Objective: The study was designed to determine metformin effects on meal pattern, gastric emptying, energy expenditure, and to identify metformin-sensitive neurons and their phenotype. Methods: This study was performed on C57BL/6J and obese/diabetic (db/db) mice. Metformin (300 mg/kg) was administered by oral gavage. Food intake, meal pattern, oxygen consumption (VO2), and carbon dioxide production (VCO2) were obtained using an Oxylet Physiocage System. Gastric emptying assay and real-time RT-PCR from dorsal vagal complex extracts were also performed. C-Fos expression was used as a marker of neuronal activation. Phenotypic characterization of activated neurons was performed using either proopiomelanocortin (POMC)-Tau-Topaz GFP transgenic mice or NUCB2/nesfatin-1 and tyrosine hydroxylase (TH) labeling. Results: Acute per os metformin treatment slowed down gastric emptying, reduced meal size, but not meal number in a leptin-independent manner, and transiently decreased energy expenditure in a leptin-dependent manner. Metformin specifically activated central circuitry within the brainstem, independently of vagal afferents. Finally, while POMC neurons seemed sparsely activated, we report that a high proportion of the c-Fos positive cells were nesfatinergic neurons, some of which coexpressing TH. Conclusions: Altogether, these results show that metformin modifies satiation by activating brainstem circuitry and suggest that NUCB2/nesfatin-1 could be involved in this metformin effect