Atrial activity extraction in Holter registers using adaptive wavelet analysis

Extraction of atrial activity (AA) is quite important in the study of different atrial arrhythmias. This work shows the possibility of AA extraction from atrial fibrillation (AF) episodes in Holter registers using only one lead with a new technique, the adaptive wavelet analysis (AWA). The principal aim is to adapt automatically the discrete packet wavelet transform (DPWT) depending on the shape of the signal considered in each moment. The more suitable wavelet functions are selected to obtain several ideal wavelet forms. Every register is divided in different blocks and a DPWT is applied using the corresponding wavelet in each block. So, the analysis is adapted to the behaviours and properties of the lead under analysis. After the mentioned process, the AA is reconstructed using the best coefficients of the obtained wavelet decomposition. The AWA should be applicable in arrhythmia detection and analysis, like paroxismal atrial fibrillation, which have to be usually detected from Holter systems.