The memory-improving action of glucose has now been studied for almost 20 years and the study of this phenomenon has led to a number of important developments in the understanding of memory, brain physiology and pathological consequences of impaired glucose tolerance. Glucose improvement of memory appears to involve two optimal doses in animals (100 mg/kg and 2 g/kg) that may correspond to two physiological mechanisms underlying glucose effects on memory. In humans, there have been few dose-response studies so the existence of more than one effective dose in humans is uncertain. Many tasks are facilitated by glucose in humans but tasks that are difficult to master or involve divided attention are improved more readily that easier tasks. There are a number of hypotheses about the physiological bases of the memory-improving action of glucose. Peripheral glucose injections could alleviate localized deficits in extracellular glucose in the hippocampus. These localized deficits may be due to changes in glucose transporters in that structure. Because certain neurotransmitters such as acetylcholine are directly dependent on the glucose supply for their synthesis, glucose is thought to facilitate neurotransmitter synthesis under certain circumstances. However, these hypotheses cannot account for the specificity of the dose-response effect of glucose. A number of peripheral mechanisms have been proposed, including the possibility that glucose-sensitive neurons in the brain or in the periphery may serve as glucose sensors and eventually produce neural changes that would facilitate memory processing. These latter results could be of importance because the mechanisms they suggest appear to be dose-dependent, a crucial characteristic to explain the dose-dependent effects of glucose. There may be an advantage to develop hypotheses that include both peripheral and central actions of glucose. There is evidence that impaired glucose regulation is associated with impaired cognition, particularly episodic memory. This impairment is minimal in young people but increases in older people (65 years and over) where it may compound other aging processes leading to reduced brain function. A small number of studies showed that glucose improvement of memory is associated with poor glucose regulation although this may not be the case for diabetic patients. Results of a few studies also suggest that drug treatments that improve glucose regulation also produce cognitive improvement in diabetic patients.