For the last month we have been running our first team competition on Stall Catchers, to celebrate the premiere of The Crowd & The Cloud documentary on citizen science, which features Stall Catchers and EyesOnALZ, and to kick-off our brand new Teams (beta) feature!
We hope you all enjoyed the competition so far - but some of you might be wondering what's next? Don't worry - the end of #CrushALZ is not the end!
Regular challenges & competitions
The teams feature will of course stay up and running on Stall Catchers, with daily, weekly and all time leaderboards spotlighting the leaders!
We also have a number of research challenges and competitions planned for the near future, where we will ask teams & individual catchers to sprint for a specific research goal, like we did now during #CrushALZ. These will be announced on the blog, the forum & the Stall Catchers game interface!
And during the "calm" periods, we will still post regular updates on the most active teams, so don't forget to keep an eye on the blog!
More research datasets
We are currently about half way through the research dataset that was uploaded to Stall Catchers specially for the #CrushALZ competition.
It has been our largest dataset by far, and even though it doesn't look like we will finish it in time for the competition finale, we have already done an amazing amount of work!!
After the competition ends, we will of course continue to analyse this dataset until we get to 100%. This will help researchers determine whether stalls occur closer to amyloid plaques than flowing vessels in the brain in Alzheimer's.
When the dataset is complete, we will tackle the next treatment-centered research question with data provided to us by the Schaffer-Nishimura Laboratory, Cornell University. We have been developing a data pipeline to support this, which will allow us to add new data into Stall Catchers after an old dataset is finished. That way, as long as there are unresolved questions about reduced blood flow in the brain (and enough imaging power at the Schaffer-Nishimura Laboratory!) we are unlikely to run out of vessels to analyse!
Dynamic consensus methods
We have already hinted on a couple of occasions that we are cooking up something new in the way we determine crowd answers ... 🙃
As you know, the blue sensitivity bar (next to the vessel movie) measures your sensitivity based on how well you identify stalled & flowing vessels. We are currently working on a method to assign weight to assign weight to each catcher's response based on their sensitivity, which basically means we might need a lot fewer annotations per vessel than we do now!
Based on our validation study, we are collecting 20 answers per vessel to be sure. But using the new dynamic methods might allow us to reduce this number to just a few vessels, which means we could be analysing each dataset a lot faster!
This is our highest research priority at the moment, and the dynamic methods might bring some ground-breaking changes to Stall Catchers very soon!
Redeem feature becomes active!
Yes - I know what you are thinking - the "Redeem" button has been teasing us all for a while with the "No points yet!" note and a million pending movies, right?!
Well, this is about to change.
For the redeem to work there firstly to be enough new movies in Stall Catchers - we had that uploaded for #CrushALZ, CHECK! And secondly, a way to automatically validate them & determine the crowd answer (for now that means our systems need to count 20 annotations per vessel & generate a consensus answer from them).
We have been actively developing the latter, as up till now the only way to validate movies was manual. The updated redeem feature is coming soon, and you will be able to redeem all your waiting points as early as next week!
You are wondering how many thousands and millions you have accumulated by now, right? 🤑
Keep giving us feedback!
We have received a lot of helpful feedback from you during the period of #CrushALZ & have been making changes and developing features accordingly!