Longitudinal Stability of Genetic and Environmental Influences on the Association between Diurnal Preference and Sleep Quality in Young Adult Twins and Siblings

Abstract

Overlapping genetic influences have been implicated in diurnal preference and subjective sleep quality. Our overall aim was to examine overlapping concurrent and longitudinal genetic and environmental effects on diurnal preference and sleep quality over ~5 years. Behavioral genetic analyses were performed on data from the longitudinal British G1219 study of young adult twins and nontwin siblings. A total of 1556 twins and siblings provided data on diurnal preference (Morningness-Eveningness Questionnaire) and sleep quality (Pittsburgh Sleep Quality Index) at time 1 (mean age = 20.30 years, SD = 1.76; 62% female), and 862 participated at time 2 (mean age = 25.30 years, SD = 1.81; 66% female). Preference for eveningness was associated with poorer sleep quality at both time points (r = 0.25 [95% confidence intervals {CIs} = 0.20-0.30] and r = 0.21 [CI = 0.15-0.28]). There was substantial overlap in the genetic influences on diurnal preference and sleep quality individually, across time (genetic correlations [rAs]: 0.64 [95% CI = 0.59-0.67] and 0.48 [95% CI = 0.42-.053]). There were moderate genetic correlations between diurnal preference and sleep quality concurrently and longitudinally (rAs = 0.29-0.60). Nonshared environmental overlap was substantially smaller for all cross-phenotype associations (nonshared environmental correlations (rEs) = -0.02 to 0.08). All concurrent and longitudinal associations within and between phenotypes were largely accounted for by genetic factors (explaining between 60% and 100% of the associations). All shared environmental effects were nonsignificant. Nonshared environmental influences played a smaller role on the associations between phenotypes (explaining between -0.06% and 40% of the associations). These results suggest that to some extent, similar genes contribute to the stability of diurnal preference and sleep quality throughout young adulthood but also that different genes play a part over this relatively short time frame. While there was evidence of genetic overlap between phenotypes concurrently and longitudinally, the possible emergence of new genetic factors (or decline of previously associated factors) suggests that molecular genetic studies focusing on young adults should consider more tightly specified age groups, given that genetic effects may be time specific.

Keywords

chronotype diurnal preference genetic longitudinal sleep quality twins

The leading theory of sleep and wakefulness posits that the timing of our sleep-wake system is controlled by 2 processes. One is circadian and the other homeostatic (Borbély, 1982). The circadian system maintains 24-h daily physiological rhythms by a self-sustained oscillator located in the suprachiasmatic nucleus of the hypothalamus. The circadian system is synchronized to external time cues and influences the timing of sleep. It is possible to obtain a crude index of individual differences in circadian rhythms by subjectively assessing preferences for the timing of sleep-wake activities. Diurnal preference has repeatedly been assessed using the Morningness-Eveningness Questionnaire (MEQ; Horne and Östberg, 1976), which provides a subjective account of one’s tendency toward morning or evening activity. Diurnal preference as assessed by the MEQ has been shown to appropriately differentiate morning and evening types in terms of the timing of the circadian phase markers of melatonin, core body temperature, and subjective sleepiness (Lack et al., 2009). The homeostatic process refers to the rise in sleep drive, which accumulates with wakefulness and dissipates during sleep (Borbély and Achermann, 2011). Slow-wave sleep (SWS), often considered the “refreshing” stage of sleep, is used as a marker of homeostatic sleep pressure. Although it may be difficult to obtain a subjective measure of sleep homeostasis, assessment of the subjective quality of one’s sleep could provide an indirect measure of SWS and, by proxy, of the efficiency of the homeostatic drive. The Pittsburgh Sleep Quality Index (PSQI: Buysse et al., 1989) assesses 7 components of sleep quality pertaining to subjective sleep quality, sleep onset latency, sleep duration, sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. While indices of the circadian system and homeostatic drive may be considered in some ways to be independent, it is possible that they are interlinked in terms of the behavior they dictate as well as their underlying mechanisms. For example, it is possible that diurnal preference could be considered to be an index of the homeostatic drive; that, a faster buildup of sleep propensity could drive people toward a morning preference whereby they go to bed and arise earlier. Not surprisingly, these subjective assessments of diurnal preference and sleep quality are often associated (Barclay et al., 2010a; Megdal and Schernhammer, 2007; Merikanto et al., 2012; Soehner et al., 2011; Tzischinsky and Shochat, 2011). In these studies, eveningness (relative to morningness) is typically associated with various indices of poor sleep quality.

Behavioral genetic studies have repeatedly demonstrated that genetic influences account for about half the variability in diurnal preference (Barclay et al., 2010a; Drennan et al., 1992; Hur et al., 1998; Koskenvuo et al., 2007; Vink et al., 2001) and about 30% to 40% in sleep quality (Barclay et al., 2010a, 2010b; Genderson et al., 2013; Heath et al., 1990; Partinen et al., 1983). Analyses of the current data set have demonstrated that the association between diurnal preference and sleep quality is largely influenced by shared (also referred to as common; overlapping) genetic factors (Barclay et al., 2010a). However, little is known about the stability of this genetic overlap over time.

The organization of sleep and the circadian system change across developmental periods (Bliwise, 2005). The association between diurnal preference and sleep quality, as well as the contribution of genetic and environmental influences on this association, may also change over particular developmental periods. Accumulating evidence suggests that genetic and environmental influences on diurnal preference change with age (Barclay et al., 2014; Hur, 2007; Jones et al., 2007). A handful of studies have focused on the stability of the genetic and environmental influences on sleep quality over time (Barclay et al., 2015; Gregory et al., 2009), yet we know relatively little about the association between diurnal preference and sleep quality over time. It is important to address this question during young adulthood when physiology continues to develop and social responsibilities tend to increase (Arnett, 2000; Dahl and Lewin, 2002), factors that can affect both sleep timing and quality.

The current study assessed associations between diurnal preference and sleep quality concurrently and longitudinally in a sample of young adult twins and siblings from the G1219 study (for a description of this study, see McAdams et al., 2013). We examined 1) stability of diurnal preference and sleep quality over time; 2) associations between diurnal preference and sleep quality concurrently and longitudinally; 3) stability of genetic and environmental influences on diurnal preference and sleep quality, within phenotypes, over time; and 4) the extent of genetic and environmental overlap between phenotypes concurrently and longitudinally.

