Introduction: Genetic underpinnings of sleep architecture in newborns are largely unknown. One of the first findings on the genetics of REM sleep in newborns was by Gould et al. at 1978. Here we report on the results of classical twin study in monozygotic (MZ) and dizygotic (DZ) pairs examine the sleep macrostructure during the first three months of life. Methods: Polysomnographic recordings were obtained in Children.’s Sleep Laboratory, University Hospital Split. 15 pairs of MZ and 15 pairs of DZ twins were included, all born prematurely, first recorded at the 37th(range 36.–38) postmenstrual age (PMA), second at the 44 th (range 44 .– 45) PMA and third at 52th PMA. Visual analysis and spectral analyses were made for every PSG recording. Results: Correlations of total sleep time (TST) between MZ pairs among three measures ranges from r=0.09 to r=0.6 whereas on DZ pairs from r=-0.02 to r=0.37. Sleep efficiency (SF) showed higher correlations between MZ pairs (0.02 to 0.5) then DZ pairs (-0.15 to -0.35). In active sleep (AS), MZ pairs showed similar correlations (-0.22 to 0.73) as in DZ pairs (0.04 to 0.74). There was no significant difference between correlations of MZ and DZ pairs (p>0.05) in quite sleep (QS) time at the 52th PMA whereas on the 37th and 52th there were inconsistency. These preliminary study showed higher correlations between MZ pairs then DZ pairs only at sleep efficiency parameter and lots of inconsistency in other parameters of sleep macrostructure. Conclusion: Our study indicates that visual sleep state scoring during first three months can reveal only limited information about sleep physiology of the newborns, and to obtain more detailed insight, spectral analyses of the EEG signal is needed.