Table 2

Heritability of sleep phenotypes

nh2 (SE)LRT P valueFixed-effects covariates
P value
AgeSexOSA (yes/no)Medication
Total sleep time4600.50.00010.008nsns
Sleep efficiency (%)460.32 (0.40)0.1860.0020.025nsns
Sleep latency (min)460.41 (0.25)0.054nsnsnsns
Total wake (min)460.22 (0.57)0.444nsnsnsns
Total wake (%)460.24 (0.58)0.4180.073nsnsns
WASO (min)4500.50.033nsnsns
AHI (events/h)4500.50.005nsns
N1 (min)4600.5nsnsnsns
N1 (%)4600.5nsnsnsns
N2 (min)4600.5ns0.032nsns
N2 (%)4600.50.006nsns0.026
SWS min460.43 (0.26)0.058<0.0001nsnsns
SWS %460.49 (0.24)0.0300.0007nsnsns
REM min460.65 (0.44)0.052nsnsnsns
REM %460.44 (0.30)0.069nsnsnsns
Delta power (µV2)480.50 (0.22)0.0110.00010.017nsns
Theta power (µV2)480.93 (0.15)0.0001§ns0.0004ns0.085
Alpha power (µV2)480.51 (0.41)0.102ns0.027nsns
Sigma power (µV2)480.71 (0.26)0.023ns0.004nsns
  • SEs are shown for all heritability estimates greater than zero. LRT P value: P values from LRT indicating whether the heritability estimated is significantly different from zero. Fixed-effects covariates P value: reported P values from LRT indicating whether the variable significantly impacted the heritability model. Age and sex were retained in the model regardless of their LRT P value, while OSA, BMI, and medication were dropped if P > 0.1 (reported in table as not significant [ns]) to result in a parsimonious model. Sleep stages are expressed as minutes as well as percent of total sleep time. Significantly different data appear in boldface type. Trends for statistical significance appear in italics. WASO, wake after sleep onset.

  • §P values exceeding the Bonferroni corrected threshold for significance of P = 0.003 (corrected for the number of traits analyzed).