Materials and Methods

Participants

The present analyses focused on wave 4 and wave 5 (later referred to as times 1 and 2 for simplicity of presentation) of the G1219 and G1219Twins longitudinal studies (for a description of the sample, see McAdams et al., 2013).

At time 1 (wave 4), a total of 1556 individuals participated. Approximately 5 years later, at time 2 (wave 5), a total of 862 individuals participated. The mean age at time 1 was 20.30 years (SD = 1.76, range = 18-27 years), and 62% of the sample were female. The mean age at time 2 was 25.30 years (SD = 1.81, range = 22-32 years), and 66% of the sample were female. Following the study design, the majority of participants were close in age within each time point (at time 1, 90% of the participants were aged 18-22 years; at time 2, 90% of the participants were aged 22-27 years), but the inclusion of siblings inevitably created some age spread. Zygosity was established through a questionnaire measure completed by mothers at waves 2 and 3 assessing physical similarities between twins (Cohen et al., 1975). The number of monozygotic (MZ), dizygotic (DZ), and sibling (Sib) individuals as well as full pairs participating at each wave and providing data on each variable is displayed in Table 1.

Table 1.

Means (SD) of raw diurnal preference and sleep quality scores and number of individuals (N) and full pairs (FP).^a

	Total	Males	Females	MZ	DZ	Sibs
Time 1 MEQ	48.57 (8.12)^b	47.22 (8.39)^c	49.56 (7.77)	50.23 (7.67)^d	47.88 (8.16)	48.58 (8.25)
n	1376	554	822	398	695	283
FP				189	327	127
Time 1 PSQI	5.66 (3.02)^b	5.56 (2.99)	5.74 (3.03)	5.45 (2.86)	5.74 (3.11)	5.71 (2.95)
n	1354	550	804	388	684	282
FP				181	322	127
Time 2 MEQ	52.29 (8.39)^b	50.67 (8.71)	53.27 (8.05)	53.28 (8.71)	51.88 (8.07)	52.22 (8.58)
n	771	268	503	218	368	185
FP				103	170	84
Time 2 PSQI	5.31 (2.85)^b	5.24 (2.78)	5.35 (2.88)	5.14 (2.85)	5.32 (2.96)	5.48 (2.63)
n	757	262	495	213	360	184
FP				103	163	83

MEQ = diurnal preference (higher scores indicate a greater tendency toward morningness); PSQI = sleep quality (higher scores indicate poorer sleep quality); MZ = monozygotic twins; DZ = dizygotic twins; Sibs = siblings. Analyses were performed on untransformed (i.e., raw) data in Mx.

Significant differences within phenotypes across time.

Significant sex differences within phenotypes.

Significant zygosity differences within phenotypes.

Ethical approval for the collection of data at different waves was provided by the Research Ethics Committees of the Institute of Psychiatry, South London and Maudsley NHS Trust, and Goldsmiths, University of London. The protocol for data collection conformed to international ethical standards in accordance with the Declaration of Helsinki and Portaluppi and colleagues (2010).

Predictors of Attrition

Logistic regression, focusing on variables previously associated with attrition in this sample, demonstrated that responders at time 2 were significantly more likely than nonresponders to be female (odds ratio = 0.60, p < 0.001), have more highly educated mothers (odds ratio = 1.14, p < 0.001, where mother’s education was rated from the lowest level 0 to 8), and have a greater tendency toward morningness at time 1 (odds ratio = 1.02, p = 0.01). There were no significant differences between responders and nonresponders in terms of housing, sleep quality, or prosocial behavior at earlier waves (all p’s > 0.05).

Measures

The MEQ (Horne and Östberg, 1976) is a widely used measure of diurnal preference. It contains 19 self-report items measuring preference for the timing of daytime activities, sleeping habits, hours of peak performance, and times of “feeling best” and maximum alertness. A total score (range = 16-86) is derived in which higher scores indicate greater morningness. For the current study, this scale was reversed so that we could decompose a positive correlation for ease of interpretation for the reader. The MEQ demonstrated good internal reliability in the present sample at time 1 (Cronbach’s α = 0.78) and time 2 (α = 0.80).

The PSQI (Buysse et al., 1989) was used to examine sleep quality over the past month. The PSQI contains 18 scored items, which include both free-response questions (e.g., “During the past month, when have you usually gone to bed at night?”) and fixed-choice questions (e.g., “During the past month, how would you rate your sleep quality overall?” “Very good,” “Fairly good,” “Fairly bad,” or “Very bad”). The total score (range = 0-21) provides an overall sleep quality index, with higher scores indicating poorer quality. The PSQI demonstrated good internal reliability in the present sample at both time points (Cronbach’s α = 0.71 at both time points). Barclay et al. (2010a) provide further information on these measures.

Statistical Analyses

Data Preparation

Skew was not problematic for the MEQ or PSQI at either time point (time 1 MEQ skew = -0.16, [SE = 0.06], PSQI skew = 0.91, [SE = 0.06]; time 2 MEQ skew = -0.07, [SE = 0.08], PSQI skew = 0.93 [SE = 0.09]). Prior to analysis, the data were age and sex regressed, as has been recommended for twin modeling (McGue and Bouchard, 1984). Outliers of 3 or more standard deviations above or below the mean were omitted (n = 34 cases).

Genetic Analytic Principles Using Twin Data

Genetic models were fitted using maximum likelihood in Mx (Neale, 1997). Twin studies compare the similarity within MZ twin pairs to the similarity within DZ twin pairs (or full siblings) to estimate genetic influences on traits. Since MZ twins share 100% of their genes while DZ twins and full siblings share 50% of genetic effect on average, this information estimates the relative contribution of 3 sources of variance affecting a phenotype: additive genetic influences (A) (where alleles at a locus “add up” to influence behavior), shared environmental influences (C) (environmental influences that act to make twins within a pair similar), and nonshared environmental influences (E) (environmental influences that act to make twins within a pair different; this source of variance also incorporates error). The twin model assumes that correlations between DZ twins and siblings are relatively equal. If this assumption is upheld, parameters can be confidently estimated in genetic models by constraining these estimates to be equal between groups. If the correlation between MZ pairs is greater than that of DZ/sibling pairs, genetic influences are indicated. MZ twin correlations equivalent to DZ/sibling correlations indicate the importance of the shared environment and the extent to which MZ twin correlations are smaller than 1 indicates the magnitude of nonshared environmental factors.

