Supplementary Material for: Comparison of Frequency- and Time-Domain Autoregulation and Vasoreactivity Indices in a Piglet Model of Hypoxia-Ischemia and Hypothermia

Introduction: The optimal method to detect impairments in cerebrovascular pressure autoregulation in neonates with hypoxic-ischemic encephalopathy (HIE) is unclear. Improving autoregulation monitoring methods would significantly advance neonatal neurocritical care. Methods: We tested several mathematical algorithms from the frequency and time domains in a piglet model of HIE, hypothermia, and hypotension. We used laser Doppler flowmetry and induced hypotension to delineate the gold standard lower limit of autoregulation (LLA). Receiver operating characteristics curve analyses were used to determine which indices could distinguish blood pressure above the LLA from that below the LLA in each piglet. Results: Phase calculation in the frequency band with maximum coherence, as well as the correlation between mean arterial pressure (MAP) and near-infrared spectroscopy relative total tissue hemoglobin (HbT) or regional oxygen saturation (rSO2), accurately discriminated functional from dysfunctional autoregulation. Neither hypoxia-ischemia nor hypothermia affected the accuracy of these indices. Coherence alone and gain had low diagnostic value relative to phase and correlation. Conclusion: Our findings indicate that phase shift is the most accurate component of autoregulation monitoring in the developing brain, and it can be measured using correlation or by calculating phase when coherence is maximal. Phase and correlation autoregulation indices from MAP and rSO2 and vasoreactivity indices from MAP and HbT are accurate metrics that are suitable for clinical HIE studies.