Figure 1
(a) Membrane potential snapshots, spaced by 143 ms. Regular fast pacing in the centre of a large (\(12\!\times {12}\,{\text{cm}}\)) slice of tissue creates regular waves radiating out toward the upper left corner. A spiral persists in the bottom right corner, and transient rotors appear in the lower left and upper right corners. (b) Tissue maps for the RQA measures from left to right, \(L_{max}\), RATIO, DET, and LAM. \(L_{max}\) takes on high values where activation is strictly regular, in the upper left corner and where rotor cores persist. RATIO highlights similar regions (via lower values), but identifies both rotor tips and the regular activation dynamics surrounding them. Measures DET and LAM highlight (via lower values) the regions where activation dynamics are most chaotic, characterised by wavelets that form where the main propagating waves interact. DET emerges as most appropriate for this purpose, as LAM also reaches low values around the rotor tip in the lower right corner. (c) Blurring the original maps via 2D convolution (see main text) makes the regions identified by the different RQA measures more visually distinguishable. In particular, the boundaries between regions of regular and fibrillatory activation detected by measure DET become much more pronounced.
