To determine the neuroprotective effect of fasting after traumatic brain injury (TBI) and to elucidate the potential underlying mechanisms, we used a controlled cortical impact (CCI) injury model to induce either a moderate or a severe injury to adult male Sprague Dawley rats. Tissue-sparing assessments were used to determine the level of neuroprotection of fasting, hypoglycemia (insulin 10 U), or ketone (1.66 mmoles/kg/day or 0.83 mmoles/kg/day; D-beta-hydroxtbutyrate) administration. Mitochondrial isolation and respiratory studies were utilized to determine the functionality of mitochondria at the site of injury. We also investigated biomarkers of oxidative stress, such as lipid/protein oxidation, reactive oxygen species (ROS) production, and intramitochondrial calcium load, as a secondary measure of mitochondrial homeostasis. We found that fasting animals for 24 hr, but not 48 hr, after a moderate (1.5 mm), but not severe (2.0 mm), CCI resulted in a significant increase in tissue sparing. This 24-hr fast also decreased biomarkers of oxidative stress and calcium loading and increased mitochondrial oxidative phosphorylation in mitochondria isolated from the site of injury. Insulin administration, designed to mimic the hypoglycemic effect seen during fasting did not result in significant tissue sparing after moderate CCI injury and in fact induced increased mortality at some injection time points. However, the administration of ketones resulted in increased tissue sparing after moderate injury. Fasting for 24 hr confers neuroprotection, maintains cognitive function, and improves mitochondrial function after moderate (1.5 mm) TBI. The underlying mechanism appears to involve ketosis rather than hypoglycemia. © 2008 Wiley-Liss, Inc.