Bivariate twin correlations indicate the relative sources of variance accounting for the concurrent and longitudinal phenotypic correlations. These were modeled using a correlated factors approach (see Figure 1). This model provides information on the univariate estimates and the overlap in the genetic and environmental influences on each phenotype at each time point. This information can be used to calculate the proportion of genetic and environmental influences accounting for the concurrent and longitudinal associations.

Figure 1.

Multivariate correlated factors model. A = additive genetic influence; E = nonshared environmental influence. Figure is shown for 1 twin/sibling only.

The fit statistic provided by Mx for raw data modeling is -2LL (minus twice the log likelihood of the observations). Saturated models, which estimate the maximum number of parameters required to describe the variance-covariance matrix and means of observed variables, providing a perfect fit to the data, are first approximated to the data. The -2LL of a saturated model is then subtracted from the -2LL of the genetic model, with lower values indicating better genetic model fit. The difference is χ² distributed; nonsignificance indicates that the genetic model does not fit the data less well than the saturated model and therefore provides a good description of the data. We first tested a scalar-sex limitation ACE model that allows quantitative sex differences. This was then followed by a homogeneity model in which male and female parameters were equated. Finally, we ran an AE homogeneity model to determine whether the C parameter could be dropped without significantly reducing model fit.

Results

Descriptives

Table 1 (and Supplementary Figures S1-S12) shows the means and standard deviations of diurnal preference and sleep quality scores from time 1 and 2, split by sex and zygosity. Compared with females, males tended toward eveningness at times 1 (fully constrained model compared with sex differences model: Δχ² = 27.94, Δdf = 2, p < 0.001) and 2 (Δχ² = 17.44, Δdf = 2, p < 0.001). Sleep quality did not differ between males and females. There were zygosity differences in diurnal preference (Δχ² = 17.51, Δdf = 4, p < 0.001), where MZ twins tended toward morningness compared with DZ twins and siblings at time 1 but not time 2, and no zygosity differences in sleep quality. There was a significant increase in morningness across time (Δχ² = 162.44, Δdf = 2, p < 0.001) and a significant improvement in sleep quality (Δχ² = 9.79, Δdf = 2, p < 0.001).

Zero-Order Phenotypic Correlations

As previously noted (Barclay et al., 2010a), time 1 eveningness was associated with poorer sleep quality (r = 0.25 [95% confidence intervals [CIs] = 0.20-0.30]), and this was similar at time 2 (r = 0.21 [CI = 0.15-0.28]). There was substantial stability for each phenotype across time (diurnal preference: r = 0.64 [CI = 0.59-0.67]; sleep quality: r = 0.48 [CI = 0.42-0.53]). The cross-trait longitudinal correlations indicated that diurnal preference at time 1 was significantly associated with sleep quality at time 2 (r = 0.14 [CI = 0.07-0.20]). Likewise, time 1 sleep quality was significantly associated with time 2 diurnal preference (r = 0.20 [CI = 0.14-0.26]).

Univariate Twin Correlations

As shown in Table 2, all univariate MZ twin correlations were significant (range r = 0.24-0.52), as were 3 of the univariate DZ twin correlations (range r = 0.17-0.26). All univariate sibling correlations were nonsignificant (range r = 0.02-0.16).

Table 2.

Univariate and bivariate twin correlations (95% confidence intervals) for diurnal preference and sleep quality at time 1 and 2.^a

	Twin/Sibling 1
	MEQ Time 1	PSQI Time 1	MEQ Time 2	PSQI Time 2
MZ twins
Twin 2
MEQ Time 1	0.51 (0.39 to 0.60)
PSQI Time 1	0.24 (0.16 to 0.32)	0.43 (0.30 to 0.54)
MEQ Time 2	0.47 (0.37 to 0.56)	0.20 (0.09 to 0.30)	0.52 (0.37 to 0.65)
PSQI Time 2	0.13 (0.02 to 0.24)	0.24 (0.11 to 0.36)	0.24 (0.12 to 0.35)	0.24 (0.05 to 0.41)
DZ twins
Twin 2
MEQ Time 1	0.11 (0.00 to 0.21)
PSQI Time 1	0.04 (–0.04 to 0.12)	0.23 (0.13 to 0.33)
MEQ Time 2	0.09 (–0.02 to 0.19)	0.07 (–0.03 to 0.16)	0.17 (0.02 to 0.31)
PSQI Time 2	0.00 (–0.10 to 0.09)	0.12 (0.01 to 0.22)	0.12 (0.01 to 0.23)	0.26 (0.10 to 0.41)
Siblings
Sibling 2
MEQ Time 1	0.16 (0.00 to 0.31)
PSQI Time 1	0.06 (–0.06 to 0.17)	0.02 (–0.14 to 0.19)
MEQ Time 2	0.11 (–0.03 to 0.25)	0.03 (–0.10 to 0.16)	0.11 (–0.09 to 0.30)
PSQI Time 2	0.09 (–0.06 to 0.22)	0.28 (0.13 to 0.41)	0.02 (–0.13 to 0.16)	0.05 (–0.16 to 0.26)

MEQ = diurnal preference (this variable has been reversed so that higher scores indicate a greater tendency toward eveningness); PSQI = sleep quality (higher scores indicate poorer sleep quality); MZ = monozygotic twins; DZ = dizygotic twins; Sibs = siblings. Analyses were performed on age- and sex-regressed data in Mx. The model was constrained where appropriate. For example, the twin correlations are constrained so that those of the randomly selected twin 1’s are the same as the randomly selected twin 2’s.

For the cross-twin within-trait correlations for diurnal preference at both time points, MZ correlations were significantly greater than the corresponding DZ correlations, suggesting the presence of genetic effects. For the cross-twin within-trait correlations for sleep quality, while the ratio of MZ:DZ twin correlations at time 1 were indicative of genetic effects, at time 2 twin correlations for MZ and DZ twins were similar, suggesting a greater influence of the environment on sleep quality at this time point. Of note, however, the sibling correlation was much smaller than the MZ twin correlation, which was more in line with expectations and highlighted possible genetic effects. Equating DZ twin and sibling correlations for sleep quality at time 2 did not significantly reduce model fit compared with the model in which MZ, DZ, and sibling correlations were free to differ (Δχ² = 2.46, Δdf = 1, p = 0.12), indicating no significant differences in the correlations between DZ twins and siblings.

