Challenge Results

These are all the challenges results that have been posted on Mozak.

  • Date: May 24, 2017 UTC
  • Consensus length: 54661 (69%) Gold Standard length: 79597 This cell has great dendrite quality, but very weak axon in many areas. The consensus dendrites look nearly perfect, but the consensus axon failed to capture the low intensity axon. There was very little false positive, which was helpful.
  • Length of community consensus: 6766μm
  • Distance contributed to consensus: 52085μm
  • Top contributors:
  • UserContribution
    Length
    Percentage
    cell_porker_856766μm100.0%
    glol6270μm92.7%
    Gil6150μm90.9%
    Oniwakamaru5213μm77.0%
    gdnskye4871μm72.0%
  • Date: May 24, 2017 UTC
  • Consensus length: 44665 (82%) Gold Standard length: 54700 This is a moderately challenging cell due to quality. However, the consensus was very accurate with dendrites, missing only two short branches. The consensus captured most of the more obvious axon. However, most of the deeper, fainter axon was missed. Overall, it was a useful starting point. The main negatives were the false positive signal scattered around the soma, which was took a long time to separate from the real signal, and the tiny breaks in long straight segments.
  • Length of community consensus: 5699μm
  • Distance contributed to consensus: 39782μm
  • Top contributors:
  • UserContribution
    Length
    Percentage
    Rock On5699μm100.0%
    glol5182μm90.9%
    Gil4910μm86.2%
    BendBob4801μm84.2%
    Oniwakamaru4010μm70.4%
  • Date: May 24, 2017 UTC
  • Consensus length: 73911 (103%) Gold Standard length: 71540 This is a nice, high quality cell. The consensus looks great in the dendrites. There is a little bit of missed axon, but not very much. The main lesson that can be taken from this comparison is in regards to the false positive tracing. You can see a green cluster near the top of the image and a few green streaks near the bottom left corner that were not drawn in the gold standard. These are places where more than one Mozak player traced something, but the thing that they traced was not part of the target cell. The cluster of signal near the top of the image is where players mistook the edge of the brain (and some debris) for real signal. The streaks near the bottom of the image are actual branches, but they are part of a different filled cell. In order to maximize the use of each slice of tissue, multiple cells are patched. Occasionally, two of these cells are close enough together that the branches from one will be present in the image of another. In order to avoid tracing false positive signal, the rule that players should follow is to avoid tracing anything that cannot be connected back to some part of the target cell.
  • Date: May 24, 2017 UTC
  • Consensus length: 88334 (114%) Gold Standard length: 77557 The challenge for this cell was the leakage. You can see in the image below that the biocytin leakage is very high. In the consensus there were a few missed branches in the soma area where the leakage was the worst. However, I also found a missed branch in the gold standard that was captured in the consensus (well done citizens!). There was essentially no axon to be found in this image, so this is a dendrite-only reconstruction. The consensus had surprisingly little false-positive signal, which is awesome given the severity of the leakage. Great work!

IMPORTANT!

The browser you are currently using may not be compatible. Please use the most recent Chrome or Firefox for the best experience.