Table 2. Univariate heritability estimates for sleep and spindle traits.

Phenotype	White CFS families		Black CFS families
	h2	P-value	h2	P-value
Sleep stage duration (min)
N1 sleep	0.26	0.002	0.40	2 × 10−5
N2 sleep	0.21	0.03	0.06	0.27
N3 sleep	0.43	1 × 10−5	0.22	0.008
REM sleep	0.20	0.003	0.19	0.016
Total sleep time	0.21	0.01	0.24	0.009
Spectral band power
Slow	0.26	0.008	0.20	0.006
Delta	0.26	0.004	0.28	0.0001
Theta	0.29	0.009	0.30	3 × 10−5
Alpha	0.41	0.0002	0.34	2 × 10−6
Sigma	0.74	3 × 10−10	0.49	2 × 10−8
Beta	0.43	0.0005	0.43	4 × 10−8
Spindle traits (N2 sleep)
Density	0.45	8 × 10−6	0.43	3 × 10−6
Amplitude	0.48	7 × 10−6	0.39	3 × 10−6
Duration	0.39	0.0003	0.23	0.01
Frequency	0.39	0.0008	0.33	0.0003
Oscillations	0.41	0.0002	0.20	0.02
Symmetry	0.31	0.001	0.11	0.10

CFS, Cleveland Family Study; REM, rapid eye movement; SNP , single-nucleotide polymorphism.

Separately for black and white individuals from the CFS, heritability (h²) was estimated using mixed models applied to SNP microarray data. Owing to the relatively small sample sizes for this type of analysis, confidences intervals will be broad and so stochastic fluctuations in reported h² values are to be expected, and should not be overinterpreted. Most aspects of sleep architecture (minutes in N1, N3 and REM) showed significant heritability. Spectral band power for all frequencies showed significant genetic influence, in particular sigma activity. (Similar results were obtained using relative rather than absolute band power, data not shown.) Spindle traits during N2, in particular density and amplitude, also showed highly significant (P<10⁻⁵) estimates for heritability, ∼0.4–0.5. Intraclass correlations based on first siblings were also consistent with heritable influences on spindle traits (Supplementary Table 24).