Sleep and twins
J Sleep Res. (2014) 23, 153–158
Sleep duration and personality in Croatian twins A N A B U T K O V I C , T E N A V U K A S O V I C and D E N I S B R A T K O Department of Psychology, Faculty of Humanities and Social Sciences, University of Zagreb, Zagreb, Croatia
Keywords five-factor personality model, heritability, sleep length, twin study Correspondence Ana Butkovic, PhD, Department of Psychology, Faculty of Humanities and Social Sciences, University of Zagreb, Ivana Lucica 3, 10 000 Zagreb, Croatia. Tel.: +385-(1)-612-01-94; fax: +385-(1)-612-00-37; e-mail: [email protected] Accepted in revised form 5 September 2013; received 14 May 2013 DOI: 10.1111/jsr.12101
SUMMARY
The objective of this study was to examine which genetic and environmental influences contribute to individual differences in sleep duration in a sample of Croatian adolescent/early adult twins, as well as to investigate the relationship between personality and sleep duration. Participants included 339 twin pairs (105 monozygotic and 234 dizygotic) aged between 15 and 22 years. They reported on their average sleep duration and personality. The broad heritability estimate (additive and non-additive genetic influences) for sleep duration was 0.63, while personality estimates ranged between 0.47 and 0.62. Significant negative phenotypic associations with neuroticism and openness were mainly genetically mediated 100 and 80%, respectively. Only 6% of the sleep duration variance was explained by genetic influences shared with neuroticism and openness. In regression analysis, age, gender and five personality traits explained 5% of sleep duration variance, with neuroticism and openness as significant predictors. Comparison of short, moderate and long sleepers showed that participants in the short sleepers group had significantly higher neuroticism scores than groups of moderate and long sleepers, as well as a significantly higher openness score than the group of long sleepers. This indicates that personality traits of neuroticism and openness contribute to the prediction of sleep duration due to overlapping genetic influences that contribute to both these personality traits and sleep duration. However, as phenotypic overlap of personality and sleep duration is relatively weak, heritability of sleep duration is not only related to individual differences in personality traits, so future research needs to examine other phenotypic correlates of sleep duration.
INTRODUCTION Sleep is important for normal human functioning, but little is still known about its underlying mechanisms, interactions with different conditions and long-term effects on health and life. Changes in modern society, which include longer work hours and shiftwork, have also impacted upon the duration of sleep for many people. Research has shown that lack of sleep is associated with metabolic, endocrine and immune consequences, including the development of metabolic syndromes affecting both children and adults (AlDabal and BaHammam, 2011), as well as changes in hormone levels and in parameters associated with immune response (MaurovichHorvat et al., 2008). Two recent meta-analyses have indicated that sleep duration is associated with increased mortality. Cappuccio et al. (2010) have examined the association between sleep ª 2013 European Sleep Research Society
duration and all-cause mortality from prospective studies, and found that both short (12%) and long (30%) duration of sleep are associated with greater risk of death in prospective population studies. Gallicchio and Kalesan (2009) have examined the associations between sleep duration and allcause and cause-specific mortality. The results show that there is a 10% increased risk of death in short sleepers and 23% in long sleepers. They also found an increased risk of cardiovascular-related (38%) and cancer-related (21%) death in long sleepers. Accumulating evidence indicates that sleep duration is important for our health, but specific mechanisms behind the association between both short and long sleep duration and mortality and morbidity are as yet unclear. Another important question is why people differ in their sleep duration. Two large twin studies have indicated that individual differences in sleep duration could be attributed to both genetic and
153
154
A. Butkovic et al.
environmental influences. Partinen et al. (1983) have compared self-reported sleep length and sleep quality in monozygotic (MZ) and dizygotic (DZ) twin pairs from the Finnish Twin Cohort Study, and found a heritability estimate for sleep length of 0.44. Recently, Hublin et al. (2013) investigated the contribution of genetic factors to stability and change of sleep length among adults over a 15-year period using the Finnish Twin Cohort. They found heritability of approximately 0.30 for sleep length that was quite stable across three measurement points and high genetic correlations, from 0.76 to 0.93, between three measurement points. Heath et al. (1990) examined the aetiological role of genetic and environmental influences in sleep pattern (daytime napping, habitual bedtime and sleep duration), subjective sleep quality and sleep disturbance using a general population sample of adult Australian twin pairs aged 17–88 years. Broad heritability estimates, which include both additive genetic influences or direct effects of genes and non-additive genetic influences or gene interactions, of the sleep pattern variables were remarkably consistent, ranging between 38 and 46%. For sleep duration specifically, the broad heritability estimate was 0.40, with evidence for non-additive genetic influences being larger than additive genetic influences (0.31 versus 0.09). However, not all behavioural–genetic studies of sleep duration have found evidence of the importance of genetic influences. Barclay et al. (2010) have collected data from a UK sample of twins and siblings on the Pittsburgh Sleep Quality Index and found that genetic factors were important for all sleep variables except sleep duration. Because 90% of the sample was aged 18–22 years, it could be that studying at the university may have attenuated the impact of genetic influences on sleep length as a possible environmental pressure. Another study, examining associations and differences in sleep–wake patterns among children and their parents from Taiwan, found no correlation between nocturnal sleep duration of children and parents (Gau and Merikangas, 2004). The relationship between personality and sleep duration was investigated in a number of studies, but the results were inconsistent. Studies differed considerably regarding age and composition of the sample, as well as personality and sleep measures. In early studies, Webb and Friel (1971) found no personality differences between short and long sleepers, while Hartmann (1973) concluded, based on clinical observations, that short sleepers were ‘non-worriers’ and long sleepers were ‘worriers’. However, two studies that followed showed that anxiety (a facet of neuroticism) is higher in short than long sleepers (Hicks and Pellegrini, 1977; Kumar and Vaidya, 1984). In more recent studies, two found no significant relationship between personality and sleep duration using different personality measures (Gray and Watson, 2002; Soehner et al., 2007), one found a significant positive relationship between sleep duration on weekdays and agreeableness and conscientiousness (Randler, 2008) and two studies using different personality measures found that people with high neuroticism had shorter sleep duration (Gau, 2000; Vincent et al., 2009).
