Abstract Introduction Shortened mean sleep latency and multiple Sleep Onset REM Period (SOREMP) on MSLT has been used as a characteristic marker for narcolepsy and incorporated in all editions of International Classification of Sleep Disorders. Epidemiologic studies have shown that SOREMP could occur in general population especially those with sleep insufficiency and shift work. However the physiological meaning of SOREMP besides the diagnostic marker is not well clarified. We searched for subjective and objective sleep variables independently associated with the number of SOREMPs. Methods Participants were 769 consecutive sleepy patients or controls who gave written informed consent and underwent PSG followed by MSLT in Seiwa Hosiptal or Koishikawa Tokyo Hospital from October 2014 to November 2021. We excluded those with obvious nocturnal sleep disturbances (AHI≥10, PLMI≥15), with medication affecting sleep and with severe first-night effect at the time of PSG. We analyzed data from resultant 580 participants (male/female=312/268, age 25.3±8.4 years old, BMI 21.6±3.3). They were subdivided into 4 groups according to the number of SOREMPs on MSLT (used as ordinal scale). Subjects with 3 or more SOREMPs were merged into one group to increase number enough for statistical analyses. We dichotomized participants into narcolepsy type 1 (NT1) (n=50) and others (n=530) to adjust the effect of NT1 diagnosis. Epworth Sleepiness Scale (ESS) as well as conventional sleep variables on PSG and MSLT were compared by ANOVA and variables associated with the number of SOREMPs were further analyzed by regression analyses to determine the independent association after adjustment of covariates (age, sex, BMI and dichotomized diagnosis). Results Number of SOREMPs on MSLT was positively associated with ESS score (p=0.010), Total Sleep Time (TST) (p=0.03), REM %TST(p<0.001), and negatively associated with Sleep Latency (p<0.01), REM Latency (p<0.001), MSLT mean Sleep Latency (p<0.001) and MSLT REM Latency (p<0.001) and subjective sleep latency on PSG night. Conclusion Our data indicate that the number of SOREMPs have physiological significance besides narcolepsy diagnosis. More SOREMPs reflects higher sleep propensity and REM propensity as well as subjective sleepiness. The number of SOREMPs can be a marker reflecting one aspect of sleepiness. Our study will also contribute to better understand the meaning of SOREMP in diagnostic criteria of central disorders of hypersomnolence. Support (If Any) This study was partly supported by JSPS KAKENHI Grant Number JP21H02856.