Chilling or freezing of dough are widely used in industrial bakeries to facilitate bread production. Such intermediary stages induce significant spatial and temporal temperature gradients over the dough. Non-invasive MRI temperature mapping would permit to measure these gradients and to contribute to the better understanding of the chilling process and of its impact on bubble inflation. In this study, 2D-SE, 2D-GE and 3D-GE two-point T1-weighed MRI methods were evaluated in order to maximise the accuracy of temperature mapping. The optimal parameters of the methods were determined empirically on a low-field MRI and compared to the results from the theoretical estimation of uncertainty [1]. The ultimate objective was to select the best MRI method to monitor temperature during dough chilling. MATERIALS AND METHODS The study was performed on a 0.2 T MRI scanner (Open, Siemens). Different TR and α of the two-point 2D- and 3D-GE sequences and different TR of the two-point 2D-SE sequence were tested. Uncertainties were propagated along each experimental procedure and were used for the comparison between methods. The experimental uncertainties were also compared to the theoretical estimations. Temperature mapping in pre-fermented dough during chilling prior to freezing was finally achieved with the best method found and compared locally to optical fibre measurements. RESULTS AND DISCUSSION All the methods with comparable acquisition times minimised the temperature uncertainty to the same extent. The 2D-SE method was chosen for further applications on pre-fermented dough because of its lower sensitivity to susceptibility differences in porous media. The experimental values of uncertainties agreed to a great extent with the theoretical prediction, for both SE and GE methods. Finally, temperatures as measured by MRI or optical fibre during dough chilling were in good agreement.