Objectives: This fMRI study evaluated the cognitive mechanisms and the cerebral substrates when evaluating the healthiness of food products from nutritional information displayed either with a traffic light (TL) system, a colored nutritional label, or with a guideline daily amount (GDA) system, a numeric label. We postulated that TL label would recruit emotional processes and activation of subjacent cerebral regions (e.g. insula and amygdala). On the contrary, the nutritional information presented in a GDA label, would recruit, due to its numeric format and higher complexity, supplementary cognitive processes and activation of related brain regions (e.g. middle and superior frontal as well as parietal cortices). Methods: We examined 50 healthy participants during an evaluation task on the healthiness of real food products from nutritional information only. Per total, 60 food products nutritional labels have been presented, with either colored (TL) or numeric (GDA) nutritional information and three levels of complexity of nutritional information. Results: In line with our predictions, evaluations based on GDA recruited prefrontal and parietal regions reported for analytic processes. Contrary to our predictions, the same network has been recruited when evaluations were based on TL. Finally, we found significant correlation between response time and the superior parietal lobule in the GDA condition. Discussion: Our results suggested that TL did not have an effect on the used strategy compared to GDA, based on calculation and arithmetic processes. Correlations between response time and brain activations suggested a significant involvement of the arithmetic mechanisms in the evaluation of food healthiness.