The aim of this study was to examine which genetic and environmental influences contribute to individual differences in sleep duration in a sample of Croatian adolescent/early adult twins, as well as to investigate the relationship between personality and sleep duration. The use of the twin method makes it possible to examine genetic and environmental influences on the phenotypic associations between personality and sleep duration. Based on previous studies, we hypothesize that individual differences in sleep duration will be attributed to both genetic and environmental influences, that sleep duration will be correlated negatively on a phenotypic level with neuroticism and that this phenotypic association will be explained mainly by overlapping genetic factors. As for the other four personality factors, only one study found a positive correlation with agreeableness and conscientiousness, so we hypothesize that the other four personality factors will not be correlated significantly with sleep duration. Finally, we hypothesize that people with shorter (≤6 h) and longer (≥9 h) sleep duration will have a significantly different neuroticism score, in the direction of people with shorter sleep duration having higher neuroticism scores. METHODS The sample used in this research was formed in 2007 based on the register of citizens from the Zagreb (capital of Croatia) area. From that register, twin pairs born between 1985 and 1992 were identified. Data were collected via mail. From 2005 contacted individuals, 732 (36.5%) returned completed questionnaires. From those individuals, 339 twin pairs were identified (105 MZ and 234 DZ). Their age varied between 15 and 22 years [mean = 18.62, standard deviation (SD) = 2.31]. Zygosity was determined by a questionnaire constructed for the purpose of this research. It included 11 items, six evaluating physical similarities (e.g. facial appearance, hair colour) and five evaluating twin confusion by parents, other family members, teachers, casual friends and strangers. These items have been shown to be valid indicators of zygosity in a number of studies and have been used in different zygosity questionnaires (e.g. Chen et al., 1999; Goldsmith, 1991). Twins’ self-reports about their zygosity were coded so that the lower score indicated monozygosity. In a number of studies (e.g. Reed et al., 2005; Spitz et al., 1996), the use of questionnaires for zygosity determination has been shown to be accurate to approximately 95%. Sleep duration was measured with one item: ‘How long do you sleep on average every night?’. Responses ranged between 4 and 12 h per night (mean = 7.74; SD = 1.17); sleep duration did not differ between MZ and DZ twins (7.81 versus 7.71 h; t = 0.951, P = 0.342). We divided our participants into three groups in terms of sleep duration: those sleeping ≤6 h (n = 72, or 10%), from 6.5 to 8.5 h (n = 466, or 70%) and ≥9 h (n = 133, or 20%). In each group there were approximately twice as many DZ than MZ twins, which is comparable to the composition of the total sample. ª 2013 European Sleep Research Society
Sleep duration and personality in Croatian twins Personality traits were measured with the NEO-Five Factor Inventory (NEO-FFI), a short version of the NEO-Personality Inventory (Costa and McCrae, 1992). NEO-FFI has 60 items and measures neuroticism (N), extraversion (E), openness (O), agreeableness (A) and conscientiousness (C). Cronbach’s a coefficients were 0.81, 0.72, 0.57, 0.66 and 0.81 for neuroticism, extraversion, openness, agreeableness and conscientiousness, respectively. RESULTS Descriptive statistics are presented in Table 1. Because multivariate analysis of variance (MANOVA) showed some sex and age effects in our sample, age and sex were regressed from the scores before genetic model-fitting using an adjustment procedure proposed by McGue and Bouchard (1984). In our sample there were significant sex differences in neuroticism (F1,651 = 51.09, P < 0.001), with females scoring higher than males (MF = 22.02 versus MM = 17.14). For age, significant differences were found in conscientiousness (F7,651 = 5.76, P < 0.001), which increased with age. Twin intraclass correlations were also calculated, and represent the proportion of total variance due to variance between pairs and are a more appropriate measure of twin pair similarity than Pearson’s r. The pattern of MZ–DZ correlations indicates the relative importance of different sources of variance. The structural equation model-fitting program Mx was used for genetic model-fitting analyses (Neale et al., 2003). The goodness-of-fit of each genetic model was measured relative to the phenotypic (saturated) model. Based on the obtained pattern of MZ–DZ correlations shown in Table 1, the first tested model included additive genetic effects (A), nonadditive genetic (D) or shared environmental effects (C) and non-shared environmental effects (E). A series of nested models were run to test if any of the parameters could be equated or dropped from the model without significant worsening of fit. The results of the univariate analyses are presented in Table 2. As our sample composition and size has a limited power to detect non-additive genetic (D) or shared environmental (C)
Table 1 Descriptive statistics for sleep duration and personality traits Males
Females
Intraclass r
Measures
Mean
SD
Mean
SD
MZ
DZ
Sleep duration Neuroticism Extraversion Openness Agreeableness Conscientiousness
7.76 17.14 30.29 23.04 29.82 30.59
1.15 7.47 6.55 6.06 6.30 7.48
7.73 22.02 30.48 24.10 30.67 32.06
1.18 8.77 6.72 6.37 5.98 7.37
0.62 0.64 0.52 0.67 0.50 0.50
0.20 0.21 0.12 0.25 0.27 0.27
MZ, monozygotic twins, DZ, dizygotic twins; SD, standard deviation.
ª 2013 European Sleep Research Society
155
influences, which is also apparent from large confidence intervals of our parameter estimates reported in Table 2, we report the fit and parameter estimates from full ADE or ACE models. The results indicate that the broad heritability estimate, which includes both additive genetic (A) and non-additive genetic (D) influences, is 0.63 for sleep duration in our sample, with non-shared environmental influences explaining the rest of the variance. As for personality, broad heritability estimates range between 0.47 and 0.62, which is in line with previous findings on heritability of personality (Plomin et al., 2008). Phenotypic correlations between sleep duration and personality traits, as well as results from bivariate analyses, are reported in Table 3. As can be seen, sleep duration was correlated negatively with neuroticism (r = 0.14, P = 0.001) and openness (r = 0.16, P < 0.001). In order to assess how much these phenotypic relations are due to overlapping genetic and environmental influences, bivariate analyses were computed using bivariate Cholesky decomposition, which was then transformed into a correlated factors model to enable easier interpretation (Loehlin, 1996). Our models included additive genetic (A) and non-shared environmental effects (E), based on the results of univariate analyses. A significant genetic correlation of 0.26 was obtained between sleep duration and neuroticism, and 0.22 between sleep duration and openness. By comparing bivariate heritabilities (products of genetic correlation and the square root of each univariate heritability) with phenotypic correlations, a percentage of each phenotypic correlation explained by shared genetic influences was obtained and indicated that a whole phenotypic association with neuroticism is genetically mediated as well as 80% of the phenotypic association with openness. We also ran trivariate Cholesky decompositions, which included both personality traits (neuroticism and openness) in varying order and sleep duration. The results of these analyses were in line with bivariate analyses and the same genetic correlations were obtained. Individual differences in sleep duration were explained mainly by specific genetic (52%) and environmental (42%) influences, with approximately 6% of the variance explained by genetic influences shared with neuroticism and openness. Next, we compared neuroticism and openness personality scores in three groups of participants divided by the length of their sleep in groups of short sleepers (≤6 h), moderate sleepers (6.5–8.5 h) and long sleepers (≥9 h) using analysis of variance (ANOVA). As we had twins in our sample, we adjusted degrees of freedom to calculate the F-value significance by dividing them by two (McGue et al., 1993). The results were significant for both neuroticism (F2,326 = 7.05, P = 0.001) and openness (F2,326 = 6.55, P = 0.002). A post test showed that participants in the short hoc Scheffe sleepers group had significantly higher neuroticism scores than groups of moderate and long sleepers (23.28 versus 19.84 and 18.62). Also, the short sleepers group had a significantly higher openness score than the long sleepers group (25.56 versus 22.28).
