[Background]: Maintenance of cerebral blood flow during cardiopulmonary resuscitation (CPR) is a critical factor affecting the neurological outcome of patients after cardiac arrest. However, dynamic indicator of cerebral blood flow during CPR is not established. Near infrared spectroscopy (NIRS) is a non-invasive technique capable of detecting changes in cerebral blood volume continuously. The aim of this study is to compare cerebral perfusion during CPR obtained by continuous measure of carotid blow flow (CBF) using flow sensor with continuous cerebral perfusion indices (CPI) assessment by NIRS in a porcine model of ventricular fibrillation (VF) cardiac arrest. [Methods]: CBF during CPR were measured in 21 swine (25± 3 kg) using extra-vascular flow probes positioned around the common carotid arteries. The optodes of the NIRS were attached to the skull bone. Following baseline measurements, VF was electrically induced. After the interval of untreated VF, CPR was continued for 25 min, during which mechanical and manual CPR was alternated at every 5 min intervals. CBF and CPI were measured continuously during CPR. The mean volume of CBF and CPI was calculated average 10 beats in each phase. Correlations between CBF volume and CPI were analyzed by Spearman's rank correlation. [Results]: The followability of CBF and CPI during CPR was good. The Spearman's rank correlation coefficient between CBF volume and CPI was different by sections and it was suggested that the CBF volume was too small to assess the correlation in many sections. If the sections with CBF volume>0.4 ml were extracted and reanalyzed, the rate of sections correlated between CBF volume and CPI were increased (Spearman’s correlation coefficient≧0.4 was 41.3%, Spearman’s correlation coefficient≧0.7 was 17.5%).(Table) [Conclusions]: It was suggested that CPI by NIRS can reflect the CBF during CPR in the sections where enough CBF is.