The brain is a complex system whose function relies on a dynamic information exchange between trillions of neural connections organised hierarchically: local neuronal circuits are interconnected to form large-scale functional networks spanning several brain areas. Neural oscillations are the result of this multilevel interaction and regulate vital processes, from sleep to attention. Neurophysiological techniques, such as electrophysiological recordings or brain imaging, can only investigate separately the micro- and macro-circuits that, together, generate global activity patterns. To date, despite significant recent technical advancements, the causal roles between local and global brain activity, and between global dynamics and overall brain function, remain largely unknown. In this context, complementary computational approaches can dramatically improve the understanding of the multilevel functional organization of the brain. This project aims to develop the first model of whole-brain slow oscillations based on the integration of multi-scale neural activity. Slow neural oscillations (<1Hz) regulate key functions such as synaptic plasticity, memory consolidation and sensory processing. Moreover, abnormalities in this brain rhythm have been linked to the pathogenesis of autistic spectrum disorders. The novelty of my model lays in three aspects. It will include pyramidal neurons, parvalbumin interneurons and somatostatin interneurons, following on the experimental results defining their differential roles in regulating slow waves; it will be based on the integration of multi-scale experimental data acquired in-house (local field potential recordings and fMRI); it will be used to model the slow dynamics alterations occurring in genetically defined autistic-like disorders. This solid and highly credible computational tool will advance our understanding of the physiology of brain oscillations and will potentially impact the diagnostic and therapeutic paths for autism.