Bivariate Twin Correlations

As noted previously (Barclay et al., 2010a), the cross-twin cross-trait concurrent association between diurnal preference and sleep quality at time 1 appeared to be influenced by genetic factors (Table 2: MZ twin correlation was greater than DZ twin and sibling correlations). A similar pattern of twin correlations emerged at time 2. The stability over time in diurnal preference appeared to be influenced by genetic factors; inspection of the longitudinal cross-twin within-trait correlations showed that the MZ correlation was greater than DZ twin/sibling correlation. The longitudinal association between sleep quality at time 1 and 2, however, appeared to be influenced by both genetic and environmental (shared and nonshared) factors (MZ correlation > DZ twin correlation and similar to the sibling correlation). Equating DZ twin and sibling correlations for the association between sleep quality at time 1 and 2 did not significantly reduce model fit. For the cross-twin cross-trait associations longitudinally, genetic factors appeared to influence the association between diurnal preference at time 1 and sleep quality at time 2. Genetic factors also appeared to contribute to the association between sleep quality at time 1 and diurnal preference at time 2 (MZ correlation > DZ twin/sibling correlations).

Multivariate Correlated Factors Model

The full ACE model allowing sex differences was first run to estimate the influence of each source of variance (full scalar sex-limitation ACE model fit compared with the saturated model: Δχ² = 335.48, Δdf = 246, p < 0.001, AIC = -156.52). Equating estimates for males and females (homogeneity model) resulted in a significant worsening of fit Δχ² = 32.08, Δdf = 12, p < 0.001). However, a reduced AE homogeneity model (where C parameters were fixed at 0) did not fit significantly worse than the ACE homogeneity model Δχ² = 1.86, Δdf = 10, p = 0.99, AIC = -18.14) nor the AE scalar sex-limitation model Δχ² = 14.96, Δdf = 8, p = 0.06, AIC = -1.04). Thus, the AE homogeneity model was selected as the most parsimonious best-fitting model.

Univariate and Bivariate Estimates

The AE homogeneity multivariate correlated factors model (see Figure 1) estimates the relative contribution of genetic and environmental influences on the individual traits, and their associations, at each time point. Table 3 displays the univariate parameter estimates on the individual traits at each time point. Table 4 (top part) shows the genetic and environmental overlap between phenotypes concurrently and longitudinally. For example, there were moderate genetic correlations between diurnal preference and sleep quality concurrently at both time points (time 1: rA = 0.48 [95% CI = 0.31-0.65]; time 2: rA = 0.60 [95% CI = 0.34-0.94]). Longitudinally, the cross-trait genetic correlations were also significant, although smaller (rA’s = 0.29-0.40). The within-trait cross-time correlations demonstrated a substantial degree of stability in the genetic influences on diurnal preference and sleep quality individually, across time (rA’s: 0.90 [95% CI = 0.77-1.00] and 0.80 [95% CI = 0.57-1.00], respectively). Nonshared environmental overlap was substantially smaller for all associations (rE’s range from -0.02 to 0.41).

Table 3.

Univariate parameter estimates for diurnal preference and sleep quality (with 95% confidence intervals).^a

	A	E
Time 1 MEQ	45% (95% CI = 0.34-0.55)	55% (95% CI = 0.45-0.66)
Time 1 PSQI	42% (95% CI = 0.31-0.52)	58% (95% CI = 0.48-0.69)
Time 2 MEQ	46% (95% CI = 0.33-0.58)	54% (95% CI = 0.42-0.67)
Time 2 PSQI	29% (95% CI = 0.15-0.42)	71% (95% CI = 0.58-0.85)

MEQ = diurnal preference; PSQI = sleep quality; A = additive genetic influences; E = nonshared environmental influences; CI = 95% confidence intervals. Analyses were performed on age- and sex-regressed data in Mx. Estimates from the homogeneity AE multivariate correlated factors model. Note that slight discrepancies in the parameter estimates presented here compared with those published in Barclay et al. (2010a) are due to differences in the structure and fit of the multivariate correlated factors model compared with the purely univariate model. This is because multivariate models use all information (including that from other variables) to provide univariate estimates, and this additional information can lead to slight fluctuations in results.

Table 4.

Parameter estimates (with 95% confidence intervals) from the homogeneity AE multivariate correlated factors model.^a

	MEQ Time 1	PSQI Time 1	MEQ Time 2	PSQI Time 2
Additive genetic and Non-shared environmental overlap between phenotypes
MEQ Time 1	/
PSQI Time 1	rA 0.48 (0.31 to 0.65)rE 0.08 (−0.04 to 0.20)	—
MEQ Time 2	rA 0.90 (0.77 to 1.00)rE 0.41 (0.28 to 0.53)	rA 0.40 (0.19 to 0.62)rE 0.04 (−0.12 to 0.19)	—
PSQI Time 2	rA 0.29 (0.01 to 0.59)rE 0.05 (−0.10 to 0.20)	rA 0.80 (0.57 to 1.00)rE 0.30 (0.15 to 0.43)	rA 0.60 (0.34 to 0.94)rE −0.02 (−0.17 to 0.14)	—
Proportion of phenotypic correlation due to A and E influences
MEQ Time 1	—
PSQI Time 1	A 0.83 (0.53 to 1.11)E 0.17 (−0.11 to 0.47)	—
MEQ Time 2	A 0.64 (0.49 to 0.78)E 0.36 (0.22 to 0.51)	A 0.90 (0.46 to 1.37)E 0.10 (−0.37 to 0.54)	—
PSQI Time 2	A 0.76 (0.04 to 1.54)E 0.24 (−0.54 to 0.96)	A 0.60 (0.38 to 0.80)E 0.40 (0.20 to 0.62)	A 1.05 (0.60 to 1.58)E −0.06 (−0.58 to 0.40)	—

MEQ = diurnal preference; PSQI = sleep quality; rA = additive genetic correlations; rE = nonshared environmental correlations; A = additive genetic influence; E = nonshared environmental influence. Analyses were performed on age- and sex-regressed data in Mx.

The lower part of Table 4 shows the proportions of the phenotypic associations accounted for by genetic and environmental influences. For example, the concurrent associations between diurnal preference and sleep quality at time 1 and time 2 were largely accounted for by genetic factors (A = 83%–1.05%). The longitudinal associations within and between phenotypes were also largely accounted for by genetic factors (60%–90%).

