I am curious to know in which part of the analyzed datasets the heart failure information was hidden. I would have expected that they reanalyzed heart time traces but it was not mentioned explicitly, was it?