The quality of sleep has recently come to the forefront of public health concerns in industrialized nations. Indeed, voluntary sleep curtailment is widespread, sleep disorders are increasingly recognized and both correlate with the current epidemiology of diabetes and obesity (Van Cauter and Knutson, 2008). Sleep fragmentation (SF) is a periodic disruption found in highly prevalent sleep-related breathing and movement disorders, where the sleeper might be aroused several hundred times per night. By roughly summing the prevalence of obstructive sleep apnea (OSA) syndromes and periodic limb movement disorders, two of the most common sleep disorders, we predict a total prevalence for SF of nearly 15% in the European active population. Overlooked cases and insufficient concern for the problematic bear a major socioeconomic burden because of the broad impact of excessive daytime sleepiness on productivity (Hossain and Shapiro, 2002), motor vehicle accidents (Aldrich, 1989) and medical complications (Spiegel et al., 2009). Recurrent interference with the natural architecture of sleep leads to sleepiness, difficulty in concentration and cognitive impairments such as decreased reaction times and poor working memory equivalent to a full night of sleep deprivation in severe cases (Bonnet, 1985, 1989; Stepanski, 2002). In addition, recent experimental evidence in humans has unraveled the unprecedented link between sleep of poor quality and dramatic endocrine imbalances such as deregulation of appetite controlling hormones, glucocorticoids and increase in sympathetic tone, ultimately leading to the development of insulin resistance (Spiegel et al., 2009; Stamatakis and Punjabi, 2010). Current animal models for the study of sleep pathology mostly involve short intervention often biased by stress induction and usually involving a complete period of sleep deprivation. These models do not tightly reflect the most common clinical pattern characterized by sleep perturbation over long periods of time. Therefore, paying attention to maximally avoid methodological stress, we designed a new device aiming at performing instrumental SF for fourteen days in mice. Our model mimics SF observed in sleep disorders since it has no major impact on circadian timing and total amount of sleep but clearly shortens individual sleep episodes. With this method, we approached three problematics that were insufficiently addressed in the literature. In our first study, we addressed the question of whether SF could trigger electrophysiological and molecular sleep homeostasis mechanisms in the short-term and whether these effects could be maintained in the long-term. We showed that SF leads to an anticipated homeostatic regulation of Slow Wave Activity (SWA), the sleep EEG hallmark of sleep homeostasis. Furthermore, EEG spectral analysis revealed an unexpected power increase in 4-40 Hz frequencies during sleep with a clear increase in the amplitude of slow waves and spindles. We concluded that sustained sleep pressure in SF leads to a global increase in cortical synchronization. We next explored whether these electrophysiological changes could impact on the expression of transcripts known to be regulated by sleep homeostasis (immediate early genes, clock-genes and heat shock proteins). Only heat shock proteins showed a consistent induction after one day but normalized after fourteen days of SF. These results point to the fact that electrophysiological and molecular aspects of sleep homeostasis are dissociable and that long-term sleep pressure triggers allostatic cellular processes of adaptation. In our second study, we explored the effects of long-term SF on general metabolism. We showed that SF led to a general imbalance in energy intake and expenditure, possibly through a slight elevation of brain temperature at rest. We also observed that glucocorticoids were slightly elevated at given circadian points and were accompanied by glucose intolerance. Altogether, this study brings evidence for the participation of SF in addition to intermittent hypoxia in the development of a metabolic syndrome in OSA patients. In the third study, we combined our chronic model with a spatial learning task spanning over several days (Morris Water Maze) in order to study the impact of SF on memory and metabolic plasticity. We showed that SF impaired proper memory consolidation supporting the idea that continuous sleep is required for specific mechanisms of memory trace post-processing. We also provide preliminary evidence that brain energy metabolism is not optimally responsive to neural activation after SF and that this effect might underlie the observed cognitive impairment. In conclusion, we showed that our animal SF model is suitable for studying the specific mechanisms underlying cognitive impairments and metabolic imbalance observed in sleep medicine. Our results underscore the existence of a global allostatic load in a wide range of systems during SF. Sensitization of the population to the prevalence of SF and its stealthy mode of presentation might improve its detection. Increased recognition of the problematic would be especially beneficial to cognitive development in school children with sleep apnea and might reduce accidents related to sleepiness. More specifically, we propose that screening sleep quality and effective correction of SF in patients with endocrine disorders has the potential to improve glucose metabolism and energy balance.