Non-invasive cardiac imaging based on just the standard 12-lead signals?

Recently, we presented progress in the development of an inverse procedure based on the equivalent double layer at the surface of the heart for non-invasive imaging of the activation sequence on this surface. The number of signals used was 64. In this paper we compared the activation sequences estimated from fewer signals from those obtained when using the 64 signals (the activation times estimated from the 64 signals are taken as the gold standard). The lead systems studied involved 32 (Lux), 23 (Barr), eight (standard 12-lead) or six independent (un-weighted Frank) signals. The inverse procedure proved to be remarkably robust with respect to this reduction in the number of leads, the 12-lead ECG yielding the poorest results. We attribute this robustness to the fact that the initial estimate of the solution was based on the general electrophysiology of activation propagation.