AMplitude Spectrum Area (AMSA) is a metric derived from electrocardiogram (ECG) waveforms during ventricular fibrillation (VF). AMSA has been widely demonstrated as a good predictor of successful defibrillation and return of spontaneous circulation (ROSC) during cardiopulmonary resuscitation (CPR). Although several studies have raised concern that chest compression artifacts can interfere with AMSA calculation, there is no consensus on whether AMSA should be estimated from compression pause periods only. We hypothesize that AMSA can be affected by chest compression artifacts, but the influence of such artifacts does not change ROSC predictive value of AMSA. To test our hypothesis, a swine model of pediatric cardiac arrest (N=71; 1- to 2-month-old, 8-13 kg) was used. For each subject, a pair of 4-second ECG segments during VF were obtained, one during the chest compression pause period closest to the 10th minute of CPR and another during an adjacent period of compressions. AMSA was compared between compressions and pause periods, as well as between animals that achieved ROSC (n=46) and those that did not (No ROSC, n=25). AMSA during pause periods was higher than AMSA during compressions, both for ROSC (1.6 [1.1-2.0] vs. 1.9 [1.5-2.5]; P<0.01) and No ROSC animals (1.1 [0.9-1.4] vs. 1.4 [0.9-1.9]; P<0.01; median [25th-75th] percentiles). Nevertheless, AMSA for ROSC animals was higher than No ROSC animals both during pauses (1.6 [1.1-2.0] vs. 1.1 [0.9-1.4]; P=0.003) and compressions (1.9 [1.5-2.5] vs. 1.4 [0.9-1.9]; P=0.004). The area under the receiver operating characteristic curve (AUC) for discriminating ROSC and No ROSC subjects was the same for AMSA during pauses and compressions (AUC=0.71). Although AMSA values were affected by compression artifacts, the ability of AMSA to discriminate ROSC from No-ROSC animals did not change due to compressions. Our finding in this study supports future continuous monitoring of AMSA throughout CPR for ROSC prediction.