Spontaneous thalamic waves of activity propagate to the cortex during embryonic development, influencing the organization of cortical structures. We have shown that structured patterns of neuronal activity in the thalamus of mouse embryos sculpt the functional columns in the cortex and the concomitant functional somatotopic map during immature cortical stages. We identified that the fundamental columnar organization of the thalamocortical somatotopic map already exists before birth. Moreover, our lab has recently demonstrated that sensory circuits emerge as nonsegregated modules and that at birth these circuits become segregated and sensory modalities specified. This segregation takes place in the superior colliculus, in a process that depends on the earliest activity from the retina. In my talk I will show these data and discuss unpublished results on how patterns of spontaneous activity in sensory stations might be used as a tool to predict circuit development and early sensory plasticity.