Although simple animal models relevant to human disease are lacking, many recent clinical and experimental studies have focused on the pathophysiology of cerebral malaria. Evidence of sequestration of parasitized erythrocytes is found in the cerebral capillaries of patients dying from cerebral malaria. Cytoadherence of parasitized red blood cells to endothelium is an essential process. However, no correlation was found between in vitro cytoadherence of clinical isolates and the presence of cerebral symptoms. Rosetting is defined by the agglutination of nonparasitized erythrocytes around red cells containing mature forms of the parasite, and probably contributes to the intravascular sequestration of erythrocytes. This phenomenon occurs in vitro, and isolates from patients with cerebral malaria appear to have increased rosetting properties. The excellent recovery of most survivors, even after a deep and lengthy coma, suggests that microvascular obstruction causing cerebral hypoxia is not the main factor contributing to cerebral dysfunction. Raised intracranial pressure with or without cerebral oedema is common in children, and contributes to mortality. The nonspecific immune inflammatory response of the host to the malarial parasite, with release of various mediators, seems to be of paramount importance. Among cytokines, tumor necrosis factor (TNF) appears to be associated with mortality; although plasma levels of TNF are not correlated with neurological dysfunction, this does not exclude a role of this cytokine at a paracrine level. Cytoadherence of parasitized red blood cells to endothelium and concomitant activation of mononuclear blood cells may be responsible for a local synthesis of cytokines, or even neurotransmitters that remain to be identified.