Objective: Early identification of infants with hypoxic-ischemic encephalopathy who have adverse outcomes despite neuroprotection with therapeutic hypothermia (TH) is urgently needed. Recent studies have found limited value of amplitude integrated EEG (aEEG) for predicting short-term outcomes in this population. Other quantitative electroencephalography (EEG) variables reflecting EEG amplitude, such as EEG power, could provide early stratification of a high-risk cohort in this population. The aim of the study was to evaluate and compare early EEG power and aEEG as predictors of magnetic resonance imaging (MRI) injury in neonatal hypoxic-ischemic encephalopathy. Study Design: We conducted a retrospective cohort analysis of 78 encephalopathic infants treated with TH between January 2009 and April 2013. About 56 infants had no/mild injury on MRI (group A), whereas 22 had moderate/severe MRI injury (group B). Total EEG power (TEP) and aEEG were obtained soon after initiation of hypothermia and then compared for their ability to predict future MRI injury. Results: TEP, calculated at a mean age of 8.9 h, was significantly higher in infants in group A as compared to group B (71.6±64.8 vs 26.9±65.3, P=0.02). Odds ratios for predicting moderate-severe MRI injury for TEP<10 μV2, TEP<20 μV2, burst Suppression or worse aEEG pattern were 55 (confidence interval (CI) 6.4 to 471), 12.5 (CI 3.8 to 40.7) and 6.7 (CI 2.0 to 19.8), respectively. Conclusion: Early TEP is a reliable predictor of moderate-severe MRI injury in encephalopathic infants undergoing TH and may enable early stratification of infants who may benefit from adjuvant therapeutic interventions.