What is the central question of this study? What is the role of carbon dioxide in the cerebral blood flow (CBF) response to the cold pressor test (CPT)? What is the main finding and its importance? The CBF response was elevated during the isocapnic (controlled CO ) CPT in the middle cerebral artery and the internal carotid artery compared with the poikilocapnic (uncontrolled CO ) CPT, owing to ventilation-associated reductions in end-tidal CO . Furthermore, the common carotid artery vasodilated to a greater extent during the isocapnic compared with the poikilocapnic CPT, whereas the internal carotid artery vasoconstricted during both CPTs. Our data highlight the importance of CO control when investigating the CBF response to the CPT. In addition to increasing sympathetic nervous activity, blood pressure and cerebral blood flow (CBF), the cold pressor test (CPT) stimulates pain receptors, which may increase ventilation above metabolic demand; this response is likely to reduce the partial pressure of end-tidal carbon dioxide (P ET ,CO2) and will attenuate elevations in CBF. Our hypotheses were as follows: (i) the CPT will elicit hyperventilation, effectively lowering P ET ,CO2; (ii) the CBF response will be elevated during an isocapnic (controlled P ET ,CO2) compared with a poikilocapnic CPT (uncontrolled P ET ,CO2); and (iii) in response to the CPT, the common carotid artery (CCA) will vasodilate, while the internal carotid artery (ICA) will remain unchanged to help regulate CBF. Using a new, randomized experimental design, we measured the cerebrovascular response in the middle cerebral artery (MCA), CCA and internal carotid artery (ICA), during an isocapnic and poikilocapnic CPT in 15 participants. Blood pressure and cardiac output (finger photoplethysmography), heart rate (ECG), MCA mean velocity (transcranial Doppler ultrasound) and CCA and ICA CBF (Duplex ultrasound) were recorded during both CPT trials. Our findings were as follows: (i) ventilation increased, which reduced P ET ,CO2 (-5.3 ± 6.4 mmHg) during the poikilocapnic compared with the isocapnic CPT; (ii) the CBF response was elevated during the isocapnic compared with the poikilocapnic CPT in the MCA and ICA, but not in the CCA; and (iii) the CCA dilated to a greater extent during the isocapnic compared with the poikilocapnic CPT, and the ICA vasoconstricted during both trials. Our data emphasize the importance of P ET ,CO2 control in the CBF response to the CPT and in the differential vasomotor regulation between the CCA and ICA.