By providing a non-invasive method for continuous display of mean flow velocity (Vmean) in the cerebral arteries, transcranial Doppler (TCD) ultrasound supplements evaluation of cerebral perfusion. Dynamic exercise increases middle cerebral artery (MCA) Vmean from approximately 55 to 65 cm s-1 dependent on work rate, and even more when corrected for changes of the arterial carbon dioxide tension. Evaluation of Vmean corresponds to that of cerebral blood flow as determined with the 133Xenon clearance technique, and reflects regional cortical regulation of the active muscles with important afferent nervous influence. Concomitant increases of mean arterial pressure (MAP) and heart rate is only of minor importance as illustrated during static exercise and post-exercise muscle ischaemia, where Vmean is not significantly elevated. During sustained head-up tilt, the Vmean remained unchanged at a MAP approximately 83 mmHg. Below this level, it decreased in parallel with MAP until MAP reached 50 mmHg. At an even lower MAP, Vmean seemed to approach a lower limit approximately 25 cm s-1, but at a diastolic pressure of 21 mmHg there was no flow in the MCA. Conversely, during post-exercise muscle ischemia, an increase in MAP to 140 mmHg did not influence Vmean. This is in contrast to patients operated for carotid artery stenosis and who develop ipsilateral headache. In these patients the ipsilateral MCA Vmean changed in parallel with MAP, and autoregulation was re-established only after one to two weeks. In patients with severe carotid stenosis and poor collateral circulation, the CO2-reactivity as expressed by Vmean was the lowest, and could be negative on the ipsilateral side. During carotid endarterectomy, a Vmean clamp/Vmean pre-clamp ratio below 0.6 identified patients with a cerebral blood flow below 20 ml 100 g-1 min-1. Furthermore, when the ratio was below 0.4 pathological electroenchephalographic changes developed. Thus, Vmean of the large basal cerebral arteries reflects cerebral perfusion with respect to regional flow distribution, autoregulatory response, and CO2-reactivity in normal man and patients with limited cerebral flow.