156
A. Butkovic et al.
Table 2 Univariate model-fitting results for sleep duration and personality traits: fit statistics for saturated and full model with parameter estimates for the full model Measures
Model
Sleep duration N
Full Full Full Full Full Full Full Full Full Full Full Full
E O A C
saturated saturated saturated saturated saturated saturated
2LL (df) 2029.771 2035.393 4564.541 4575.603 4313.221 4316.864 4186.212 4191.805 4190.515 4198.763 4425.226 4431.062
(661) (666) (648) (653) (648) (653) (648) (653) (648) (653) (648) (653)
v2(df)
P
A
D
5.622 (5)
0.35
0.16 (0.00–0.51)
0.47 (0.00–0.71)
0.37 (0.28–0.50)
6.026 (5)
0.30
0.23 (0.00–0.66)
0.39 (0.00–0.71)
0.38 (0.28–0.49)
3.643 (5)
0.60
0.00 (0.00–0.51)
0.50 (0.00–0.62)
0.50 (0.38–0.65)
5.593 (5)
0.35
0.44 (0.00–0.69)
0.18 (0.00–0.68)
0.38 (0.29–0.50)
8.248 (5)
0.14
0.56 (0.27–0.66)
0.00 (0.00–0.19)
0.44 (0.34–0.59)
5.836 (5)
0.32
0.47 (0.10–0.62)
0.03 (0.00–0.30)
0.50 (0.38–0.65)
C
E
2LL, minus twice the log-likelihood of the data; df, degrees of freedom; v (df), 2LL (and df) difference between current and previous model; A, additive genetic variance; D, non-additive genetic variance; C, shared environmental variance; E, non-shared environmental variance; 95% confidence intervals shown in parentheses. 2
Table 3 Phenotypic correlations between sleep duration and personality traits and bivariate model-fitting results: fit statistics and correlation estimates for correlated factors AE model
Sleep duration with
Phenotypic correlation
2LL
df
Correlation estimates
Bivariate heritability
Neuroticism
0.14*
6603.217
1321
rA = 0.26 ( 0.44 to 0.08) rE = 0.03 ( 0.16 to 0.21)
0.15
Extraversion Openness
0.04 0.16*
6221.737
1321
rA = rE =
0.13
Agreeableness Conscientiousness
0.01 0.01
0.22 ( 0.41 to 0.04) 0.01 ( 0.19 to 0.17)
2LL, minus twice the log-likelihood of the data; df, degrees of freedom; rA, genetic correlation; rE, non-shared environmental correlation; 95% confidence intervals shown in parentheses, significant genetic and environmental correlations are shown in bold type. All variance estimates from these analyses were within 0.06 of the univariate estimates given in Table 2. *P < 0.01.
Finally, in order to assess if sleep duration can be predicted from the age and sex demographic variables and personality traits, we ran hierarchical regression analyses with age and sex entered into the first step, and personality traits into the second step. The results are presented in Table 4. Demographic variables have not explained any of the sleep duration variance, but both neuroticism (b = 0.17, P < 0.001) and openness (b = 0.16, P < 0.001) were significant predictors, with the whole model explaining 5% of sleep duration variance. DISCUSSION The first contribution of this study is that, in a sample of Croatian twins, it estimates the genetic and environmental influences which contribute to individual differences in sleep duration. In line with studies investigating the heritability of sleep duration in Finnish and Australian twins (Heath et al., 1990; Hublin et al., 2013; Partinen et al., 1983), we found
Table 4 Results of the hierarchical regression analysis Predictors Model 1 Age Gender Model 2 Age Gender Neuroticism Extraversion Openness Agreeableness Conscientiousness
Beta
t
Summary of the model
0.09 0.02
2.28 0.38
Adjusted R² = 0.01 F2,324 = 2.67
0.08 0.06 0.17* 0.03 0.16* 0.06 0.03
2.03 1.33 3.81 0.62 3.95 1.50 0.71
Adjusted R² = 0.05 F7,321 = 5.17* DR² = 0.04 FDR²(5,321) = 6.13*
*P < 0.01.
that genetic influences contribute to individual differences in sleep duration. Our heritability estimate is somewhat higher than in previous studies (63% versus 44%, 40%, approximately ª 2013 European Sleep Research Society
Sleep duration and personality in Croatian twins 30%), but with large confidence intervals. Also in line with Australian results, our results indicate that both additive and non-additive genetic influences might be important for individual differences in sleep duration. If non-additive genetic influences are important for sleep duration, this could explain why Gau and Merikangas (2004) found no parent–offspring similarity in sleep duration in a Taiwanese sample, as parents and offspring share only additive genetic influences. The main focus of this study was to estimate putative overlapping genetic influences of personality and sleep duration. This is a new and original aspect of the study, because it goes beyond correlational analysis between sleep duration and personality on a phenotypic level and examines the genetic and environmental factors contributing to those associations. However, as a sample size of thousands of twins, which is difficult to recruit, is needed to obtain precise estimates, the observed results have large confidence intervals. This should be taken into account when interpreting the results. Therefore, the second contribution of this study is that it explores the relationship between personality and sleep duration in a genetically informative sample which makes it possible to examine genetic and environmental influences on these phenotypic associations. In line with our hypothesis and previous studies (Gau, 2000; Vincent et al., 2009), we found a significant negative phenotypic association between sleep duration and neuroticism. We also found a significant negative phenotypic association between sleep duration and openness which was not found in previous studies exploring the relationship between personality and sleep duration. Because six of 12 NEO-FFI openness items indicate intellectual curiosity and appreciation of art and beauty, it seems that those individuals who, for example, enjoy and spend a great deal of time thinking and reading about different ideas are sleeping less. This finding is in line with Geiger et al.’s (2010) finding that there is a negative association between sleep duration and measures of intelligence. Also in line with our hypothesis, we found a significant difference in neuroticism score between groups of people sleeping ≤6 h, which had the highest neuroticism score, and groups of people sleeping between 6.5 and 8.5 h and ≥9 h, that did not differ in neuroticism score. We did not have a hypothesis regarding openness, but as there was a significant phenotypic correlation we also compared three sleep groups for openness score, and found that people who sleep ≤6 h have a significantly higher openness score than people sleeping ≥9 h. Bivariate and trivariate Cholesky decompositions indicate that phenotypic associations between personality and sleep duration are due mainly to overlapping genetic factors. Results of the trivariate analysis show that approximately 6% of the sleep duration variance is accounted for by genetic influences shared with neuroticism and openness. This percentage is similar to the amount of sleep duration variance explained by personality in the regression analysis, which was 5%. Taken together, this indicates that personality traits of neuroticism and openness contribute to the prediction of ª 2013 European Sleep Research Society
157
sleep duration due to overlapping genetic influences that contribute to both these personality traits and sleep duration. Sleep duration is influenced by a number of variables, and personality traits account for only a small proportion of sleep duration variance. However, as sleep duration has implications for different medical conditions such as diabetes or cardiovascular disease, as well as for increased mortality, it seems important to find stable psychological characteristics that are related to sleep duration. In this way, our study contributes to the understanding of which personality traits are related to sleep duration and why. The results of some studies that used different sleep variables indicate that personality might be related even more to sleep quality than to sleep duration (Gray and Watson, 2002; Soehner et al., 2007). Of course, there are limitations to our study. The specific age composition of our sample has strengths, but we do not know if our results would be the same in older twin cohorts. Heritability estimates, as well as relations with personality traits, might differ for sleep duration among different age groups. Sleep duration was measured with only one selfreport item, which asked for average amount of sleep per night. People usually sleep for different amounts of time on weekdays and at weekends. Also, an amount of other sleep information could be collected to understand more clearly the relationship between sleep and personality, such as selfreported bedtime and rise time on weekdays and at weekends, sleep habits, sleep problems and sleep quality. However, even with this limited information about sleep we can see that individual differences in sleep duration exist, and that they can be explained significantly by genetic and non-shared environmental differences. Also, in line with some other studies, we found that short sleepers have a higher neuroticism score. In a slightly younger adolescent sample (aged 12–15 years) it was shown that a high neuroticism score is associated with going to bed late on school days, short sleep duration, different sleep habits, sleep problems and impaired daytime function in comparison with a low neuroticism score (Gau, 2000). Ramsawh et al. (2011) investigated the relationship between childhood adversity and adult sleep quality in college students and found a significant relationship between childhood adversity and adult sleep quality which was mediated by neuroticism. The authors conclude that neuroticism may represent a potential target for change in future insomnia interventions. Because, in our study, neuroticism was related to sleep duration both on phenotypic and genetic levels, and as both neuroticism and sleep duration have been found in previous studies to be related to mortality and health problems, a future direction could be to design interventions to change neuroticism in order to diminish different sleeping problems. This is in line with novel ideas in personality psychology that changing personality traits should be the focus of interventions, as personality trait change can improve people’s lives (Roberts, 2012). Therefore, we believe that irrespective of the sample size limitations, the observed results are of clinical relevance.
