Although less often investigated than the positive blood-oxygen-level-dependent (BOLD) signal, a negative BOLD signal has also been observed under certain conditions, which could have a neuronal origin. To further investigate the negative BOLD signal from a metabolic point of view, a functional 1H magnetic resonance spectroscopy (fMRS) study was conducted using visual stimulations to trigger a negative BOLD response. The latter was linked to decreases of [Glu] and [Lac] with also a decrease in GABA concentration. These results suggest that the negative BOLD signal is linked to a neuronal deactivation and maybe to a local increase of inhibitory activity. Although both glycolytic and oxidative metabolisms are involved during neuronal activity, it is still unclear how they are respectively modulated by the latter. To investigate this, a functional magnetic resonance imaging (fMRI)-fMRS study was carried out on participants performing a finger tapping task at either 1Hz, 2Hz or 3Hz. In addition to the BOLD signal, the CBF signal was also obtained. As previously observed, the BOLD and CBF signals increased for larger finger tapping frequencies confirming the overall increase of metabolic demand. [Glu] and [Lac] were used as biomarker of oxidative and glycolytic metabolisms, respectively. [Glu] changes were the largest for 3Hz while [Lac] changes were more important at 2Hz. These results highlights the different involvement of the two metabolic pathways. The positive BOLD signal differs in its timings and shape across brain regions. Using a high temporal resolution sequence combined with short stimulations, the positive BOLD responses of six motor regions, including the cerebellum, were extracted. The main difference between motor regions was the positive peak and undershoot amplitudes that were the smallest in the cerebellum, which also showed delayed onsets. In addition to the BOLD response, the cerebellum has some very specific characteristics such as the representation of body parts in both the anterior and posterior lobes. In order to obtain more precise and complete maps of the cerebellar somatotopy, an fMRI motor task was performed testing for the eyes, the tongue, the little fingers, the thumbs or the toes. The anterior lobe showed a consistent and robust somatotopic gradient with also the presence of such a gradient in the posterior lobe, albeit less obvious. These results show that multiple representations of the body are present in the cerebellum and organized in an orderly manner. In addition, the cerebellum can be parcellated based on several features like cell packing, olive fibre input or molecular biomarker. To test whether such tissue characteristics can be observed at macroscale level, T1 and T2* mapping was performed on cerebellar surfaces generated at three different cortical depths. T1 surfaces showed an alternation of lower and higher T1 values when going from the median to the lateral part of the cerebellar hemispheres with shorter T1 values observed in the deeper grey matter layers. T2* maps showed a similar longitudinal pattern, but no change related to the cortical depths. These patterns reflect potential myelin and iron content variations over the cerebellar cortex. Whether these tissue variations are linked to the somatotopic organization is not clear.