Transcranial Doppler sonography (TCD) is a simple, noninvasive bedside procedure that can be repeated any time for the measurement of cerebral blood flow velocity in the great basal cerebral arteries. It is practicable in most severely head-injured patients in critical care. Flow patterns and pulsatility index (PI) resulting from maximal systolic and diastolic flow velocities and representing cerebrovascular resistance give quite an accurate impression of potential intracranial hypertension and the dependent cerebral perfusion pressure (CPP). With increasing intracranial pressure (ICP) and decreasing CCP, diastolic flow is progressively reduced. If ICP reaches the systemic diastolic blood pressure level, diastolic flow disappears. Oscillating (reverberating) flow patterns are seen when ICP increases further up to the arterial mean pressure level. The authors' own studies on 20 comatose patients with raised ICP showed typical changes in TCD parameters following different therapies for intracranial hypertension. Under continuous TCD monitoring of the middle cerebral artery, increases in maximal flow velocity (from 4% up to 102%, on average 27%) and mean flow velocity (from 18% up to 153%, on average 73%) were always observed after osmotherapy. In addition, a variable increase in negative frequencies was noted, probably due to increased turbulences. After barbiturate administration (thiopentone bolus of 0.3 g) a flow reduction was always seen [from - 2% up to - 25% (on average - 13%) for maximal flow velocity and from - 9% up to - 30% (on average - 19%) for mean flow velocity]. The efficiency of all therapeutic measures for decreasing intracranial hypertension, controlled by continuous ICP monitoring, was found to be closely related to the degree of drop in Pl. In constant respiratory conditions with mild hyperventilation a clear, but only intraindividual, linear relation between (raised) ICP over 20 mmHg and PI could be observed, in some cases over a period of several days. On the other hand, there was a large degree of interindividual variability in ICP and PI. Thus, in patients with an ICP level of 30 mmHg, a wide range of PI, from 1.0 up to 3.5, was found. In conclusion, TCD proved to be a useful method for assessing the effectiveness of osmotherapy and barbiturate administration in cases of intracranial hypertension, especially if ICP monitoring is not available. TCD is also useful for the diagnosis of posttraumatic vasospasm. Disturbed cerebrovascular autoregulation can be assessed by determination of CO, reactivity. Oscillating (reverberating) flow patterns and systolic spikes indicate intracranial circulatory arrest and can be used as an additional method of determining brain death.