158
A. Butkovic et al.
AUTHOR CONTRIBUTIONS AB contributed to study design, data collection, data analysis, interpretation of results and preparation of the paper; TV contributed to data analysis, interpretation of results and preparation of the paper; DB contributed to study design, data collection, interpretation of results and preparation of the paper. CONFLICTS OF INTEREST No conflicts of interest declared. REFERENCES AlDabal, L. and BaHammam, A. S. Metabolic, endocrine, and immune consequences of sleep deprivation. Open Respir. Med. J., 2011, 5: 31–43. Barclay, N. L., Eley, T. C., Buysse, D. J., Rijsdijk, F. V. and Gregory, A. M. Genetic and environmental influences on different components of the Pittsburgh Sleep Quality Index and their overlap. Sleep, 2010, 33: 659–668. Cappuccio, F. P., D’Elia, L., Strazzullo, P. and Miller, M. A. Sleep duration and all-cause mortality: a systematic review and metaanalysis of prospective studies. Sleep, 2010, 33: 585–592. Chen, W. J., Chang, H.-W., Wu, M.-Z. et al. Diagnosis of zygosity by questionnaire and polymarker polymerase chain reaction in young twins. Behav. Genet., 1999, 29: 115–123. Costa, P. T. Jr and McCrae, R. R. Revised NEO Personality Inventory (NEO PI-R) and the NEO Five-Factor Inventory (NEOFFI) Professional Manual. Psychological Assessment Resources, Odessa, FL, 1992. Gallicchio, L. and Kalesan, B. Sleep duration and mortality: a systematic review and meta-analysis. J. Sleep Res., 2009, 18: 148–158. Gau, S. Neuroticism and sleep-related problems in adolescence. Sleep, 2000, 23: 1–8. Gau, S. S. and Merikangas, K. R. Similarities and differences in sleep–wake patterns among adults and their children. Sleep, 2004, 27: 299–304. Geiger, A., Achermann, P. and Jenni, O. G. Association between sleep duration and intelligence scores in healthy children. Dev. Psychol., 2010, 46: 949–954. Goldsmith, H. H. A zygosity questionnaire for young twins: a research note. Behav. Genet., 1991, 21: 257–269. Gray, E. K. and Watson, D. General and specific traits of personality and their relation to sleep and academic performance. J. Pers., 2002, 70: 177–206. Hartmann, E. Sleep requirement: long sleepers, short sleepers, variable sleepers, and insomniacs. Psychosomatics, 1973, 14: 95–103. Heath, A. C., Kendler, K. S., Eaves, L. J. and Martin, N. G. Evidence for genetic influences on sleep disturbance and sleep pattern in twins. Sleep, 1990, 13: 318–335.
Hicks, R. A. and Pellegrini, R. J. Anxiety levels of short and long sleepers. Psychol. Rep., 1977, 41: 569–570. Hublin, C., Partinen, M., Koskenvuo, M. and Kaprio, J. Genetic factors in evolution of sleep length—a longitudinal twin study in Finnish adults. J. Sleep Res., 2013, 22: 513–518. Kumar, A. and Vaidya, A. K. Anxiety as a personality dimension of short and long sleepers. J. Clin. Psychol., 1984, 40: 197–198. Loehlin, J. C. The Cholesky approach: a cautionary note. Behav. Genet., 1996, 26: 65–69. €cher, T. and Sonka, Maurovich-Horvat, E., Pollma K. The effects of sleep and sleep deprivation on metabolic, endocrine and immune parameters. Prague Med. Rep., 2008, 109: 275–285. McGue, M. and Bouchard, T. J. Adjustment of twin data for the effects of age and sex. Behav. Genet., 1984, 14: 325–343. McGue, M., Bacon, S. and Lykken, D. Personality stability and change in early adulthood: a behavioral genetic analysis. Dev. Psychol., 1993, 29: 96–109. Neale, M. C., Boker, S. M., Xie, G. and Maes, H. H. M.. Mx: Statistical Modelling, 6th edn. Department of Psychiatry, Richmond, VA, 2003. Partinen, M., Kaprio, J., Koskenvuo, M., Putkonen, P. and Langinvainio, H. Genetic and environmental determination of human sleep. Sleep, 1983, 6: 179–185. Plomin, R., DeFries, J., McClearn, G. E. and McGuffin, P. Behavioral Genetics, 5th edn. Freeman, New York, 2008. Ramsawh, H. J., Ancoli-Israel, S., Sullivan, S. G., Hitchcock, C. A. and Stein, M. B. Neuroticism mediates the relationship between childhood adversity and adult sleep quality. Behav. Sleep Med., 2011, 9: 130–143. Randler, C. Morningness–eveningness, sleep–wake variables and big five personality factors. Pers. Individ. Dif., 2008, 45: 191–196. Reed, T., Plassman, B. L., Tanner, C. M., Dick, D. M., Rinehart, S. A. and Nichols, W. C. Verification of self-report of zygosity determined via DNA testing in a subset of the NAS–NRC Twin Registry 40 later. Twin Res. Hum. Genet., 2005, 8: 362–367. Roberts, B. W. Personality trait change and improving people’s lives: changing personality traits should be the focus of interventions [Abstract]. In: L. Di Blas, A. Carnaghi, D. Ferrante and V. Piccoli (Eds) Book of Abstracts: 16th European Conference on Personality Psychology. Universita Degli Studi di Trieste, Trieste, 2012: 17. Soehner, A. M., Kennedy, K. S. and Monk, T. H. Personality correlates with sleep–wake variables. Chronobiol. Int., 2007, 24: 889–903. Spitz, E., Moutier, R., Reed, T. et al. Comparative diagnoses of twin zygosity by SSLP variant analysis, questionnaire, and dermatoglyphic analysis. Behav. Genet., 1996, 26: 55–63. Vincent, N., Cox, B. and Clara, I. Are personality dimensions associated with sleep length in a large nationally representative sample? Compr. Psychiatry, 2009, 50: 158–163. Webb, W. B. and Friel, J. Sleep stage and personality characteristics of ‘natural’ long and short sleepers. Science, 1971, 171: 587–588.
