Table 1 Detailed description of boxplots showing the distribution of analyzed measures in Figs. 6 and 7.

\(\textrm{RATIO}\): This measure reaches a low value for anchored rotation. The measure is also able to distinguish a traveling rotor from fibrillation, including its combination with anchored rotation (see Fig. 6 and Supplementary Table S2). The spatial correlation reaches lower values for anchored rotation (see Fig. 7 and Supplementary Table S3). In the traveling rotor, RATIO reaches high values in regions between rotor meandering (see Fig. 5)

\(\textrm{DET}\): In Fig. 6 and Supplementary Table S2 it can be seen that the regular rotational movement of the AP around the scar tissue reaches high values of DET. On the contrary, chaotic fibrillation results in a reduction in its value (but these differences are very small). This situation is also reflected in the spatial correlation of this measure (see Fig. 7 and Supplementary Table S3). This phenomenon can be also noticed in Fig. 5. In this figure an even distribution of DET in the map for AP anchored rotation propagation is shown as opposed to other propagation types

\(\textrm{DIV}\): Anchored rotation reaches a reduced average value. The spatial correlation of this measure cannot distinguish between different types of AP propagation

\(\mathrm {L_{max}}\), \(\mathrm {L_{mean}}\), and \(\textrm{ENTR}\): regular movement of the AP around the scar tissue results in a slightly increased value of these measures for anchored rotation. The spatial correlation of this measure cannot distinguish between different types of AP propagation

\(\textrm{LAM}\): the average value of LAM is increased not only for anchored rotation but also for the travelling rotor. As a result, these two types of APs can be distinguished from propagation where fibrillation is present. Spatial correlation is comparable for all types of APs. RQA maps of individual AP propagations are very similar to the measure RATIO, the region around which the traveling rotor meanders is highlighted (see Fig. 5)

\(\mathrm {V_{max}}\), and \(\mathrm {V_{mean}}\): anchored rotation can be very well distinguished by both the mean value (see Fig. 6 and Supplementary Table S2) and the spatial correlation (see Fig. 7 and Supplementary Table S3). Other types of AP propagation cannot be distinguished from each other

\(\textrm{OI1}\), and \(\textrm{OI2}\): using the mean value, the traveling rotor and anchored rotation can be well distinguished (see Fig. 6 and Supplementary Table S2). Regular rotation results in a large amount of power placed in the first two highest peaks in the frequency spectrum. This phenomenon is manifested in high OI1 and OI2 anchored rotation. In contrast, the very low value of these measures for the traveling rotor indicates irregular AP propagation. Due to the analysis of the short time series of the modeled AP (1 s), this irregularity is not observed. Spatial correlation of these measures indicates the ability to distinguish between all analyzed types of AP propagation (see Fig. 7 and Supplementary Table S3). These differences are relatively small

\(\textrm{eigen1}\), and \(\textrm{eigen2}\): the mean value of the first and second highest eigenvalue of the recurrence plot for anchored rotation (\(<40\)) achieves the highest discrimination from other AP types (\(>10\)) among all measures tested. Other types of AP propagation can also be distinguished using this measure. However, the differences in the mean values of eigen1 and eigen2 between these types of propagation are minor (approximately 1). Spatial correlation of these measures provides distinction only for anchored rotation measured at eigen1

\(\textrm{ApEnt}\), and \(\textrm{SampEnt}\): the mean approximate entropy calculation is able to distinguish all analyzed types of AP propagation, except the traveling rotor. However, this type of AP propagation can be distinguished by SampEnt. SampEnt and ApEnt show the opposite direction of entropy increase. SampEnt reaches the lowest level for anchored rotation and the highest for travelling rotor, while for ApEnt, this trend is reversed (see boxplots in Fig. 6). These boxplots suggest that SampEnt is more appropriate for this data type. It can be assumed that regular movement of the AP around the scar should achieve less complexity than irregular fibrillation or a more complex rotor meandering. This assumption is also supported by the work of Montesinos et al.⁵⁴ In this paper, the authors examined the time series of centres of the pressure during a posturography test of different groups of adults. They found that SampEnt is more appropriate to distinguish these groups. Spatial correlation of any entropy is the same for all types of APs