Discussion

Our results provide 5 main conclusions. First, both diurnal preference and sleep quality were relatively stable over time (i.e., both phenotypes exhibited moderate correlations across time). Second, at both time points, a preference for eveningness was associated with poorer sleep quality. Third, genetic influences contributed to both diurnal preference and sleep quality at both time points, although they were somewhat smaller for sleep quality at time 2 than time 1, with a compensatory increase in the importance of the nonshared environment. Fourth, there was moderate to substantial overlap in the genetic influences contributing to both diurnal preference and sleep quality concurrently and longitudinally. Fifth, concurrent and longitudinal associations between diurnal preference and sleep quality were largely accounted for by genetic effects, whereas the nonshared environment contributed to their differences. These results are discussed further below with reference to our initial aims set forth in our introduction.

Stability of Diurnal Preference and Sleep Quality over Time

The phenotypic correlations indicated substantial stability in diurnal preference from time 1 to time 2. This resonates with other studies that demonstrate temporal stability of diurnal preference using a range of self-report tools over short time periods (Boudebesse et al., 2013; Caci et al., 2000; Di Milia et al., 2004; Greenwood, 1994; Randler, 2009). Inspection of the means indicates a significantly greater tendency toward morningness from time 1 to time 2. Numerous studies demonstrate an age-related shift toward morningness across adulthood (Barclay et al., 2014; Carrier et al., 1997; Hur et al., 1998; Kramer et al., 1999; Paine et al., 2006; Taillard et al., 1999). Sleep quality also demonstrated some degree of longitudinal stability, although to a lesser extent—a finding that has also been observed in relation to the stability of sleep problems from early childhood through to adolescence (Gregory and O’Connor, 2002) and in relation to the stability of insomnia symptoms in adolescents over the course of 1 year (Roberts et al., 2008). Again, inspection of the means indicates that sleep quality improved, albeit marginally, across time. Although increasing age is typically associated with poorer subjective and objective sleep quality (for example, greater number of and duration of awakenings, decreased amount of SWS, and increases in lighter stage 1 and 2 sleep; Bliwise, 2005; Feinberg et al., 1967; Landolt et al., 1996; Van Cauter et al., 2000), our sample reflected young adulthood at both time points, and it is likely that declines in sleep quality are more relevant to different developmental periods (i.e., young, mid-, and late adulthood).

Associations between Diurnal Preference and Sleep Quality Concurrently and Longitudinally

As previously noted (Barclay et al., 2010a), at time 1 there was a significant association between diurnal preference and sleep quality, such that a greater tendency toward eveningness was associated with poorer sleep quality. This pattern was also demonstrated at time 2. We previously hypothesized that this association is explained by differences in the phase angle between intrinsic properties of the circadian system and wake time between morning and evening types. In essence, this means that evening types awaken closer to their body temperature nadir, when alertness is low (Barclay et al., 2010a). It is possible that low alertness upon awakening contributes to subjective feelings of having slept poorly. Our finding here of significant longitudinal associations, such that diurnal preference at time 1 predicted sleep quality at time 2 (and vice versa), suggests 2 additional possible mechanisms. The constraints of our social world impose restrictions on our sleep, forcing us to wake early to attend school/university or work. Given that evening types typically initiate sleep late, it is likely that their sleep duration is curtailed when social schedules are inflexible. Thus, a preference for eveningness at time 1 may contribute to poor sleep quality at time 2 from the accumulation of sleep debt over time. It would be beneficial for future research to address this proposition using longitudinal samples of diurnal preference and sleep quality/duration using the Munich Chronotype Questionnaire (Roenneberg et al., 2003), which accounts for differences in preference for daily activity on work and free days. Such examination would allow us to determine the extent to which the timing of sleep and its quality/duration is affected by social constraints. The converse association is a little harder to explain. Difficulty sleeping may lead one to delay bedtime in an attempt to avoid prolonged sleep-onset latency. Over time, this may result in a phase shift that delays the circadian rhythm, which is also reflected in (and maintained by) one’s subjective diurnal preference.

Stability of Genetic and Environmental Influences on Diurnal Preference and Sleep Quality, Within Phenotypes, over Time

The univariate estimates from the correlated factors model demonstrate that genetic influences accounted for about half the variability in both diurnal preference and sleep quality at time 1. These estimates have been presented previously (Barclay et al., 2010a) and are in line with estimates from numerous other studies of both phenotypes (Heath et al., 1990; Hur, 2007; Hur et al., 1998; Koskenvuo et al., 2007; Partinen et al., 1983; Vink et al., 2001). The longitudinal aspect of this study allowed us to identify the extent to which these effects contributed to stability in these phenotypes. We found that there was substantial genetic overlap for diurnal preference across time and sleep quality across time. This suggests that to a large extent, the same genes contribute to the stability of these phenotypes throughout young adulthood. A growing amount of evidence points to the importance of “clock” genes in predicting differences in the circadian rhythm as well as diurnal preference (e.g., Archer et al., 2003; Barclay et al., 2011; Carpen et al., 2005; Carpen et al., 2006; Ellis et al., 2009; Ferrante et al., 2015; Hu et al., 2016; Jones et al., 2007; Jones et al., 2016; Katzenberg et al., 1998), and particularly, those related to the serotonin system (among others) have been implicated in sleep quality, although inconsistently (Barclay et al., 2011; Barclay and Gregory, 2013; Brummett et al., 2007). Recent genome-wide association studies (GWAS) have pointed to many other genetic candidates implicated in these phenotypes (see Parsons, 2015, for a review). However, the fact that these genetic correlations were not absolute suggests that new genes may come into play (or that genes that previously were important do not continue to exert such a strong effect) during young adulthood to influence these phenotypes. Regarding the nonshared environment, overlap was substantially smaller for both phenotypes across time. Thus, while it is likely that there are some nonshared environmental factors that continue to contribute to diurnal preference over time, the majority are time specific. This is perhaps not surprising given that environmental factors that influence sleep are often likely to be transient. For example, living in a noisy household may hinder sleep, but this is unlikely to continue to affect sleep over time once the environmental factor has surpassed, which is likely in young adults who may move house or have a change in life circumstances during this time period.

