IntroductionBrain metabolism is provided almost exclusively by glucose oxidation. The high metabolic demand of the brain, associated with its inability to store energy, requires an important and stable cerebral perfusion, in order to guarantee a normal cerebral function. Mean cerebral blood flow (CBF) values range around 50 cc per 100 g and per minute. They are, however, higher in young people and lower in the elderly. Regional cerebral blood flows (rCBF) are twice or thrice superior in the gray matter, compared with those in the white matter. rCBF are increased in activated cerebral areas. The rCBF changes relating to cerebral activity modifications may reach 10±20 %, and up to 40 % in case of extreme metabolic conditions, such as coma or convulsion [1, 2].Complex autoregulation processes ensure both the adjustment of rCBF to local energetic needs, determined by the activity level of local neurons and the stability of CBF despite changes in systemic arterial pressure. This rCBF autoregulation is controlled by intricate neurobiochemical mechanisms, involving sensitivity to blood pressure, blood pCO 2 and pH. Brain vascular autoregulation notably allows for a vascular dilatation when the systemic pressure tends to lower, in order to keep a constant CBF. This vascular dilatation leads in turn to an increased cerebral blood volume (CBV) [3, 4].The rCBF alterations are encountered in a variety of pathological conditions, the most frequent being strokes. Strokes affect 500,000 patients every year in the United States and represent the third cause of death in Eur. Radiol. (2001) Abstract Viability of the cerebral parenchyma is dependent on cerebral blood flow (CBF), which is usually kept in a very narrow range due to efficient autoregulation processes and can be altered in a variety of pathological conditions. An accurate method allowing for a quantitative assessment of regional cerebral blood flows (rCBF) and available for the routine clinical practice would, for sure, greatly contribute to improving the management of patients with cerebrovascular diseases. Different imaging techniques are now available to evaluate rCBF: positron emission tomography; single photon emission CT; stable-xenon CT; perfusion CT; and perfusion MRI. Each of these imaging techniques uses an indicator, with specific biological properties, and is supported by a model, which consists of a few simplifying assumptions, necessary to state and solve the equations giving access to rCBF. The obtained results are more or less reliable, depending on whether modeling hypotheses are fulfilled by the used indicator. The purpose of this article is to review the various supporting models in the assessment of rCBF, with special emphasis on perfusion CT studies at low injection rates and on iodinated contrast material used as an indicator.