ª 2013 European Sleep Research Society
J Sleep Res. (2014) 23, 153–158
Sleep duration and personality in Croatian twins A N A B U T K O V I C , T E N A V U K A S O V I C and D E N I S B R A T K O Department of Psychology, Faculty of Humanities and Social Sciences, University of Zagreb, Zagreb, Croatia
Keywords five-factor personality model, heritability, sleep length, twin study Correspondence Ana Butkovic, PhD, Department of Psychology, Faculty of Humanities and Social Sciences, University of Zagreb, Ivana Lucica 3, 10 000 Zagreb, Croatia. Tel.: +385-(1)-612-01-94; fax: +385-(1)-612-00-37; e-mail: [email protected] Accepted in revised form 5 September 2013; received 14 May 2013 DOI: 10.1111/jsr.12101
SUMMARY
The objective of this study was to examine which genetic and environmental influences contribute to individual differences in sleep duration in a sample of Croatian adolescent/early adult twins, as well as to investigate the relationship between personality and sleep duration. Participants included 339 twin pairs (105 monozygotic and 234 dizygotic) aged between 15 and 22 years. They reported on their average sleep duration and personality. The broad heritability estimate (additive and non-additive genetic influences) for sleep duration was 0.63, while personality estimates ranged between 0.47 and 0.62. Significant negative phenotypic associations with neuroticism and openness were mainly genetically mediated 100 and 80%, respectively. Only 6% of the sleep duration variance was explained by genetic influences shared with neuroticism and openness. In regression analysis, age, gender and five personality traits explained 5% of sleep duration variance, with neuroticism and openness as significant predictors. Comparison of short, moderate and long sleepers showed that participants in the short sleepers group had significantly higher neuroticism scores than groups of moderate and long sleepers, as well as a significantly higher openness score than the group of long sleepers. This indicates that personality traits of neuroticism and openness contribute to the prediction of sleep duration due to overlapping genetic influences that contribute to both these personality traits and sleep duration. However, as phenotypic overlap of personality and sleep duration is relatively weak, heritability of sleep duration is not only related to individual differences in personality traits, so future research needs to examine other phenotypic correlates of sleep duration.
INTRODUCTION Sleep is important for normal human functioning, but little is still known about its underlying mechanisms, interactions with different conditions and long-term effects on health and life. Changes in modern society, which include longer work hours and shiftwork, have also impacted upon the duration of sleep for many people. Research has shown that lack of sleep is associated with metabolic, endocrine and immune consequences, including the development of metabolic syndromes affecting both children and adults (AlDabal and BaHammam, 2011), as well as changes in hormone levels and in parameters associated with immune response (MaurovichHorvat et al., 2008). Two recent meta-analyses have indicated that sleep duration is associated with increased mortality. Cappuccio et al. (2010) have examined the association between sleep ª 2013 European Sleep Research Society
duration and all-cause mortality from prospective studies, and found that both short (12%) and long (30%) duration of sleep are associated with greater risk of death in prospective population studies. Gallicchio and Kalesan (2009) have examined the associations between sleep duration and allcause and cause-specific mortality. The results show that there is a 10% increased risk of death in short sleepers and 23% in long sleepers. They also found an increased risk of cardiovascular-related (38%) and cancer-related (21%) death in long sleepers. Accumulating evidence indicates that sleep duration is important for our health, but specific mechanisms behind the association between both short and long sleep duration and mortality and morbidity are as yet unclear. Another important question is why people differ in their sleep duration. Two large twin studies have indicated that individual differences in sleep duration could be attributed to both genetic and
153
154
A. Butkovic et al.
environmental influences. Partinen et al. (1983) have compared self-reported sleep length and sleep quality in monozygotic (MZ) and dizygotic (DZ) twin pairs from the Finnish Twin Cohort Study, and found a heritability estimate for sleep length of 0.44. Recently, Hublin et al. (2013) investigated the contribution of genetic factors to stability and change of sleep length among adults over a 15-year period using the Finnish Twin Cohort. They found heritability of approximately 0.30 for sleep length that was quite stable across three measurement points and high genetic correlations, from 0.76 to 0.93, between three measurement points. Heath et al. (1990) examined the aetiological role of genetic and environmental influences in sleep pattern (daytime napping, habitual bedtime and sleep duration), subjective sleep quality and sleep disturbance using a general population sample of adult Australian twin pairs aged 17–88 years. Broad heritability estimates, which include both additive genetic influences or direct effects of genes and non-additive genetic influences or gene interactions, of the sleep pattern variables were remarkably consistent, ranging between 38 and 46%. For sleep duration specifically, the broad heritability estimate was 0.40, with evidence for non-additive genetic influences being larger than additive genetic influences (0.31 versus 0.09). However, not all behavioural–genetic studies of sleep duration have found evidence of the importance of genetic influences. Barclay et al. (2010) have collected data from a UK sample of twins and siblings on the Pittsburgh Sleep Quality Index and found that genetic factors were important for all sleep variables except sleep duration. Because 90% of the sample was aged 18–22 years, it could be that studying at the university may have attenuated the impact of genetic influences on sleep length as a possible environmental pressure. Another study, examining associations and differences in sleep–wake patterns among children and their parents from Taiwan, found no correlation between nocturnal sleep duration of children and parents (Gau and Merikangas, 2004). The relationship between personality and sleep duration was investigated in a number of studies, but the results were inconsistent. Studies differed considerably regarding age and composition of the sample, as well as personality and sleep measures. In early studies, Webb and Friel (1971) found no personality differences between short and long sleepers, while Hartmann (1973) concluded, based on clinical observations, that short sleepers were ‘non-worriers’ and long sleepers were ‘worriers’. However, two studies that followed showed that anxiety (a facet of neuroticism) is higher in short than long sleepers (Hicks and Pellegrini, 1977; Kumar and Vaidya, 1984). In more recent studies, two found no significant relationship between personality and sleep duration using different personality measures (Gray and Watson, 2002; Soehner et al., 2007), one found a significant positive relationship between sleep duration on weekdays and agreeableness and conscientiousness (Randler, 2008) and two studies using different personality measures found that people with high neuroticism had shorter sleep duration (Gau, 2000; Vincent et al., 2009).
