Mouse vs Zebrafish: Scientific face-off

If you’ve played Eyewire for awhile you may be familiar with our two datasets – e2198 and the zfish dataset (available to Mystic players). Our original dataset comes from the retina of Harold, an adult C57BL/6 mouse, and the zfish dataset comes from the hindbrain of a larval zebrafish (thus far unnamed, we’ll take suggestions if you have them!).

These animals are not unique to Eyewire, they are commonly used in many scientific studies. But why? Why not an octopus or a black bear? Well, let’s learn a little more about these creatures and their places in the scientific sphere!


Mice are the most commonly used of all research mammals! One of the main reasons mice are useful in scientific studies, is that they are surprisingly similar to humans. They are from the same clade as humans – the Euarchontoglires clade. They have similar physiology, anatomy, and genetics to humans, meaning that it is more likely that tests performed on lab mice will have similar outcomes in humans. After sequencing, the mouse genome was found to have many human homologues. 

Mice are also easy to house and maintain due to their small size. They reproduce quickly, and have relatively short lifespans, making it easy to study the effects of aging on them and to study multiple generations at a time. For all of these reasons, mice are excellent candidates for many scientific studies, and Eyewire is one of them!

Lab mice even have a statue dedicated to them in Siberia, Russia, which depicts a lab mouse knitting together 2 strands of DNA.

What have we discovered from our mouse dataset?

The Seung Lab has published multiple papers using research data taken from Harold’s retina, and traced by YOU, our great Eyewire community!

Our first paper was published in Nature in 2014, and answered the question “How does the mammalian retina detect motion?” using reconstructed Off-type starburst amacrine cells (SACs) and bipolar cells (BCs). Based on where different BC types were seen to be forming a synaptic connection with SAC dendrites, we were able to show how signals from different types of BCs were able to combine into a strong signal, due to the spacial orientation of the BCs and their known signal lag based on type.

If you want to learn more about this finding, check out the Nature paper here, and read a longer breakdown on the Eyewire blog here.

Eyewire’s second paper was published in Cell in 2018. This paper paper was based on Eyewire’s Museum, an interactive digital tool that displays and offers data analysis of 400 ganglion cells reconstructed by Eyewirers. Based on the visualizations in the Eyewire Museum, scientists found that:

  • The inner plexiform layer divides into four sublaminae defined by anatomical criteria
  • The aggregate neurite density of a ganglion cell type is approximately uniform
  • Inner marginal ganglion cells exhibit significantly more sustained visual responses

This digital museum allowed for a robust classification of ganglion cell types, including 6 new ones!

You can read the Cell paper here, read a less dense breakdown of the findings on our blog, and visit the Eyewire Museum here.


The zebrafish hasn’t been in the lab as long as the mouse has, but is has become the hot new research animal in recent years, with NIH grants for zebrafish projects outpacing those of fruit flies, C. elegans and frogs.

They were first used in the laboratory by American molecular biologist George Streisinger in the 1970s and 80s. Streisinger worked on zebrafish clones, which were among the earliest successful vertebrate clones created.

Zebrafish have multiple factors that make them good research animals. For example, they have similar reactions to humans and other mammals in toxicity testing, and have a similar sleep cycle to mammals. They also transparent during embryonic development, so their internal organs can be observed in living animals and without invasive procedures. Like the mouse, the zebrafish genome has also been fully sequenced.

And if you thought mice were short lived and good reproducers, then you haven’t met a zebrafish! An adult zebrafish can reproduce about every 10 days and lays between 50 and 300 eggs each time. Mice generally have up to 10 pups at a time, with only a few litters in a lifetime.

What have we discovered from our zebrafish dataset?

Like zebrafish as a research animal, our zfish dataset is much newer than our mouse dataset. However, we have still learned a lot so far, and a preprint paper from Seung Lab postdoc Ashwin Vishwanathan is currently out now!

In this paper, cells Eyewirers reconstructed from the larval zebrafish brainstem were analyzed. It was discovered that one group of neurons were responsible for eye movements, while another was responsible for body movements. It was also observed that eye movement neurons preferred to connect with other neurons in their group. Using theoretical models from the past and combining them with connectivity patterns from the reconstructions, accurate predictions of eye movement were able to be made.

You can read the full paper here. You can also read additional analysis from Ashwin on the Eyewire blog, here and here!

Choose your team!

Will you be team mouse or team zebrafish? Which little research hero will come out on top? One thing’s for sure – it will certainly be… for science! The fun starts at 11:00 AM ET on 6/7 and goes for 48 hours.


  • Starting Lineup – top 3 players on each team, who earn 75% of their score in bonus points
  • All Other Players – earn 50% of their score in bonus points
  • Winning Team – 20,000 additional points
  • Each Team’s MVP – 5,000 additional points

The winning team is determined by average points per player, with 2x weight given to Starting Lineup players. To qualify for any of the above bonuses or affect the team score, players must earn a minimum baseline of 2,000 points.

Images from Amy Sterling and Ashwin Vishwanathan