Sleep spindles are a distinctive EEG phasic feature of NREM sleep and are prevalent during stage 2 as compared to slow wave sleep. While the neurophysiological mechanisms of spindle generation, that involves thalamic and corticothalamic networks, have been recently delineated and are briefly reviewed, their definitive functional meaning still remains to be elucidated. This review summarizes the present knowledge on visually scored and automatically detected spindles, as well as the literature on EEG power in the sigma band. Among the factors known to affect sleep spindles and sigma activity, their intra-cycle temporal dynamics, their time-course across sleep cycles, the reciprocal relationship with delta activity, the effects of sleep deprivation, of circadian factors and of ageing, and their role in information processing have been discussed. Moreover, specific attention has been paid to the existence of functionally and topographically distinct slow- and fast-spindles, also taking into account the presence of large inter-individual differences. Nevertheless, several fundamental issues remain to be elucidated: the physiological mechanisms controlling age-related changes in spindle parameters; the role of melatonin as a spindle-promoting agent; the relationships between plastic mechanisms (after stroke, or as a consequence of learning) and modifications in spindle activity; the possibility of using some spindle parameters as an index of the severity of developmental disorders in abnormal maturational processes.