Transcranial Doppler ultrasound is a noninvasive technique for monitoring the velocity of blood flow in the main intracranial arteries, particularly those in the circle of Willis. Our aim was to assess whether changes in cerebral arterial blood flow in pediatric patients under sevoflurane anesthesia demonstrated by pulsed Doppler ultrasound correlate with changes in the bispectral (BIS) index and electroencephalographic state and response entropy (ES and ER, respectively). Prospective, blinded observational study of 36 pediatric patients (age range, 5 to 11 years; ASA physical status classification, 1-2) under sevoflurane anesthesia for minor surgical procedures. Anesthesia was induced with sevoflurane and maintained with 2.5% sevoflurane in an inspired oxygen fraction of 50% in air. A continuous perfusion of remifentanil was provided for analgesia. In all patients we monitored hemodynamic and respiratory patterns, gases, temperature, and hypnosis (BIS, ES and ER) as well as cerebral blood flow estimated by pulsed Doppler ultrasound in the middle cerebral artery. The resistance index, pulsatility index, mean velocity, and estimated baseline cerebral blood flow were calculated from the Doppler sonogram. Correlations (Pearson's r) were calculated between BIS, ES, ER, the pulsatility index, resistance index, mean flow velocity, estimated cerebral blood flow, fraction of end-tidal carbon dioxide, and temperature. A regression model was constructed. Induction caused a pattern of high velocity (elevated mean velocity and normal or reduced pulsatility index) until the lowest BIS and ES values of 31 and 29, respectively, were reached. During maintenance, the Doppler sonogram pattern was slower (normalization of the pulsatility index, the resistance index, and mean velocity). Changes in flow and absolute entropy and BIS values were statistically correlated (Pearson's r values > or = 0.91); there was 95.6% agreement between Doppler values and BIS and agreement between BIS and ES values of 35 to 45. On awakening, flow velocities approached baseline values when BIS and ES rose to between 90 and 98. The estimated cerebral blood flow underwent fluctuations coinciding with an approximately concomitant increase or decrease in BIS (r > 0.95); the BIS response occurred with a slight delay of no more than a minute. The entropy measurements did not reflect the fluctuations. We show Doppler ultrasound patterns during anesthetic induction with sevoflurane. Systems for monitoring hypnosis could be considered indirect measurements of cerebral blood flow; BIS measurements are more sensitive to change. Transcranial Doppler ultrasound facilitates the observation of changes in blood flow that occur at different levels of hypnosis during anesthesia.