“…The ECG analysis concept in this study brings a new perspective to the state of the art. The algorithm monitors the rhythm during continuous CPR, not limited to analysis only during CC [14,16,17,20,22,24,25,27,[29][30][31][35][36][37]51,52] or only during clean ECG [17,31,39,41,49,53,54], as summarized in Figure 10. Although most published algorithms have been optimized and report the performance for either CPR-ECG or fully Clean-ECG signal parts during OHCA, the shock advisory performance of our algorithm CNN-CPR (10 s) is in the high range for both signal parts: CPR-ECG (Se = 92-94.4%, Sp = 92.2-99.5%) for decision time in the range [−15; 0 s]; fully Clean-ECG (Se = 98.7%, Sp = 98.9-100%) for decision time equal to 10 s. The difficult task for CPR-ECG analysis was managed here without additional processing (ECG prefiltering, and additional sensors), which can be found in some studies [14,16,17,20,36,37,51].…”