The aim of this study was to examine which genetic and environmental influences contribute to individual differences in sleep duration in a sample of Croatian adolescent/early adult twins, as well as to investigate the relationship between personality and sleep duration. The use of the twin method makes it possible to examine genetic and environmental influences on the phenotypic associations between personality and sleep duration. Based on previous studies, we hypothesize that individual differences in sleep duration will be attributed to both genetic and environmental influences, that sleep duration will be correlated negatively on a phenotypic level with neuroticism and that this phenotypic association will be explained mainly by overlapping genetic factors. As for the other four personality factors, only one study found a positive correlation with agreeableness and conscientiousness, so we hypothesize that the other four personality factors will not be correlated significantly with sleep duration. Finally, we hypothesize that people with shorter (≤6 h) and longer (≥9 h) sleep duration will have a significantly different neuroticism score, in the direction of people with shorter sleep duration having higher neuroticism scores. METHODS The sample used in this research was formed in 2007 based on the register of citizens from the Zagreb (capital of Croatia) area. From that register, twin pairs born between 1985 and 1992 were identified. Data were collected via mail. From 2005 contacted individuals, 732 (36.5%) returned completed questionnaires. From those individuals, 339 twin pairs were identified (105 MZ and 234 DZ). Their age varied between 15 and 22 years [mean = 18.62, standard deviation (SD) = 2.31]. Zygosity was determined by a questionnaire constructed for the purpose of this research. It included 11 items, six evaluating physical similarities (e.g. facial appearance, hair colour) and five evaluating twin confusion by parents, other family members, teachers, casual friends and strangers. These items have been shown to be valid indicators of zygosity in a number of studies and have been used in different zygosity questionnaires (e.g. Chen et al., 1999; Goldsmith, 1991). Twins’ self-reports about their zygosity were coded so that the lower score indicated monozygosity. In a number of studies (e.g. Reed et al., 2005; Spitz et al., 1996), the use of questionnaires for zygosity determination has been shown to be accurate to approximately 95%. Sleep duration was measured with one item: ‘How long do you sleep on average every night?’. Responses ranged between 4 and 12 h per night (mean = 7.74; SD = 1.17); sleep duration did not differ between MZ and DZ twins (7.81 versus 7.71 h; t = 0.951, P = 0.342). We divided our participants into three groups in terms of sleep duration: those sleeping ≤6 h (n = 72, or 10%), from 6.5 to 8.5 h (n = 466, or 70%) and ≥9 h (n = 133, or 20%). In each group there were approximately twice as many DZ than MZ twins, which is comparable to the composition of the total sample. ª 2013 European Sleep Research Society
Sleep duration and personality in Croatian twins Personality traits were measured with the NEO-Five Factor Inventory (NEO-FFI), a short version of the NEO-Personality Inventory (Costa and McCrae, 1992). NEO-FFI has 60 items and measures neuroticism (N), extraversion (E), openness (O), agreeableness (A) and conscientiousness (C). Cronbach’s a coefficients were 0.81, 0.72, 0.57, 0.66 and 0.81 for neuroticism, extraversion, openness, agreeableness and conscientiousness, respectively. RESULTS Descriptive statistics are presented in Table 1. Because multivariate analysis of variance (MANOVA) showed some sex and age effects in our sample, age and sex were regressed from the scores before genetic model-fitting using an adjustment procedure proposed by McGue and Bouchard (1984). In our sample there were significant sex differences in neuroticism (F1,651 = 51.09, P < 0.001), with females scoring higher than males (MF = 22.02 versus MM = 17.14). For age, significant differences were found in conscientiousness (F7,651 = 5.76, P < 0.001), which increased with age. Twin intraclass correlations were also calculated, and represent the proportion of total variance due to variance between pairs and are a more appropriate measure of twin pair similarity than Pearson’s r. The pattern of MZ–DZ correlations indicates the relative importance of different sources of variance. The structural equation model-fitting program Mx was used for genetic model-fitting analyses (Neale et al., 2003). The goodness-of-fit of each genetic model was measured relative to the phenotypic (saturated) model. Based on the obtained pattern of MZ–DZ correlations shown in Table 1, the first tested model included additive genetic effects (A), nonadditive genetic (D) or shared environmental effects (C) and non-shared environmental effects (E). A series of nested models were run to test if any of the parameters could be equated or dropped from the model without significant worsening of fit. The results of the univariate analyses are presented in Table 2. As our sample composition and size has a limited power to detect non-additive genetic (D) or shared environmental (C)
Table 1 Descriptive statistics for sleep duration and personality traits Males
Females
Intraclass r
Measures
Mean
SD
Mean
SD
MZ
DZ
Sleep duration Neuroticism Extraversion Openness Agreeableness Conscientiousness
7.76 17.14 30.29 23.04 29.82 30.59
1.15 7.47 6.55 6.06 6.30 7.48
7.73 22.02 30.48 24.10 30.67 32.06
1.18 8.77 6.72 6.37 5.98 7.37
0.62 0.64 0.52 0.67 0.50 0.50
0.20 0.21 0.12 0.25 0.27 0.27
MZ, monozygotic twins, DZ, dizygotic twins; SD, standard deviation.
ª 2013 European Sleep Research Society
155
influences, which is also apparent from large confidence intervals of our parameter estimates reported in Table 2, we report the fit and parameter estimates from full ADE or ACE models. The results indicate that the broad heritability estimate, which includes both additive genetic (A) and non-additive genetic (D) influences, is 0.63 for sleep duration in our sample, with non-shared environmental influences explaining the rest of the variance. As for personality, broad heritability estimates range between 0.47 and 0.62, which is in line with previous findings on heritability of personality (Plomin et al., 2008). Phenotypic correlations between sleep duration and personality traits, as well as results from bivariate analyses, are reported in Table 3. As can be seen, sleep duration was correlated negatively with neuroticism (r = 0.14, P = 0.001) and openness (r = 0.16, P < 0.001). In order to assess how much these phenotypic relations are due to overlapping genetic and environmental influences, bivariate analyses were computed using bivariate Cholesky decomposition, which was then transformed into a correlated factors model to enable easier interpretation (Loehlin, 1996). Our models included additive genetic (A) and non-shared environmental effects (E), based on the results of univariate analyses. A significant genetic correlation of 0.26 was obtained between sleep duration and neuroticism, and 0.22 between sleep duration and openness. By comparing bivariate heritabilities (products of genetic correlation and the square root of each univariate heritability) with phenotypic correlations, a percentage of each phenotypic correlation explained by shared genetic influences was obtained and indicated that a whole phenotypic association with neuroticism is genetically mediated as well as 80% of the phenotypic association with openness. We also ran trivariate Cholesky decompositions, which included both personality traits (neuroticism and openness) in varying order and sleep duration. The results of these analyses were in line with bivariate analyses and the same genetic correlations were obtained. Individual differences in sleep duration were explained mainly by specific genetic (52%) and environmental (42%) influences, with approximately 6% of the variance explained by genetic influences shared with neuroticism and openness. Next, we compared neuroticism and openness personality scores in three groups of participants divided by the length of their sleep in groups of short sleepers (≤6 h), moderate sleepers (6.5–8.5 h) and long sleepers (≥9 h) using analysis of variance (ANOVA). As we had twins in our sample, we adjusted degrees of freedom to calculate the F-value significance by dividing them by two (McGue et al., 1993). The results were significant for both neuroticism (F2,326 = 7.05, P = 0.001) and openness (F2,326 = 6.55, P = 0.002). A post test showed that participants in the short hoc Scheffe sleepers group had significantly higher neuroticism scores than groups of moderate and long sleepers (23.28 versus 19.84 and 18.62). Also, the short sleepers group had a significantly higher openness score than the long sleepers group (25.56 versus 22.28).
