The association of low blood glucose with central nervous system (CNS) injury was first described in 1937 by Hartmann and Jaudon. In the early 60 years since publication of these observations the effects of hypoglycemia upon the brain remain poorly understood. Technology capable of accurately determining plasma glucose concentrations has been developed. Investigators have sought to establish critical values below which glucose levels should not be allowed to fall. Despite these efforts the definitive level of glucose capable of producing brain injury in any particular patient remains unknown. Glucose homeostasis within the neonatal CNS represents a dynamic process consisting of many interrelated variables including gestational and chronologic age, genotype, relative health, blood flow, metabolic rate and availability of other suitable substrates. New technique for assessing the glucose delivery: consumption ratio and directly monitoring the cellular consequences of glucose deprivation within discrete regions of the brain will help to answer the question 'How long is too low and how long is too long?'