Scientists have learned a lot about the brain by measuring it from non-invasive imaging, like MRI. But as we collect larger datasets, our measurements become less accurate. Current computer algorithms are able to make measurements quickly, but are not very accurate. Expert neuroanatomists are accurate, but it takes them a long time. The goal of the Medulina project is to train citizen scientists in neuroanatomy, and combine their measurements with computer algorithms to get the most accurate measurements of the brain in a short amount of time.
Medulina is an open source game to crowdsource brain image segmentation. Its still in beta, but if you want to try it out, you just need a GitHub account. Currently, you can try to identify meningiomas (a type of brain tumor), stroke lesions , and the dentate gyrus , a structure in the brain that is important for forming new memories.
Medulina was started by Anisha Keshavan while procrastinating on writing her PhD disseration. Dylan Nielson and Adam Thomas became collaborators when they all met at the 2017 OHBM Hackathon.
What hardware do I need to play?
Medulina works on a computer, tablet and even a cell phone (but you should only play if you have a large screen, and you are connected to wifi/have unlimited data).
Can I crowdsource my data on Medulina?
We are actively developing Medulina, so it is not "live" yet. If you are interested in using Medulina for your data, please post on our Github Issues Page or email anishakeshavan at gmail dot com .
I found a bug! What do I do?
Please post on our Github Issues Page !
Want to join the team?
Join us on GitHub and on the #medulina brainhack slack channel!