156
A. Butkovic et al.
Table 2 Univariate model-fitting results for sleep duration and personality traits: fit statistics for saturated and full model with parameter estimates for the full model Measures
Model
Sleep duration N
Full Full Full Full Full Full Full Full Full Full Full Full
E O A C
saturated saturated saturated saturated saturated saturated
2LL (df) 2029.771 2035.393 4564.541 4575.603 4313.221 4316.864 4186.212 4191.805 4190.515 4198.763 4425.226 4431.062
(661) (666) (648) (653) (648) (653) (648) (653) (648) (653) (648) (653)
v2(df)
P
A
D
5.622 (5)
0.35
0.16 (0.00–0.51)
0.47 (0.00–0.71)
0.37 (0.28–0.50)
6.026 (5)
0.30
0.23 (0.00–0.66)
0.39 (0.00–0.71)
0.38 (0.28–0.49)
3.643 (5)
0.60
0.00 (0.00–0.51)
0.50 (0.00–0.62)
0.50 (0.38–0.65)
5.593 (5)
0.35
0.44 (0.00–0.69)
0.18 (0.00–0.68)
0.38 (0.29–0.50)
8.248 (5)
0.14
0.56 (0.27–0.66)
0.00 (0.00–0.19)
0.44 (0.34–0.59)
5.836 (5)
0.32
0.47 (0.10–0.62)
0.03 (0.00–0.30)
0.50 (0.38–0.65)
C
E
2LL, minus twice the log-likelihood of the data; df, degrees of freedom; v (df), 2LL (and df) difference between current and previous model; A, additive genetic variance; D, non-additive genetic variance; C, shared environmental variance; E, non-shared environmental variance; 95% confidence intervals shown in parentheses. 2
Table 3 Phenotypic correlations between sleep duration and personality traits and bivariate model-fitting results: fit statistics and correlation estimates for correlated factors AE model
Sleep duration with
Phenotypic correlation
2LL
df
Correlation estimates
Bivariate heritability
Neuroticism
0.14*
6603.217
1321
rA = 0.26 ( 0.44 to 0.08) rE = 0.03 ( 0.16 to 0.21)
0.15
Extraversion Openness
0.04 0.16*
6221.737
1321
rA = rE =
0.13
Agreeableness Conscientiousness
0.01 0.01
0.22 ( 0.41 to 0.04) 0.01 ( 0.19 to 0.17)
2LL, minus twice the log-likelihood of the data; df, degrees of freedom; rA, genetic correlation; rE, non-shared environmental correlation; 95% confidence intervals shown in parentheses, significant genetic and environmental correlations are shown in bold type. All variance estimates from these analyses were within 0.06 of the univariate estimates given in Table 2. *P < 0.01.
Finally, in order to assess if sleep duration can be predicted from the age and sex demographic variables and personality traits, we ran hierarchical regression analyses with age and sex entered into the first step, and personality traits into the second step. The results are presented in Table 4. Demographic variables have not explained any of the sleep duration variance, but both neuroticism (b = 0.17, P < 0.001) and openness (b = 0.16, P < 0.001) were significant predictors, with the whole model explaining 5% of sleep duration variance. DISCUSSION The first contribution of this study is that, in a sample of Croatian twins, it estimates the genetic and environmental influences which contribute to individual differences in sleep duration. In line with studies investigating the heritability of sleep duration in Finnish and Australian twins (Heath et al., 1990; Hublin et al., 2013; Partinen et al., 1983), we found
Table 4 Results of the hierarchical regression analysis Predictors Model 1 Age Gender Model 2 Age Gender Neuroticism Extraversion Openness Agreeableness Conscientiousness
Beta
t
Summary of the model
0.09 0.02
2.28 0.38
Adjusted R² = 0.01 F2,324 = 2.67
0.08 0.06 0.17* 0.03 0.16* 0.06 0.03
2.03 1.33 3.81 0.62 3.95 1.50 0.71
Adjusted R² = 0.05 F7,321 = 5.17* DR² = 0.04 FDR²(5,321) = 6.13*
*P < 0.01.
that genetic influences contribute to individual differences in sleep duration. Our heritability estimate is somewhat higher than in previous studies (63% versus 44%, 40%, approximately ª 2013 European Sleep Research Society
Sleep duration and personality in Croatian twins 30%), but with large confidence intervals. Also in line with Australian results, our results indicate that both additive and non-additive genetic influences might be important for individual differences in sleep duration. If non-additive genetic influences are important for sleep duration, this could explain why Gau and Merikangas (2004) found no parent–offspring similarity in sleep duration in a Taiwanese sample, as parents and offspring share only additive genetic influences. The main focus of this study was to estimate putative overlapping genetic influences of personality and sleep duration. This is a new and original aspect of the study, because it goes beyond correlational analysis between sleep duration and personality on a phenotypic level and examines the genetic and environmental factors contributing to those associations. However, as a sample size of thousands of twins, which is difficult to recruit, is needed to obtain precise estimates, the observed results have large confidence intervals. This should be taken into account when interpreting the results. Therefore, the second contribution of this study is that it explores the relationship between personality and sleep duration in a genetically informative sample which makes it possible to examine genetic and environmental influences on these phenotypic associations. In line with our hypothesis and previous studies (Gau, 2000; Vincent et al., 2009), we found a significant negative phenotypic association between sleep duration and neuroticism. We also found a significant negative phenotypic association between sleep duration and openness which was not found in previous studies exploring the relationship between personality and sleep duration. Because six of 12 NEO-FFI openness items indicate intellectual curiosity and appreciation of art and beauty, it seems that those individuals who, for example, enjoy and spend a great deal of time thinking and reading about different ideas are sleeping less. This finding is in line with Geiger et al.’s (2010) finding that there is a negative association between sleep duration and measures of intelligence. Also in line with our hypothesis, we found a significant difference in neuroticism score between groups of people sleeping ≤6 h, which had the highest neuroticism score, and groups of people sleeping between 6.5 and 8.5 h and ≥9 h, that did not differ in neuroticism score. We did not have a hypothesis regarding openness, but as there was a significant phenotypic correlation we also compared three sleep groups for openness score, and found that people who sleep ≤6 h have a significantly higher openness score than people sleeping ≥9 h. Bivariate and trivariate Cholesky decompositions indicate that phenotypic associations between personality and sleep duration are due mainly to overlapping genetic factors. Results of the trivariate analysis show that approximately 6% of the sleep duration variance is accounted for by genetic influences shared with neuroticism and openness. This percentage is similar to the amount of sleep duration variance explained by personality in the regression analysis, which was 5%. Taken together, this indicates that personality traits of neuroticism and openness contribute to the prediction of ª 2013 European Sleep Research Society
157
sleep duration due to overlapping genetic influences that contribute to both these personality traits and sleep duration. Sleep duration is influenced by a number of variables, and personality traits account for only a small proportion of sleep duration variance. However, as sleep duration has implications for different medical conditions such as diabetes or cardiovascular disease, as well as for increased mortality, it seems important to find stable psychological characteristics that are related to sleep duration. In this way, our study contributes to the understanding of which personality traits are related to sleep duration and why. The results of some studies that used different sleep variables indicate that personality might be related even more to sleep quality than to sleep duration (Gray and Watson, 2002; Soehner et al., 2007). Of course, there are limitations to our study. The specific age composition of our sample has strengths, but we do not know if our results would be the same in older twin cohorts. Heritability estimates, as well as relations with personality traits, might differ for sleep duration among different age groups. Sleep duration was measured with only one selfreport item, which asked for average amount of sleep per night. People usually sleep for different amounts of time on weekdays and at weekends. Also, an amount of other sleep information could be collected to understand more clearly the relationship between sleep and personality, such as selfreported bedtime and rise time on weekdays and at weekends, sleep habits, sleep problems and sleep quality. However, even with this limited information about sleep we can see that individual differences in sleep duration exist, and that they can be explained significantly by genetic and non-shared environmental differences. Also, in line with some other studies, we found that short sleepers have a higher neuroticism score. In a slightly younger adolescent sample (aged 12–15 years) it was shown that a high neuroticism score is associated with going to bed late on school days, short sleep duration, different sleep habits, sleep problems and impaired daytime function in comparison with a low neuroticism score (Gau, 2000). Ramsawh et al. (2011) investigated the relationship between childhood adversity and adult sleep quality in college students and found a significant relationship between childhood adversity and adult sleep quality which was mediated by neuroticism. The authors conclude that neuroticism may represent a potential target for change in future insomnia interventions. Because, in our study, neuroticism was related to sleep duration both on phenotypic and genetic levels, and as both neuroticism and sleep duration have been found in previous studies to be related to mortality and health problems, a future direction could be to design interventions to change neuroticism in order to diminish different sleeping problems. This is in line with novel ideas in personality psychology that changing personality traits should be the focus of interventions, as personality trait change can improve people’s lives (Roberts, 2012). Therefore, we believe that irrespective of the sample size limitations, the observed results are of clinical relevance.
