Spectral analysis of Electroanatomical Maps for spatial bandwidth estimation as support to ablation

Intracardiac electrograms (EGM) in Cardiac Navigation Systems (CNS) and in Electrocardiographic Imaging provide relevant information on the arrhythmia mechanism for supporting ablation. In our previous work, we proposed Manifold Harmonics Analysis (MHA) for establishing the spatial sampling rate in ElectroAnatomical Maps (EAM) accounting for anatomical and bioelectrical features (e.g., voltage or activation time). Here, we propose a theoretically founded method for spectrum representation in terms of spatial frequencies from MHA, which can determine the minimum number of EGM registered at different spatial positions for accurate EAM with a cut-off spatial bandwidth. The EAM spectrum magnitude is obtained by cross-correlation between the original spatial anatomical and bioelectrical features, and the corresponding coefficients projected onto the manifold harmonic basis. The cut-off spatial frequency is computed according to a threshold value (TH ∈ [0,1]), accounting for the EAM reconstruction quality. TH was scrutinized in high quality anatomical meshes from tomography images, and in simulated and real EAM from CNS. Experiments showed that TH>0.98 is required to obtain accurate both anatomical meshes and EAM. Strong dependence was shown on EAM with the cut-off spatial frequency in terms of the arrhythmia mechanism.