Genetic and Environmental Overlap between Diurnal Preference and Sleep Quality Concurrently and Longitudinally

There was moderate to substantial overlap in the genetic influences contributing to diurnal preference and sleep quality concurrently and longitudinally, suggesting that to some extent, similar genes influence both phenotypes. These results fit with the “general genes hypothesis”—that the same genes often contribute to associated phenotypes (Eley, 1997; Lahey and Waldman, 2003; Plomin et al., 2016), highlighting that genes often have pleiotropic effects. Furthermore, genetic influences contributed to a substantial amount of the phenotypic correlations both concurrently and longitudinally. Nonshared environmental overlap between phenotypes concurrently and longitudinally was minimal if not nonexistent, highlighting that different environmental factors contribute to these phenotypes. Furthermore, environmental factors did not explain a substantial amount of the phenotypic correlations. Together, these results suggest that largely genetic factors are responsible for the associations between diurnal preference and sleep quality concurrently, and our longitudinal data point to a biological explanation of the influence of diurnal preference on later sleep quality and vice versa. In terms of our explanations for the phenotypic associations posed above, it is likely that genes indirectly contribute to the associations between these phenotypes by first influencing our behavior toward the timing of our sleep and our attributions of the sleep obtained. That said, recent data from a GWAS failed to find significant genetic overlap between a 1-item measure of diurnal preference and several sleep phenotypes including insomnia, sleep apnea, sleep need, sleeping soundly, and sweating during sleep (Hu et al., 2016). While these sleep phenotypes are likely to be associated with sleep quality, they are more specific in nature, which may explain the genetic heterogeneity for these variables and diurnal preference.

Limitations

Our findings should be considered in light of at least 4 limitations. First, our sample size was small, which reduces our power to decompose variance into genetic and environmental components. That said, twin studies focusing on sleep phenotypes (e.g., Sletten et al., 2013) and using analyses of this kind (e.g., Gregory, 2008) have used smaller samples than this one. As the first study of its kind to include measures of diurnal preference and sleep quality in a genetically informative longitudinal design, we provide a framework of initial findings that may be refined by more powerful future studies using similar designs.

Second, our short follow-up period (5 years) limits our ability to extrapolate our findings to age-related changes in the associations between diurnal preference and sleep quality over longer time frames. That said, our results provide important information about the mechanisms underlying the stability of these phenotypes within young adulthood. This is particularly important given that young adulthood is often a time of increased responsibility related to caring for and financially supporting a family, as well as developing a career, so disturbed sleep during this period may be particularly problematic and costly (Goldman-Mellor et al., 2014).

Third, unexpectedly, MZ twins exhibited a greater tendency toward eveningness than DZ twins and siblings. However, the effect sizes were small (Cohen’s d = 0.30 and 0.21 for differences between MZ vs. DZ, and MZ vs. siblings, respectively), and this finding was not replicated at time 2. In addition, responders at time 2 exhibited a greater tendency toward morningness than nonresponders. While the absolute difference in means between zygosity groups and responders versus nonresponders is worthy of consideration in future studies, this does not preclude the computation of the relative proportions of variance accounted for by genetic and environmental factors.

Fourth concerns the interdependency of the phenotypes under study. Some of the components of the PSQI, for example, sleep duration, are likely to be influenced by the combination of diurnal preference and social schedules, which together determine the circadian phase of sleep; the MEQ itself contains some items that pertain to sleep quality (i.e., “during the first half hour after having woken in the morning, how tired do you feel?”). This brings into question the extent to which diurnal preference and sleep quality can be regarded as separate phenotypes. It is possible that this interdependency explains the apparent genetic overlap observed. That said, the apparent moderate correlations, both concurrently and longitudinally, between these phenotypes argues against this point. However, the heterogeneity in the study population in both the timing and flexibility of their social schedules possibly obscures a greater association between these phenotypes that would be detectable only if social schedules were fixed.

It is further important to consider the more general limitations that are inherent in twin modeling (e.g., the equal-environments assumption, generalizability, assortative mating, zygosity determination; see Plomin et al., 2013), yet the importance of investigating etiological mechanisms using twin modeling far outweighs these limitations.

Implications

These findings show that, over time, a combination of both overlapping, and to a lesser extent different, genes appear to be important in maintaining diurnal preference and sleep quality over time. Thus, the genetic mechanisms underlying these phenotypes appear to change to some extent across this relatively short time frame. In relation to diurnal preference, the introduction of new genetic effects (or indeed the discontinuation of effects that were previously important) may reflect the circadian stabilization that occurs after the reorganization of the circadian system during adolescence (Hagenauer et al., 2009). Given that the circadian system continues to change across the life span, it is perhaps no surprise that different genes may come into play at particular developmental time points. Not only does the circadian system undergo vast changes during adolescence but so too does the homeostatic system. For example, it has repeatedly been demonstrated that SWS (a marker of homeostatic sleep pressure) declines across adolescence (Carskadon et al., 2004). It is likely that this change is at least in part genetically controlled, and so a specific set of genes may be implicated. By young adulthood, this dramatic decline in SWS may plateau, and again, different sets of genes may come into play to maintain sleep throughout this period. Not only do the present results support these explanations, but they also suggest that similar genes may be controlling the reorganization and relative stabilization of both the circadian and homeostatic systems beyond the transition from late adolescence to young adulthood.

The present findings also increase our theoretical understanding of the potential mechanisms underlying the association between a preference for eveningness and poor sleep quality. These associations were largely due to genetic factors, and a combination of similar and different genes can be implicated. While molecular genetic studies are making some headway in determining the specific genes underlying both diurnal preference and sleep quality, results are inconsistent (Barclay and Gregory, 2013; Gehrman et al., 2011; Parsons, 2015; Parsons et al., 2014). Discrepancies between studies may be due to the etiological factors contributing to these phenotypes changing over time. Demonstrating that the stability of these phenotypes over time is partially due to unique genetic and environmental influences has the potential to direct future molecular genetic studies by considering the possibility that some sources of influence may be time specific.

Footnotes

Declaration of Conflicting Interests

The authors declare no conflicts of interests.

Funding

Waves 1 to 3 of the G1219 study were supported by grants from the W T Grant Foundation, the University of London Central Research fund, and a Medical Research Council Training Fellowship and Career Development Award to Thalia C. Eley. Wave 4 of the G1219 study was supported by grants from the Economic and Social Research Council (RES-000-22-2206) and the Institute of Social Psychiatry (06/07–11) to Alice M. Gregory. Nicola L. Barclay was supported by an ESRC studentship. Wave 5 was supported by funds from Goldsmiths University of London to Alice M. Gregory.

