A new version of the Blankenship2011 was uploaded to github. In the first version (now in the folder /hdf5-old), the spike times were exported in ASCII using a fixed number of digits. This meant that as the time increased, the number of digits after the decimal point decreased, losing precision. This has now been fixed.

Whilst checking these files, we also found several issues with the Cx36/45 double knockout data.

  1. The Cx36/45 files were identical (samples 3 and 6). A revised sample 3 is now uploaded.
  2. The old sample 5 is now named sample 4 in the new set.
  3. Sample 5 is new, and the old sample 4 gone.

Thanks to Malte Kaehne for reporting this problem with the original data.


Paper released


Maccione2014 dataset added.


Kirkby2013 dataset added.


Initial version of news file. Blankenship2011 dataset included.