I am reworking the analysis of ECGs from an old contest, and I want help from an expert.

In 2000 the Computers in Cardiology challenge (CINC2000) provided ECGs from sleep studies of 70 patients and asked contestants to identify obstructive sleep apnea based on those ECGs. I was on the team that won.

Now I am reworking the problem for the second edition of a book (see https://www.fraserphysics.com/book.pdf), and I see great variety in those ECGs (see https://www.fraserphysics.com/all_ecgs.pdf). I suspect that some of that variety is due to lead placement, and some is due to pathology, but I'm not sure.

Is anyone here willing to help me out?