Vigabatrin (VGB) causes intramyelinic edema (microvacuolation) in brain of dogs and rodent, which has encouraged development of noninvasive methods to monitor for this effect during clinical trials. We report the qualitative ex vivo magnetic resonance imaging (MRI) changes observed in a neuropathology study in dogs to detect time of onset and regression of VGB-induced intramyelinic edema. Beagles were randomly assigned to 18 groups of 6 dogs per group and administered vigabatrin orally (p.o.) at a dose of 300 mg/kg/day (2 males, 2 females) or placebo (1 male, 1 female). Animals were killed and examined at weekly intervals during the 12 weeks of treatment and at 1, 2, 4, 8, 12, and 16 weeks after discontinuation of drug treatment. Myelin microvacuolation in thalamus, hypothalamus, and fornix were noted histologically after 4-5 weeks of treatment. Increases in MRI T2 intensity were observed in hypothalamus after 4 weeks and in thalamus and columns of the fornix after 7 weeks. Both MRI T2 intensity and microvacuolation continued to increase during 12-week VGB treatment. When VGB treatment was discontinued after 12 weeks, both MRI T2 intensity and microvacuolation began to decrease. Sixteen weeks after VGB discontinuation, histopathology had returned to normal and MRI examination demonstrated a marked trend toward reversal of the increased T2 signal intensity. MRI thus has potential as a noninvasive surveillance technique in certain experimental and clinical conditions associated with intramyelinic edema.