Notes

References

Archer

Robilliard

Skene

Smits

Williams

Arendt

von Schantz

(2003) A length polymorphism in the circadian clock gene Per3 is linked to delayed sleep phase syndrome and extreme diurnal preference. Sleep 26:413-415.

Arnett

(2000) Emerging adulthood: a theory of development from the late teens through the twenties. Am Psychol 55:469-480.

Barclay

Eley

Buysse

Archer

Gregory

(2010a) Diurnal preference and sleep quality: same genes? A study of young adult twins. Chronobiol Int 27:278-296.

Barclay

Eley

Buysse

Rijsdijk

Gregory

(2010b) Genetic and environmental influences on different components of the ‘Pittsburgh Sleep Quality Index’ and their overlap. Sleep 33:659-668.

Barclay

Eley

Mill

Wong

CCY

Zavos

HMS

Archer

Gregory

(2011) Sleep quality and diurnal preference in a sample of young adults: associations with 5HTTLPR, PER3 and CLOCK 3111. Am J Med Genet B Neuropsychiatr Genet 156:681-690.

Barclay

Gehrman

Gregory

Eaves

Silberg

(2015) The heritability of insomnia progression during childhood/adolescence: results from a longitudinal twin study. Sleep 38:109-118.

Barclay

Gregory

(2013) Quantitative genetic research on sleep: a review of normal sleep, sleep disturbances and associated emotional, behavioural, and health-related difficulties. Sleep Med Rev 17:29-40.

Barclay

Watson

Goldberg

Buchwald

(2014) Moderation of the genetic and environmental influences on diurnal preference by age in a sample of adult twins. Chronobiol Int 31:222-231.

Bliwise

(2005) Normal aging. In: Kryger

Roth

Dement

, editors. Principles and practice of sleep medicine. Philadelphia (PA): Elsevier.

10.

Borbély

(1982) A two-process model of sleep regulation. Hum Neurobiol 1:195-204.

11.

Borbély

Achermann

(2011) Sleep homeostasis and models of sleep regulation. In: Kryger

Roth

Dement

, editors. Principles and practice of sleep medicine. St. Louis (MO): Elsevier Saunders.

12.

Boudebesse

Lajnef

Geoffroy

Bellivier

Nieto

Gard

Olie

Azorin

Kahn

Bougerol

. (2013) Chronotypes of bipolar patients in remission: validation of the French version of the circadian type inventory in the FACE-BD sample. Chronobiol Int 30:1042-1049.

13.

Brummett

Krystal

Ashley-Koch

Kuhn

Zuchner

Siegler

Barefoot

Ballard

Gwyther

Williams

(2007) Sleep quality varies as a function of 5-HTTLPR genotype and stress. Psychosom Med 69:621-624.

14.

Buysse

Reynolds

Monk

Berman

Kupfer

(1989) The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 28:192-213.

15.

Caci

Nadalet

Staccini

Myquel

Boyer

(2000) The composite scale of morningness: further psychometric properties and temporal stability. Eur Psychiatry 15:278-281.

16.

Carpen

Archer

Skene

Smits

von Schantz

(2005) A single-nucleotide polymorphism in the 5′-untranslated region of the hPER2 gene is associated with diurnal preference. J Sleep Res 14:293-297.

17.

Carpen

von Schantz

Smits

Skene

Archer

(2006) A silent polymorphism in the PER1 gene associates with extreme diurnal preference in humans. J Hum Genet 51:1122-1125.

18.

Carrier

Monk

Buysse

Kupfer

(1997) Sleep and morningness-eveningness in the ‘middle’ years of life (20-59y). J Sleep Res 6:230-237.

19.

Carskadon

Acebo

Jenni

(2004) Regulation of adolescent sleep. Ann NY Acad Sci 1021:276-291.

20.

Cohen

Dibble

Grawe

Pollin

(1975) Reliably separating identical from fraternal twins. Arch Gen Psychiatry 32:1371-1375.

21.

Dahl

Lewin

(2002) Pathways to adolescent health: sleep regulation and behavior. J Adolesc Health 31:175-184.

22.

Di Milia

Smith

Folkard

(2004) Refining the psychometric properties of the circadian type inventory. Pers Individ Dif 36:1953-1964.

23.

Drennan

Selby

Kripke

Kelsoe

Gillin

(1992) Morningness/eveningness is heritable. Abstr Soc Neurosci 196.

24.

Eley

(1997) General genes: a new theme in developmental psychopathology. Curr Dir Psychol Sci 6:90-95.

25.

Ellis

von Schantz

Jones

KHS

Archer

(2009) Association between specific diurnal preference questionnaire items and PER3 VNTR genotype. Chronobiol Int 26:464-473.

26.

Feinberg

Koresko

Heller

(1967) EEG sleep patterns as a function of normal and pathological aging in man. Psychiatry Res 5:107-144.

27.

Ferrante

Gellerman

Woods

Filipowicz

Jain

Bearden

Ingram

(2015) Diurnal preference predicts phase differences in expression of human peripheral circadian clock genes. J Circadian Rhythms 13:1-7.

28.

Gehrman

Byrne

Gillespie

Martin

(2011) Genetics of insomnia. Sleep Med Clin 6:191-202.

29.

Genderson

Rana

Panizzon

Grant

Toomey

Jacobson

Xian

Cronin-Golomb

Franz

Kremen

. (2013) Genetic and environmental influences on sleep quality in middle-aged men: a twin study. J Sleep Res 22:519-526.

30.

Goldman-Mellor

Gregory

Caspi

Harrington

Parsons

Poulton

Moffitt

(2014) Mental health antecedents of early midlife insomnia: evidence from a four-decade longitudinal study. Sleep 37:1767-1775.

31.

Greenwood

(1994) Long-term stability and psychometric properties of the Composite Scale of Morningness. Ergonomics 37:377-383.

32.

Gregory

(2008) A genetic decomposition of the association between parasomnias and dyssomnias in 8 year old twins. Arch Pediatr Adolesc Med 162:299-304.

33.

Gregory

O’Connor

(2002) Sleep problems in childhood: a longitudinal study of developmental change and association with behavioral problems. J Am Acad Child Adolesc Psychiatry 41:964-971.

34.

Gregory

Rijsdijk

Lau

JYF

Dahl

Eley

(2009) The direction of longitudinal associations between sleep problems and depression symptoms: a study of twins aged 8 and 10 years. Sleep 32:189-199.