158
A. Butkovic et al.
AUTHOR CONTRIBUTIONS AB contributed to study design, data collection, data analysis, interpretation of results and preparation of the paper; TV contributed to data analysis, interpretation of results and preparation of the paper; DB contributed to study design, data collection, interpretation of results and preparation of the paper. CONFLICTS OF INTEREST No conflicts of interest declared. REFERENCES AlDabal, L. and BaHammam, A. S. Metabolic, endocrine, and immune consequences of sleep deprivation. Open Respir. Med. J., 2011, 5: 31–43. Barclay, N. L., Eley, T. C., Buysse, D. J., Rijsdijk, F. V. and Gregory, A. M. Genetic and environmental influences on different components of the Pittsburgh Sleep Quality Index and their overlap. Sleep, 2010, 33: 659–668. Cappuccio, F. P., D’Elia, L., Strazzullo, P. and Miller, M. A. Sleep duration and all-cause mortality: a systematic review and metaanalysis of prospective studies. Sleep, 2010, 33: 585–592. Chen, W. J., Chang, H.-W., Wu, M.-Z. et al. Diagnosis of zygosity by questionnaire and polymarker polymerase chain reaction in young twins. Behav. Genet., 1999, 29: 115–123. Costa, P. T. Jr and McCrae, R. R. Revised NEO Personality Inventory (NEO PI-R) and the NEO Five-Factor Inventory (NEOFFI) Professional Manual. Psychological Assessment Resources, Odessa, FL, 1992. Gallicchio, L. and Kalesan, B. Sleep duration and mortality: a systematic review and meta-analysis. J. Sleep Res., 2009, 18: 148–158. Gau, S. Neuroticism and sleep-related problems in adolescence. Sleep, 2000, 23: 1–8. Gau, S. S. and Merikangas, K. R. Similarities and differences in sleep–wake patterns among adults and their children. Sleep, 2004, 27: 299–304. Geiger, A., Achermann, P. and Jenni, O. G. Association between sleep duration and intelligence scores in healthy children. Dev. Psychol., 2010, 46: 949–954. Goldsmith, H. H. A zygosity questionnaire for young twins: a research note. Behav. Genet., 1991, 21: 257–269. Gray, E. K. and Watson, D. General and specific traits of personality and their relation to sleep and academic performance. J. Pers., 2002, 70: 177–206. Hartmann, E. Sleep requirement: long sleepers, short sleepers, variable sleepers, and insomniacs. Psychosomatics, 1973, 14: 95–103. Heath, A. C., Kendler, K. S., Eaves, L. J. and Martin, N. G. Evidence for genetic influences on sleep disturbance and sleep pattern in twins. Sleep, 1990, 13: 318–335.
Hicks, R. A. and Pellegrini, R. J. Anxiety levels of short and long sleepers. Psychol. Rep., 1977, 41: 569–570. Hublin, C., Partinen, M., Koskenvuo, M. and Kaprio, J. Genetic factors in evolution of sleep length—a longitudinal twin study in Finnish adults. J. Sleep Res., 2013, 22: 513–518. Kumar, A. and Vaidya, A. K. Anxiety as a personality dimension of short and long sleepers. J. Clin. Psychol., 1984, 40: 197–198. Loehlin, J. C. The Cholesky approach: a cautionary note. Behav. Genet., 1996, 26: 65–69. €cher, T. and Sonka, Maurovich-Horvat, E., Pollma K. The effects of sleep and sleep deprivation on metabolic, endocrine and immune parameters. Prague Med. Rep., 2008, 109: 275–285. McGue, M. and Bouchard, T. J. Adjustment of twin data for the effects of age and sex. Behav. Genet., 1984, 14: 325–343. McGue, M., Bacon, S. and Lykken, D. Personality stability and change in early adulthood: a behavioral genetic analysis. Dev. Psychol., 1993, 29: 96–109. Neale, M. C., Boker, S. M., Xie, G. and Maes, H. H. M.. Mx: Statistical Modelling, 6th edn. Department of Psychiatry, Richmond, VA, 2003. Partinen, M., Kaprio, J., Koskenvuo, M., Putkonen, P. and Langinvainio, H. Genetic and environmental determination of human sleep. Sleep, 1983, 6: 179–185. Plomin, R., DeFries, J., McClearn, G. E. and McGuffin, P. Behavioral Genetics, 5th edn. Freeman, New York, 2008. Ramsawh, H. J., Ancoli-Israel, S., Sullivan, S. G., Hitchcock, C. A. and Stein, M. B. Neuroticism mediates the relationship between childhood adversity and adult sleep quality. Behav. Sleep Med., 2011, 9: 130–143. Randler, C. Morningness–eveningness, sleep–wake variables and big five personality factors. Pers. Individ. Dif., 2008, 45: 191–196. Reed, T., Plassman, B. L., Tanner, C. M., Dick, D. M., Rinehart, S. A. and Nichols, W. C. Verification of self-report of zygosity determined via DNA testing in a subset of the NAS–NRC Twin Registry 40 later. Twin Res. Hum. Genet., 2005, 8: 362–367. Roberts, B. W. Personality trait change and improving people’s lives: changing personality traits should be the focus of interventions [Abstract]. In: L. Di Blas, A. Carnaghi, D. Ferrante and V. Piccoli (Eds) Book of Abstracts: 16th European Conference on Personality Psychology. Universita Degli Studi di Trieste, Trieste, 2012: 17. Soehner, A. M., Kennedy, K. S. and Monk, T. H. Personality correlates with sleep–wake variables. Chronobiol. Int., 2007, 24: 889–903. Spitz, E., Moutier, R., Reed, T. et al. Comparative diagnoses of twin zygosity by SSLP variant analysis, questionnaire, and dermatoglyphic analysis. Behav. Genet., 1996, 26: 55–63. Vincent, N., Cox, B. and Clara, I. Are personality dimensions associated with sleep length in a large nationally representative sample? Compr. Psychiatry, 2009, 50: 158–163. Webb, W. B. and Friel, J. Sleep stage and personality characteristics of ‘natural’ long and short sleepers. Science, 1971, 171: 587–588.
ª 2013 European Sleep Research Society
