volume associated with that neuron for the player, along with an artificial intelligence's best guess for tracing the neuron through the two-dimensional images.

Publications

Accomplishments

Eyewire has been featured by Wired, Nature blog SpotOn, Forbes, Scientific American, NPR and more.