35.

Hagenauer

Perryman

Lee

Carskadon

(2009) Adolescent changes in the homeostatic and circadian regulation of sleep. Dev Neurosci 31:276-284.

36.

Heath

Kendler

Eaves

Martin

(1990) Evidence for genetic influences on sleep disturbance and sleep pattern in twins. Sleep 13:318-335.

37.

Horne

Östberg

(1976) A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. Int J Chronobiol 4:97-110.

38.

Shmygelska

Tran

Eriksson

Tung

Hinds

(2016) GWAS of 89,283 individuals identifies genetic variants associated with self-reporting of being a morning person. Nat Commun 7:10449.

39.

Hur

(2007) Stability of genetic influence on morningness–eveningness: a cross-sectional examination of South Korean twins from preadolescence to young adulthood. J Sleep Res 16:17-23.

40.

Hur

Bouchard

Lykken

(1998) Genetic and environmental influence on morningness-eveningness. Pers Individ Dif 25:917-925.

41.

Jones

KHS

Ellis

Von Schantz

Skene

Dijk

Archer

(2007) Age-related change in the association between a polymorphism in the PER3 gene and preferred timing of sleep and waking activities. J Sleep Res 16:12-16.

42.

Jones

Tyrrell

Wood

Beaumont

Ruth

Tuke

Yaghootkar

Teder-Laving

Hayward

(2016) Genome-wide association analyses in >119,000 individuals identifies thirteen morningness and two sleep duration loci. bioR xiv:031369.

43.

Katzenberg

Young

Finn

Lin

King

Takahashi

Mignot

(1998) A CLOCK polymorphism associated with human diurnal preference. Sleep 21:569-576.

44.

Koskenvuo

Hublin

Partinen

Heikkila

Kaprio

(2007) Heritability of diurnal type: a nationwide study of 8753 adult twin pairs. J Sleep Res 16:156-162.

45.

Kramer

Kerkhof

Hofman

(1999) Age differences in sleep-wake behavior under natural conditions. Pers Individ Dif 27:853-860.

46.

Lack

Bailey

Lovato

Wright

(2009) Chronotype differences in circadian rhythms of temperature, melatonin, and sleepiness as measured in a modified constant routine protocol. Nat Sci Sleep 1:1-8.

47.

Lahey

Waldman

(2003) A developmental propensity model of the origins of conduct problems during childhood and adolescence. In: Lahey

Moffitt

Caspi

, editors. Causes of conduct disorder and juvenile delinquency. New York (NY): Guilford.

48.

Landolt

Dijk

Achermann

Borbély

(1996) Effects of age on the sleep EEG: slow-wave activity and spindle frequency in young and middle-aged men. Brain Res 738:205-212.

49.

McAdams

Gregory

Rowe

Zavos

HMS

Barclay

Lau

JYF

Maughan

Eley

(2013) The Genesis 12–19 (G1219) study: a twin and sibling study of gene–environment interplay and adolescent development in the UK. Twin Res Hum Gen 16:134-143.

50.

McGue

Bouchard

(1984) Adjustment of twin data for the effects of age and sex. Behav Genet 14:325-343.

51.

Megdal

Schernhammer

(2007) Correlates for poor sleepers in a Los Angeles high school. Sleep Med 9:60-63.

52.

Merikanto

Kronholm

Peltonen

Laatikainen

Lahti

Partonen

(2012) Relation of chronotype to sleep complaints in the general Finnish population. Chronobiol Int 29:311-317.

53.

Neale

(1997) Mx: statistical modeling. Richmond (VA): Virginia Commonwealth University Department of Psychiatry,.

54.

Paine

Gander

Travier

(2006) The epidemiology of morningness/eveningness: influence of age, gender, ethnicity, and socioeconomic factors in adults (30-49 years). J Biol Rhythms 21:68-76.

55.

Parsons

(2015) On the genetics of sleep disorders: genome-wide association studies and beyond. Adv Genomics Genet 5:293-303.

56.

Parsons

Lester

Barclay

Archer

Nolan

Eley

Gregory

(2014) Polymorphisms in the circadian expressed genes PER3 and ARNTL2 are associated with diurnal preference and GNβ3 with sleep measures. J Sleep Res 23:595-604.

57.

Partinen

Kaprio

Koskenvuo

Putoken

Langinvainio

(1983) Genetic and environmental determination of human sleep. Sleep 6:179-185.

58.

Plomin

DeFries

Knopik

Neiderhiser

(2016) Top 10 replicated findings from behavioral genetics. Perspect Psychol Sci 11:3-23.

59.

Plomin

DeFries

McClearn

McGuffin

(2013) Behavioural genetics. New York (NY): Worth Publishers.

60.

Portaluppi

Smolensky

Touitou

(2010) Ethics and methods for biological rhythm research on animals and human beings. Chronobiol Int 27:1911-1929.

61.

Randler

(2009) Validation of the full and reduced Composite Scale of Morningness. Biol Rhythm Res 40:413-423.

62.

Roberts

Chan

(2008) Persistence and change in symptoms of insomnia among adolescents. Sleep 31:177-184.

63.

Roenneberg

Wirz-Justice

Merrow

(2003) Life between clocks: daily temporal patterns of human chronotypes. J Biol Rhythms 18:80-90.

64.

Sletten

Rajaratnam

Wright

Zhu

Naismith

Martin

Hickie

(2013) Genetic and environmental contributions to sleep-wake behavior in 12-year-old twins. Sleep 36:1715-1722.

65.

Soehner

Kennedy

Monk

(2011) Circadian preference and sleep-wake regularity: associations with self-report sleep parameters in daytime-working adults. Chronobiol Int 28:802-809.

66.

Taillard

Philip

Bioulac

(1999) Morningness/eveningness and the need for sleep. J Sleep Res 8:291-295.

67.

Tzischinsky

Shochat

(2011) Eveningness, sleep patterns, daytime functioning, and quality of life in Israeli adolescents. Chronobiol Int 28:338-343.

68.

Van Cauter

Leproult

Plat

(2000) Age-related changes in slow wave sleep and REM sleep and relationship with growth hormone and cortisol levels in healthy men. JAMA 284:861-868.

69.

Vink

Groot

Kerkhof

Boomsma

(2001) Genetic analysis of morningness and eveningness. Chronobiol Int 18:809-822.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB