- Email: [email protected]
Neuroticism and sleep-onset: What is the long-term connection?
Personality and Individual Differences 48 (2010) 463–468
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Neuroticism and sleep-onset: What is the long-term connection? Nanette S. Danielsson a,*, Markus Jansson-Fröjmark a, Steven J. Linton a, Göran Jutengren b, Håkan Stattin b a b
Center for Health and Medical Psychology, Örebro University, 70182 Örebro, Sweden Center for Developmental Research, Örebro University, 70182 Örebro, Sweden
a r t i c l e
i n f o
Article history: Received 22 June 2009 Received in revised form 5 November 2009 Accepted 18 November 2009 Available online 16 December 2009 Keywords: Neuroticism Sleep-onset problems Adolescence Epidemiological Longitudinal Prospective
a b s t r a c t People with sleep-onset problems often experience neuroticism. To what extent the one problem leads to the other is unknown. We used self-reported data from a Swedish longitudinal project to examine developmental links between neuroticism and sleep-onset problems. A sample of 212 people, followed from birth to midlife, was part of a cohort study spanning 37 years. Adolescent neuroticism was measured at age 16 with the High School Personality Questionnaire (HSPQ, Form A) and in midlife at age 37 with the Eysenck Personality Questionnaire (EPQ). Sleep-onset problems were measured at ages 15 to 17, 25, and 37 with items developed for the Solna Project. Adolescent neuroticism failed to predict sleep-onset problems. Instead, sleep-onset problems in adolescence and young adulthood predicted midlife neuroticism. We found that sleep-onset problems during adolescence were a direct risk for midlife neuroticism, as well as, an indirect risk through continuance of sleep-onset problems into adulthood. This study provides longitudinal support for adolescent sleep-onset problems as a potent risk factor for heightened neuroticism in midlife. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Some people with sleep-onset problems experience symptoms resembling neuroticism. Neuroticism, also known as negative affectivity or trait anxiety (Watson & Clark, 1984), is a genetically and environmentally influenced personality trait (McCrae, Jang, Livesley, Riemann, & Angleitner, 2001; Vernon, Jang, Harris, & McCarthy, 1997). Traits predispose people to perceive and experience their environment in characteristic ways (Buss & Plomin, 1984). As such, neuroticism is a tendency toward negative emotions, cognitions, and behaviours (Jang, McCrae, Angleitner, Riemann, & Livesley, 1998; Watson, Clark, & Harkness, 1994). Likewise, sleep-onset problems have emotional (e.g., Dahl & Lewin, 2002), cognitive (e.g., Roth, 2007), and behavioural components (e.g., Perlis, Giles, Mendelson, Bootzin, & Wyatt, 1997). Similarities suggest that one problem may be important for the other. Theoretically, processes involved in the development of insomnia, especially sleep-onset problems, are hierarchical. These are predisposing conditions, precipitating circumstances, and perpetuating factors (e.g., Drake & Roth, 2006). Processes may include biological (e.g., Espie, 2002) and genetic (Watson & Goldberg, 2006) predisposition; stressful life-event precipitation (Vahtera et al., 2007); and physiological (e.g., Perlis et al., 1997), cognitive (e.g.,
* Corresponding author. Address: Center for Health and Medical Psychology (CHAMP), School of Law, Psychology, and Social Work, Örebro University, 70182 Örebro, Sweden. Tel.: +46 19 301192. E-mail address: [email protected] (N.S. Danielsson). 0191-8869/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2009.11.023
Harvey, 2002), and behavioural (e.g., Perlis et al., 1997) perpetuation. Potentially the association occurs on all levels. Perhaps neuroticism predisposes some people to react toward lifestressors with negative emotions, cognitions, and behaviours incompatible with falling asleep. Predisposition alone is likely insufficient for developing sleep-onset problems (e.g., Drake & Roth, 2006). Instead, life-stressors may aggravate neuroticism, thus contributing to sleep-onset problems. On the other hand sleep-onset problems may heighten neuroticism. Typical ways people with sleep-onset problems feel, think, and react is like that for neuroticism (e.g., Dorsey & Bootzin, 1997; Freedman & Sattler, 1982; Shealy, Lowe, & Ritzler, 1980; Vahtera et al., 2007). Unknown is the nature of the association. Indeed, sleep-onset problems are associated with neuroticism. College students with sleep-onset problems reported higher neuroticism compared to controls (Shealy et al., 1980). In a laboratory, university students subjectively experiencing sleep-onset problems, reported more neuroticism than students with objective sleep-onset problems (Dorsey & Bootzin, 1997). The direction of the relation is unclear. The few studies about neuroticism and sleep-onset are crosssectional or retrospective. Which problem precedes or precipitates the other cannot be determined with these methods. Retrospective reports have the added disadvantage of memory distortion. Neuroticism may result in sleep-onset problems, or sleep-onset problems may spur emotional instability. Only longitudinally can the long-term relation be determined. Our aim is to examine developmental associations between these two problems. The longitudinal design enables us to examine
464
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
the long-term relation between neuroticism and sleep-onset problems. Specifically, is neuroticism predictive of sleep-onset problems or vice versa? To our knowledge, this is the first study to examine the long-term connection between neuroticism and sleep-onset problems. 2. Method 2.1. Participants Data were from a birth to midlife cohort project at the Clinic for the Study of Children’s Development and Health at Karolinska Hospital in Stockholm, Sweden. The study began 1955 in Solna, a suburb of Stockholm. Most pregnant women in Sweden receive prenatal care. Researchers asked every fourth pregnant woman registered for prenatal care to participate in a long-term paediatric study. The resulting sample was 212 children, 122 (58%) boys and 90 (42%) girls. The sample was shown through comparisons of socio-economic factors, parental age, maternal civil status, birth order, gestational age, and weight, in addition to registered criminality to be representative of Swedish children (Karlberg et al., 1968; Stattin & Klackenberg-Larsson, 1990). The present study included 185 participants, 107 (58%) men, 78 (42%) women. This was 91% of the original sample still alive age 37. Forty-five percent were married, 31% cohabitating, and 16% had a romantic relationship. On average, partners had been living together 10 years. Non-cohabitating romantic relationships averaged 3.5 years. Seventy-eight percent of participants had children. Thirty-eight percent of participants, and 35% of their partners held a university degree. 2.2. Procedure Researchers tracked participants’ somatic, psychological, and social development through medical examinations, interviews, and inventories (Karlberg et al., 1976). Participants were examined six times the first year, twice the second, and annually thereafter until age 18. Researchers tested participants as closely as possible to the given ages to control for differences in chronological age, usually within ±14 days. Additional data collections occurred when participants were 21, 25 and 37. 2.3. Material 2.3.1. Neuroticism measures 2.3.1.1. Adolescence. At age 16, the High School Personality Questionnaire (HSPQ, Form A, Second Edition, Cattell, 1962) was administered. In another study, a principal components analysis was conducted with a varimax rotation of the original HSPQ model (Ormerod & Billing, 1982). This was to address criticism of how many factors to include in Cattell’s 16 Personality Factor Model. We used the resulting 6 orthogonal factor model adapted from the HSPQ. The neuroticism measure comprises 14-items from three of four subscales from the anxiety/neuroticism factor. The three subscales used were emotionally stable (emotionally less stable, easily upset), tense (frustrated, overwrought), excitable (impatient, overactive). The three-item subscale, controlled, was excluded from the neuroticism scale due to its inclusion in the expedient/conscientious controlled factor (see Möller, 2004). Examples of remaining questions are ‘‘When you do a foolish thing, do you feel so badly that you wish the earth would just swallow you up?” (a = yes, b = perhaps, c = no), ‘‘Do small troubles sometimes ‘get on your nerves’ even though you know that they are not very important?” (a = yes, b = perhaps, c = no), ‘‘If your friends leave you out of something they are doing, do you (a = think they
made a mistake, b = in between, c = feel hurt and angry)?”, ‘‘When something important is coming up, such as a test or a big game, do you (a = stay very calm and relaxed, b = in between, c = get very tense and worried)?”, ‘‘Do you sometimes feel happy and sometimes depressed without real reason?” (a = yes, b = uncertain, c = no). Alpha reliability was .79. 2.3.1.2. Midlife. At age 37, participants completed Eysenck’s Personality Questionnaire (EPQ, Eysenck & Eysenck, 1975). We used the 23-item neuroticism scale. Higher scores equate with higher neuroticism. Examples were ‘‘After the fact, do you often worry about what you should not have done, or said?”, ‘‘Are you easily irritated?”, ‘‘Are your feelings easily hurt, or offended?”, ‘‘Do you often experience torment from guilt feelings?”, and ‘‘Do you consider yourself a nervous person?” (0 = no, 1 = yes). Alpha reliability was .87. 2.3.2. Sleep-onset problems Measures of sleep-onset problems were from self-report questionnaires developed for the Solna Project (Karlberg et al., 1968). Parent reports from ages 6 to 16 are reported here to establish stability. Parent-reported childhood sleep-onset problems were stable year-to-year from ages 6 to 16 (average r = .56, p < .001). Self-reported sleep-onset problems ages 15–17 showed year-to-year stability (average r = .44, p < .001) and corresponded with parentreported sleep-onset problems ages 15–16 (r = .47, p < .001). Selfreported sleep-onset problems ages 15–17 corresponded with those at ages 25 (r = .42, p < .001) and 37 (r = .31, p < .001). Self-reported sleep-onset problems at age 25 were related to continued problems at age 37 (r = .30, p < .001). Parents and their offspring report participant sleep-onset problems as stable from childhood through midlife. 2.3.2.1. Adolescence. The average score of the three annual responses to one-item, ages 15–17, formed a scale of adolescent sleep-onset problems: ‘‘Do you have difficulty falling asleep?” (1 = no, 2 = seldom; under certain circumstances, 3 = occasionally, 4 = yes, previously this year but not anymore, 5 = yes, often). The fourth scale anchor is uneven. Analyses conducted with and without the fourth scale anchor provided identical results. We retained the fourth scale anchor in the analyses to maintain population size. Reliability was .62, acceptable given the one-year intervals. 2.3.2.2. Young adulthood. One-item measured sleep-onset problems at age 25 (Karlberg et al., 1968). ‘‘Do you have difficulty falling asleep?” (1 = no, 2 = seldom; under certain circumstances, 3 = occasionally, 4 = yes, previously this year but not anymore, 5 = yes, often). See sleep-onset problems ages 15–17 for notes regarding scale anchors. 2.3.2.3. Midlife. At age 37, we averaged two-items as a measure of sleep-onset problems. The first was ‘‘Do you have difficulty falling asleep?” (1 = no, 2 = seldom; under certain circumstances, 3 = occasionally (up to once a month), 4 = sometimes (more than once a month to once a week), 5 = often (more than once a week), 6 = almost every night). The second was a broad measure of sleeping problems ‘‘Have you had difficulty sleeping within the last 3 months (difficulty falling asleep, waking during the night, poor sleep)?” (1 = no, 2 = once, 3 = once a month, 4 = once a week, 5 = several times per week, 6 = almost every day). The midlife sleep-onset measure corresponded with other sleep measures at age 37: sleep medication usage last year (r = .24, p < .001), and last 10 years (r = .27, p < .001); getting enough sleep (r = .19, p < .05); sleep duration (r = .18, p < .05); waking refreshed and rested (r = .33, p < .001); and sleep maintenance (r = .32, p < .001). This demonstrates measure construct validity. Comparisons of analyses with
465
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
or without the second sleep-onset measure were identical. Factor analysis of 11 sleep related items (sleep-onset problems, sleep problem last 3 months, sleep problem frequency, sleep maintenance, getting enough sleep, waking refreshed and rested, sleep duration, nightmares, dream recall, sleep medication usage during the last year, and previous 10 years) showed the two-items to be part of the same factor with loadings of .81. The two-item measure was used to retain population size. Alpha reliability was .82. 2.4. Analyses We used Mplus 5.0 for structural regression modelling (Muthén & Muthén, 2007). To examine and control for temporal and concurrent relations between neuroticism and sleep-onset problems, we conducted cross-lagged panel analyses (see Fig. 1). Paths included were (a) concurrent relations between sleep-onset problems and neuroticism, (b) stability paths, (c) and cross-lagged paths between earlier and later neuroticism and sleep-onset problems. Most prospective studies have some participants for which there is incomplete data at one or more time-points. Only conducting analyses on participants with complete data may bias results (Muthén, Kaplan, & Hollis, 1987). Therefore, analyses included the total sample (n = 185). We used full-information maximum likelihood estimation (ML), a robust estimator for missing data. Rather than imputing data, this estimation procedure breaks the likelihood function into components based on patterns of missing data estimated from available data (e.g., Arbuckle, 1996). Next, we consecutively removed non-significant paths stepwise from the model. Chi-square (v2) difference tests guided evaluation of retained paths. Model fit was evaluated using Root Mean Square Error of Approximation (RMSEA, Browne & Cudeck, 1993) and Comparative Fit Index (CFI, Bentler, 1990). These indexes are relatively insensitive to sample size (Fan, Thompson, & Wang, 1999). Values between .08 and .10 are considered moderate fit (e.g., Browne & Cudeck, 1993; MacCallum, Browne, & Sugawara, 1996; Mulaik et al., 1989) CFI values of .95 suggest acceptable fit, .97 good fit (Bentler, 1990). There is theoretical and correlational justification for the importance of adolescent sleep-onset problems in the development of neuroticism. Therefore, we conducted a fourth step to test a mediation model. Mediating variables account for the relation between other variables and explain how or why effects occur (Baron & Kenny, 1986). We controlled for sleep-onset problems in young adulthood by fixing the path between sleep-onset problems in young adulthood and neuroticism in midlife to zero (see path G). This was in order to see whether adolescent sleep-onset problems independently predict midlife neuroticism. Paths shown in the results are standardized.
2.5. Results Is neuroticism predictive of sleep-onset problems, or vice versa? Table 1 shows intercorrelations for variables in the model. Early neuroticism was not significantly related to later sleep-onset problems. Instead, sleep-onset problems in adolescence (r = .33, p < .001) and young adulthood (r = .24, p < .01) were associated with midlife neuroticism. Next, we tested the full model including concurrent and stability paths, and cross-lagged links. Table 2 shows fit indices and chisquare differences for consecutive stepwise procedures. First, we removed the path between adolescent neuroticism and sleep-onset problems in young adulthood (path D in Fig. 1, p = .26). This modification was acceptable in terms of model fit as indicated by the non-significant change in chi-square between steps 1 and 2 (see Table 2). In step 3, we removed the path between adolescent neuroticism and sleep-onset problems in midlife (path E in Fig. 1, p = .23). Again, the chi-square difference test indicated this was an acceptable modification (see Table 2). We made no further modifications, as all paths in this version of the model were statistically significant. The final model, had acceptable fit, v2 (3, N = 185) = 7.67, p = .05, RMSEA = .09, CFI = .95 (see Fig. 2). Sleep-onset problems appear stable from adolescence to young adulthood (b = .40, p < .001), and from young adulthood into midlife (b = .28, p < .001). Neuroticism showed stability from adolescence to midlife (b = .41, p < .001). Neuroticism and sleep-onset problems cooccurred in adolescence (b = .34, p < .001) and midlife (b = .25, p < .01). Theory suggests that neuroticism is important for the development of sleep-onset problems. Despite this, adolescent neuroticism failed to predict developing sleep-onset problems in either young adulthood or midlife. Instead sleep-onset problems in young adulthood were predictive of heightened neuroticism by midlife (b = .20, p < .01). The implication is that sleep-onset problems pose a risk for neuroticism in midlife, not vice versa. Table 1 Intercorrelations between sleep-onset problems and neuroticism in adolescence and adulthood (n = 185). Variable
1
2
3
1. 2. 3. 4. 5.
– .38*** .26** .33*** .33***
– .26** .03 .24**
– .07 .33
Sleep-onset problem ages 15–17 Sleep-onset problem age 25 Sleep-onset problem age 37 Neuroticism age 16 Neuroticism age 37
***
4
5
– .46***
–
Note: values in parentheses represent the significance level for non-significant correlations. p < .01. *** p < .001. **
Fig. 1. Cross-lagged path model for relations between sleep-onset problems and neuroticism from adolescence to midlife.
466
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
Table 2 Test of model fit and chi-square difference, removing least significant paths, and fixing the supposed mediating path in step 4 to zero.
v2
Model
df
v2/
RMSEA
CFI
Dv2
Ddf
.14 .10 .09 .11
.96 .96 .95 .89
1.26 1.44 6.64*
1 1 1
df Step Step Step Step
1. Full model 2. 3. 4. Mediation model
4.97* 6.23* 7.67 14.31**
1 2 3 4
4.86 3.12 2.56 3.58
Note: v2 = chi-square, df = degrees of freedom; CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation. * p < .05. ** p < .01.
In step 4, we tested a mediation model by controlling for sleeponset problems in young adulthood. Adolescent sleep-onset problems independently predicted midlife neuroticism (b = .20, p < .05). While a less satisfactory fit, this step suggests sleep-onset problems in young adulthood mediate the relation between adolescent sleep-onset problems and midlife neuroticism (see Table 2). 3. Discussion Our findings suggest that neuroticism is unrelated to developing sleep-onset problems. Instead, sleep-onset problems in adolescence and young adulthood pose a risk for developing neuroticism by midlife. Adolescent sleep-onset problems appear to influence midlife neuroticism directly and indirectly through continued sleep-onset problems into young adulthood. The relation seems to go in the opposite direction than might be expected from previous assumptions. Adolescent sleep-onset problems may aggravate neuroticism through changes brought on by sleep deprivation. Adolescence is a critical period for physiological and neurological changes associated with health (e.g., stress, diet, and sleep; Hyde, Mezulis, & Abramson, 2008) and gender aspects (e.g., Susman, Reiter, Ford, & Dorn, 2002). The kindling model suggests that stress sensitization occurs at the gene level (e.g., Monroe & Harkness, 2005; Post, 1992). Predisposition may lead to disorder when problems occur in timing and magnitude of gene expression (e.g., Paus, Keshavan, & Giedd, 2008). Over time, disorders become less dependent on life-stressors for recurrence. Perhaps sleep-onset problems influence neuroticism through sleep related worry. Sleep problems are associated with worry (Harvey, 2003; Harvey & Greenall, 2003). Mutual processes, such as worry, may perpetuate sleep-onset problems and increase neuroticism. Worry might elicit physiological arousal and emotional
distress contributing to both problems. Many people with sleeponset problems worry about sleep at bedtime and during the day. Thus, sleep-onset problems are a potential stressor and source of anxiety. Racing thoughts and ensuing emotions may elicit bodily tension or cognitive alertness or both. People may worry about bodily and cognitive sensations of arousal. The harder people try to sleep, the more elusive sleep may become (Ascher & Efran, 1978). Getting enough sleep may become an added concern. How people with sleep-onset problems tend to feel, think, and behave is a potential catalyst for neuroticism. Persistent sleep-onset problems may increase reactivity and reduce stress thresholds through physiological changes brought on by sleep deprivation (e.g., Meerlo, Sgoifo, & Suchecki, 2008) making it difficult to regulate emotions. Sleep deprivation is implicated in amplification of hyper-limbic response by the amygdala to negative emotional stimuli. Sleep-deprived people experienced a 60+% greater magnitude of amygdala reactivity relative to controls (Yoo, Gujar, Hu, Jolesz, & Walker, 2007). Sleep deprivation may even prime people to react to negative stimuli (Franzen, Siegle, & Buysse, 2008). In another study, sleep loss amplified negative experience of daytime disruptions and reduced positive benefits of rewarding activities (Zohar, Tzischinsky, Epstein, & Lavie, 2005). Sleep problems appear to leave people vulnerable to experiencing heightened neuroticism. Because of increased emotional arousal (amygdale, etc.) it may become difficult to continually regulate emotions. Thus worry predominates, that is, cannot be controlled or regulated. Sleep related emotional memory encoding and consolidation also appear to influence neuroticism. Sleep-deprived people retain more negative memories relative to controls (e.g., Walker & van der Helm, 2009). Modulation and processing of emotion during sleep may ‘‘reset” reactivity to emotional challenges the next day (Yoo et al., 2007). With repeated sleep cycles, especially rapid eye movement (REM), memories strengthen and associated affective tone decreases. Without decreases in affective tone, chronic anxiety may result (e.g., Walker & van der Helm, 2009). Sleep appears crucial for emotionality. We are aware of study limitations. One-item measures may result in underestimation of associations between variables. However, one-item sleep-onset measures are accepted in epidemiological sleep research (e.g., Laberge, Petit, Simard, Vitaro, & Tremblay, 2001; Morphy, Dunn, Lewis, Boardman, & Croft, 2007; Roberts, Roberts, & Chen, 2002; Roberts, Roberts, & Duong, 2008; Roberts, Shema, Kaplan, & Strawbridge, 2000). Stricter criteria for sleep-onset problems, as well as, mental and physical health measures would likely result in stronger connections between neuroticism and sleep-onset problems. Examples are sleep apnoea, restless legs syndrome, depression, and hypomania. Also, gender differences
Fig. 2. Final model testing cross-lagged path relations between sleep-onset problems and neuroticism from adolescence to midlife.
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
were unexamined, leaving the question of gender effects to later studies. Limits of this study represent areas for future research which would provide a broader picture of sleep and emotionality. Inclusion of process measures such as catastrophizing and worry would allow for examination of interaction effects. Also, person-oriented approaches examining differential pathways for developing sleep-onset problems may guide development of prevention and intervention programs. We believe that the strengths of this study outweigh the limitations. The longitudinal design provides a developmental look at neuroticism and sleep-onset problems. The sample is representative of the Swedish population in general. Impressively, 91% of the cohort still alive at the last data collection participated. This is remarkable given the 20-year time-span. Also, our neuroticism and sleep-onset measures show long-term stability. Final strengths are the high validity and reliability of the neuroticism measures, as well as having three time-points for sleep-onset. Our findings suggest that adolescent sleep-onset problems have a long-term negative impact on neuroticism. Adolescent sleep-onset problems likely influence neuroticism through emotional dysregulation as well as maladaptive cognitive and behavioural patterns that develop. These processes may be especially vulnerable to change in adolescence. Discovering when and if sleeping problems stabilize, and whether adolescence is a critical period for developing sleep-onset problems, has implications for prevention and intervention. Comorbidity with other disorders suggests sleep-onset problems pose risk for psychopathology (Roberts et al., 2002). Possibly early detection and treatment of those at risk for sleep-onset problems would have a positive impact on adolescent emotional development.
References Arbuckle, J. L. (1996). Full information estimation in the presence of incomplete data. In G. A. Marcoulides & R. E. Schumaker (Eds.), Advanced structural equation modeling: Issues and techniques (pp. 243–277). Mahwah, NJ: Lawrence Erlbaum. Ascher, L. M., & Efran, J. S. (1978). Use of paradoxical intention in a behavioral program for sleep onset insomnia. Journal of Consulting and Clinical Psychology, 46, 547–550. Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246. Browne, T. D., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury Park, CA: Sage. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Lawrence Erlbaum Associates. Cattell, R. B. (1962). Handbook for the Jr.–Sr. High School Personality Questionnaire: ‘‘The HSPQ”. Savoy, IL: Institute for Personality and Ability Testing. Dahl, R. E., & Lewin, D. S. (2002). Pathways to adolescent health: Sleep regulation and behavior. Journal of Adolescent Health, 31(Suppl.), 175–184. Dorsey, C. M., & Bootzin, R. R. (1997). Subjective and psychophysiologic insomnia: An examination of sleep tendency and personality. Biological Psychiatry, 41, 209–216. Drake, C. L., & Roth, T. (2006). Predisposition in the evolution of insomnia: Evidence, potential mechanisms, and future directions. Sleep Medicine Clinics, 1, 333–349. Espie, C. A. (2002). Insomnia: Conceptual issues in the development, persistence, and treatment of sleep disorder in adults. Annual Review of Psychology, 53, 215–243. Eysenck, H. J., & Eysenck, S. B. G. (1975). Manual for the Eysenck Personality Questionnaire. San Diego, CA: Educational and Industrial Testing Service. Fan, X., Thompson, B., & Wang, L. (1999). Effects of sample size, estimation methods, and model specification on structural equation modeling fit indexes. Structural Equation Modeling, 6, 56–83. Franzen, P. L., Siegle, G. J., & Buysse, D. J. (2008). Relationships between affect, vigilance, and sleepiness following sleep deprivation. Journal of Sleep Research, 17, 34–41. Freedman, R. R., & Sattler, H. L. (1982). Physiological and psychological factors in sleep-onset insomnia. Journal of Abnormal Psychology, 91, 380–389. Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40, 869–893.
467
Harvey, A. G. (2003). Beliefs about the utility of presleep worry: An investigation of individuals with insomnia and good sleepers. Cognitive Therapy and Research, 27, 403–414. Harvey, A. G., & Greenall, E. (2003). Catastrophic worry in primary insomnia. Journal of Behavior Therapy and Experimental Psychiatry, 34, 11–23. Hyde, J. S., Mezulis, A. H., & Abramson, L. Y. (2008). The ABCs of depression: Integrating affective, biological, and cognitive models to explain the emergence of the gender difference in depression. Psychological Review, 115, 291–313. Jang, K. L., McCrae, R. R., Angleitner, A., Riemann, R., & Livesley, W. J. (1998). Heritability of facet-level traits in a cross-cultural twin sample: Support for a hierarchical model of personality. Journal of Personality and Social Psychology, 74, 1556–1565. Karlberg, P., Klackenberg, G., Engström, L., Klackenberg-Larsson, L., Lichtenstein, H., Stensson, J., et al. (1968). The development of children in a Swedish urban community: A prospective, longitudinal study: Parts I–VI. Acta Paediatrica Scandinavica, 187(Suppl.), 9–93. Karlberg, P., Taranger, J., Engström, I., Lichtenstein, H., Svennberg-Redegren, I., Bruning, B., et al. (1976). The somatic development of children in a Swedish urban community: A prospective longitudinal study. Acta Paediatrica Scandinavica, 258(Suppl.), 1–70. Laberge, L., Petit, D., Simard, C., Vitaro, F., & Tremblay, R. E. (2001). Development of sleep patterns in early adolescence. Journal of Sleep Research, 10, 59–67. MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130–149. McCrae, R. R., Jang, K. L., Livesley, W. J., Riemann, R., & Angleitner, A. (2001). Sources of structure: Genetic, environmental, and artifactual influences on the covariation of personality traits. Journal of Personality, 69, 511–535. Meerlo, P., Sgoifo, A., & Suchecki, D. (2008). Restricted and disrupted sleep: Effects on autonomic function, neuroendocrine stress systems and stress responsivity. Sleep Medicine Reviews, 12, 197–210. Monroe, S. M., & Harkness, K. L. (2005). Life stress, ‘‘the kindling” hypothesis, and the recurrence of depression: Considerations from a life stress perspective. Psychological Review, 112, 417–445. Morphy, H., Dunn, K. M., Lewis, M., Boardman, H. F., & Croft, P. R. (2007). Epidemiology of insomnia: A longitudinal study in a UK population. Sleep, 30, 274–280. Mulaik, S. A., James, L. R., Alstine, J. V., Bennett, N., Lind, S., & Stilwell, C. D. (1989). Evaluation of goodness of fit indices for structural equation models. Psychological Bulletin, 105, 430–445. Muthén, B. O., Kaplan, D., & Hollis, M. (1987). On structural equation modeling with data that are not missing completely at random. Psychometrika, 52, 431– 462. Muthén, L. K., & Muthén, B. O. (2007). Mplus user’s guide (version 5). Los Angeles: Muthén and Muthén. Möller, K. (2004). The longitudinal and concurrent role of neuroticism for partner relationships. Scandinavian Journal of Psychology, 45, 79–83. Ormerod, M., & Billing, K. (1982). A six orthogonal factor model of adolescent personality derived from the HSPQ. Personality and Individual Differences, 3, 107–117. Paus, T., Keshavan, M., & Giedd, J. N. (2008). Why do many psychiatric disorders emerge during adolescence? Nature Reviews Neuroscience, 9, 947–957. Perlis, M. L., Giles, D. E., Mendelson, W. B., Bootzin, R. R., & Wyatt, J. K. (1997). Psychophysiological insomnia: The behavioural model and a neurocognitive perspective. Journal of Sleep Research, 6, 179–188. Post, R. M. (1992). Transduction of psychosocial stress into the neurobiology of recurrent affective disorder. American Journal of Psychiatry, 149, 999–1010. Roberts, R. E., Roberts, C. R., & Chen, I. G. (2002). Impact of insomnia on future functioning of adolescents. Journal of Psychosomatic Research, 53, 561–569. Roberts, R. E., Roberts, C. R., & Duong, H. T. (2008). Chronic Insomnia and its negative consequences for health and functioning of adolescents: A12-month prospective study. Journal of Adolescent Health, 42, 294–302. Roberts, R. E., Shema, S. J., Kaplan, G. A., & Strawbridge, W. J. (2000). Sleep complaints and depression in an aging cohort: A prospective perspective. American Journal of Psychiatry, 157, 81–88. Roth, T. (2007). Insomnia: Definition, prevalence, etiology, and consequences. Journal of Clinical Sleep Medicine, 3(Suppl. 5), S7–10. Shealy, R. C., Lowe, J. D., & Ritzler, B. A. (1980). Sleep onset insomnia: Personality characteristics and treatment outcome. Journal of Consulting and Clinical Psychology, 48, 659–661. Stattin, H., & Klackenberg-Larsson, I. (1990). The relationship between maternal attributes in the early life of the child and the child’s future criminal behavior. Development and Psychopathology, 2, 99–111. Susman, E. J., Reiter, E. O., Ford, C., & Dorn, L. D. (2002). Work group I: Developing models of healthy adolescent physical development. Journal of Adolescent Health, 31(Suppl. 6), 171–174. Vahtera, J., Kivimäki, M., Hublin, C., Korkeila, K., Suominen, S., Paunio, T., et al. (2007). Liability to anxiety and severe life events as predictors of new-onset sleep disturbances. Sleep, 30, 1537–1546. Vernon, P. A., Jang, K. L., Harris, J. A., & McCarthy, J. M. (1997). Environmental predictors of personality differences: A twin and sibling study. Journal of Personality and Social Psychology, 72, 177–183. Walker, M. P., & van der Helm, E. (2009). Overnight therapy? The role of sleep in emotional brain processing. Psychological Bulletin, 135, 731–748.
468
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465–490. Watson, D., Clark, L. A., & Harkness, A. R. (1994). Structures of personality and their relevance to psychopathology. Journal of Abnormal Psychology, 103, 18–31. Watson, N. F., & Goldberg, J. (2006). Genetic and environmental influences on insomnia, daytime sleepiness, and obesity in twins. Sleep, 29, 645–649.
Yoo, S.-S., Gujar, N., Hu, P., Jolesz, F. A., & Walker, M. P. (2007). The human emotional brain without sleep – A prefrontal amygdala disconnect? Current Biology, 17, R877–878. Zohar, D., Tzischinsky, O., Epstein, R., & Lavie, P. (2005). The effects of sleep loss on medical residents’ emotional reactions to work events: A cognitive-energy model. Sleep, 28, 47–54.
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Neuroticism and sleep-onset: What is the long-term connection? Nanette S. Danielsson a,*, Markus Jansson-Fröjmark a, Steven J. Linton a, Göran Jutengren b, Håkan Stattin b a b
Center for Health and Medical Psychology, Örebro University, 70182 Örebro, Sweden Center for Developmental Research, Örebro University, 70182 Örebro, Sweden
a r t i c l e
i n f o
Article history: Received 22 June 2009 Received in revised form 5 November 2009 Accepted 18 November 2009 Available online 16 December 2009 Keywords: Neuroticism Sleep-onset problems Adolescence Epidemiological Longitudinal Prospective
a b s t r a c t People with sleep-onset problems often experience neuroticism. To what extent the one problem leads to the other is unknown. We used self-reported data from a Swedish longitudinal project to examine developmental links between neuroticism and sleep-onset problems. A sample of 212 people, followed from birth to midlife, was part of a cohort study spanning 37 years. Adolescent neuroticism was measured at age 16 with the High School Personality Questionnaire (HSPQ, Form A) and in midlife at age 37 with the Eysenck Personality Questionnaire (EPQ). Sleep-onset problems were measured at ages 15 to 17, 25, and 37 with items developed for the Solna Project. Adolescent neuroticism failed to predict sleep-onset problems. Instead, sleep-onset problems in adolescence and young adulthood predicted midlife neuroticism. We found that sleep-onset problems during adolescence were a direct risk for midlife neuroticism, as well as, an indirect risk through continuance of sleep-onset problems into adulthood. This study provides longitudinal support for adolescent sleep-onset problems as a potent risk factor for heightened neuroticism in midlife. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Some people with sleep-onset problems experience symptoms resembling neuroticism. Neuroticism, also known as negative affectivity or trait anxiety (Watson & Clark, 1984), is a genetically and environmentally influenced personality trait (McCrae, Jang, Livesley, Riemann, & Angleitner, 2001; Vernon, Jang, Harris, & McCarthy, 1997). Traits predispose people to perceive and experience their environment in characteristic ways (Buss & Plomin, 1984). As such, neuroticism is a tendency toward negative emotions, cognitions, and behaviours (Jang, McCrae, Angleitner, Riemann, & Livesley, 1998; Watson, Clark, & Harkness, 1994). Likewise, sleep-onset problems have emotional (e.g., Dahl & Lewin, 2002), cognitive (e.g., Roth, 2007), and behavioural components (e.g., Perlis, Giles, Mendelson, Bootzin, & Wyatt, 1997). Similarities suggest that one problem may be important for the other. Theoretically, processes involved in the development of insomnia, especially sleep-onset problems, are hierarchical. These are predisposing conditions, precipitating circumstances, and perpetuating factors (e.g., Drake & Roth, 2006). Processes may include biological (e.g., Espie, 2002) and genetic (Watson & Goldberg, 2006) predisposition; stressful life-event precipitation (Vahtera et al., 2007); and physiological (e.g., Perlis et al., 1997), cognitive (e.g.,
* Corresponding author. Address: Center for Health and Medical Psychology (CHAMP), School of Law, Psychology, and Social Work, Örebro University, 70182 Örebro, Sweden. Tel.: +46 19 301192. E-mail address: [email protected] (N.S. Danielsson). 0191-8869/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2009.11.023
Harvey, 2002), and behavioural (e.g., Perlis et al., 1997) perpetuation. Potentially the association occurs on all levels. Perhaps neuroticism predisposes some people to react toward lifestressors with negative emotions, cognitions, and behaviours incompatible with falling asleep. Predisposition alone is likely insufficient for developing sleep-onset problems (e.g., Drake & Roth, 2006). Instead, life-stressors may aggravate neuroticism, thus contributing to sleep-onset problems. On the other hand sleep-onset problems may heighten neuroticism. Typical ways people with sleep-onset problems feel, think, and react is like that for neuroticism (e.g., Dorsey & Bootzin, 1997; Freedman & Sattler, 1982; Shealy, Lowe, & Ritzler, 1980; Vahtera et al., 2007). Unknown is the nature of the association. Indeed, sleep-onset problems are associated with neuroticism. College students with sleep-onset problems reported higher neuroticism compared to controls (Shealy et al., 1980). In a laboratory, university students subjectively experiencing sleep-onset problems, reported more neuroticism than students with objective sleep-onset problems (Dorsey & Bootzin, 1997). The direction of the relation is unclear. The few studies about neuroticism and sleep-onset are crosssectional or retrospective. Which problem precedes or precipitates the other cannot be determined with these methods. Retrospective reports have the added disadvantage of memory distortion. Neuroticism may result in sleep-onset problems, or sleep-onset problems may spur emotional instability. Only longitudinally can the long-term relation be determined. Our aim is to examine developmental associations between these two problems. The longitudinal design enables us to examine
464
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
the long-term relation between neuroticism and sleep-onset problems. Specifically, is neuroticism predictive of sleep-onset problems or vice versa? To our knowledge, this is the first study to examine the long-term connection between neuroticism and sleep-onset problems. 2. Method 2.1. Participants Data were from a birth to midlife cohort project at the Clinic for the Study of Children’s Development and Health at Karolinska Hospital in Stockholm, Sweden. The study began 1955 in Solna, a suburb of Stockholm. Most pregnant women in Sweden receive prenatal care. Researchers asked every fourth pregnant woman registered for prenatal care to participate in a long-term paediatric study. The resulting sample was 212 children, 122 (58%) boys and 90 (42%) girls. The sample was shown through comparisons of socio-economic factors, parental age, maternal civil status, birth order, gestational age, and weight, in addition to registered criminality to be representative of Swedish children (Karlberg et al., 1968; Stattin & Klackenberg-Larsson, 1990). The present study included 185 participants, 107 (58%) men, 78 (42%) women. This was 91% of the original sample still alive age 37. Forty-five percent were married, 31% cohabitating, and 16% had a romantic relationship. On average, partners had been living together 10 years. Non-cohabitating romantic relationships averaged 3.5 years. Seventy-eight percent of participants had children. Thirty-eight percent of participants, and 35% of their partners held a university degree. 2.2. Procedure Researchers tracked participants’ somatic, psychological, and social development through medical examinations, interviews, and inventories (Karlberg et al., 1976). Participants were examined six times the first year, twice the second, and annually thereafter until age 18. Researchers tested participants as closely as possible to the given ages to control for differences in chronological age, usually within ±14 days. Additional data collections occurred when participants were 21, 25 and 37. 2.3. Material 2.3.1. Neuroticism measures 2.3.1.1. Adolescence. At age 16, the High School Personality Questionnaire (HSPQ, Form A, Second Edition, Cattell, 1962) was administered. In another study, a principal components analysis was conducted with a varimax rotation of the original HSPQ model (Ormerod & Billing, 1982). This was to address criticism of how many factors to include in Cattell’s 16 Personality Factor Model. We used the resulting 6 orthogonal factor model adapted from the HSPQ. The neuroticism measure comprises 14-items from three of four subscales from the anxiety/neuroticism factor. The three subscales used were emotionally stable (emotionally less stable, easily upset), tense (frustrated, overwrought), excitable (impatient, overactive). The three-item subscale, controlled, was excluded from the neuroticism scale due to its inclusion in the expedient/conscientious controlled factor (see Möller, 2004). Examples of remaining questions are ‘‘When you do a foolish thing, do you feel so badly that you wish the earth would just swallow you up?” (a = yes, b = perhaps, c = no), ‘‘Do small troubles sometimes ‘get on your nerves’ even though you know that they are not very important?” (a = yes, b = perhaps, c = no), ‘‘If your friends leave you out of something they are doing, do you (a = think they
made a mistake, b = in between, c = feel hurt and angry)?”, ‘‘When something important is coming up, such as a test or a big game, do you (a = stay very calm and relaxed, b = in between, c = get very tense and worried)?”, ‘‘Do you sometimes feel happy and sometimes depressed without real reason?” (a = yes, b = uncertain, c = no). Alpha reliability was .79. 2.3.1.2. Midlife. At age 37, participants completed Eysenck’s Personality Questionnaire (EPQ, Eysenck & Eysenck, 1975). We used the 23-item neuroticism scale. Higher scores equate with higher neuroticism. Examples were ‘‘After the fact, do you often worry about what you should not have done, or said?”, ‘‘Are you easily irritated?”, ‘‘Are your feelings easily hurt, or offended?”, ‘‘Do you often experience torment from guilt feelings?”, and ‘‘Do you consider yourself a nervous person?” (0 = no, 1 = yes). Alpha reliability was .87. 2.3.2. Sleep-onset problems Measures of sleep-onset problems were from self-report questionnaires developed for the Solna Project (Karlberg et al., 1968). Parent reports from ages 6 to 16 are reported here to establish stability. Parent-reported childhood sleep-onset problems were stable year-to-year from ages 6 to 16 (average r = .56, p < .001). Self-reported sleep-onset problems ages 15–17 showed year-to-year stability (average r = .44, p < .001) and corresponded with parentreported sleep-onset problems ages 15–16 (r = .47, p < .001). Selfreported sleep-onset problems ages 15–17 corresponded with those at ages 25 (r = .42, p < .001) and 37 (r = .31, p < .001). Self-reported sleep-onset problems at age 25 were related to continued problems at age 37 (r = .30, p < .001). Parents and their offspring report participant sleep-onset problems as stable from childhood through midlife. 2.3.2.1. Adolescence. The average score of the three annual responses to one-item, ages 15–17, formed a scale of adolescent sleep-onset problems: ‘‘Do you have difficulty falling asleep?” (1 = no, 2 = seldom; under certain circumstances, 3 = occasionally, 4 = yes, previously this year but not anymore, 5 = yes, often). The fourth scale anchor is uneven. Analyses conducted with and without the fourth scale anchor provided identical results. We retained the fourth scale anchor in the analyses to maintain population size. Reliability was .62, acceptable given the one-year intervals. 2.3.2.2. Young adulthood. One-item measured sleep-onset problems at age 25 (Karlberg et al., 1968). ‘‘Do you have difficulty falling asleep?” (1 = no, 2 = seldom; under certain circumstances, 3 = occasionally, 4 = yes, previously this year but not anymore, 5 = yes, often). See sleep-onset problems ages 15–17 for notes regarding scale anchors. 2.3.2.3. Midlife. At age 37, we averaged two-items as a measure of sleep-onset problems. The first was ‘‘Do you have difficulty falling asleep?” (1 = no, 2 = seldom; under certain circumstances, 3 = occasionally (up to once a month), 4 = sometimes (more than once a month to once a week), 5 = often (more than once a week), 6 = almost every night). The second was a broad measure of sleeping problems ‘‘Have you had difficulty sleeping within the last 3 months (difficulty falling asleep, waking during the night, poor sleep)?” (1 = no, 2 = once, 3 = once a month, 4 = once a week, 5 = several times per week, 6 = almost every day). The midlife sleep-onset measure corresponded with other sleep measures at age 37: sleep medication usage last year (r = .24, p < .001), and last 10 years (r = .27, p < .001); getting enough sleep (r = .19, p < .05); sleep duration (r = .18, p < .05); waking refreshed and rested (r = .33, p < .001); and sleep maintenance (r = .32, p < .001). This demonstrates measure construct validity. Comparisons of analyses with
465
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
or without the second sleep-onset measure were identical. Factor analysis of 11 sleep related items (sleep-onset problems, sleep problem last 3 months, sleep problem frequency, sleep maintenance, getting enough sleep, waking refreshed and rested, sleep duration, nightmares, dream recall, sleep medication usage during the last year, and previous 10 years) showed the two-items to be part of the same factor with loadings of .81. The two-item measure was used to retain population size. Alpha reliability was .82. 2.4. Analyses We used Mplus 5.0 for structural regression modelling (Muthén & Muthén, 2007). To examine and control for temporal and concurrent relations between neuroticism and sleep-onset problems, we conducted cross-lagged panel analyses (see Fig. 1). Paths included were (a) concurrent relations between sleep-onset problems and neuroticism, (b) stability paths, (c) and cross-lagged paths between earlier and later neuroticism and sleep-onset problems. Most prospective studies have some participants for which there is incomplete data at one or more time-points. Only conducting analyses on participants with complete data may bias results (Muthén, Kaplan, & Hollis, 1987). Therefore, analyses included the total sample (n = 185). We used full-information maximum likelihood estimation (ML), a robust estimator for missing data. Rather than imputing data, this estimation procedure breaks the likelihood function into components based on patterns of missing data estimated from available data (e.g., Arbuckle, 1996). Next, we consecutively removed non-significant paths stepwise from the model. Chi-square (v2) difference tests guided evaluation of retained paths. Model fit was evaluated using Root Mean Square Error of Approximation (RMSEA, Browne & Cudeck, 1993) and Comparative Fit Index (CFI, Bentler, 1990). These indexes are relatively insensitive to sample size (Fan, Thompson, & Wang, 1999). Values between .08 and .10 are considered moderate fit (e.g., Browne & Cudeck, 1993; MacCallum, Browne, & Sugawara, 1996; Mulaik et al., 1989) CFI values of .95 suggest acceptable fit, .97 good fit (Bentler, 1990). There is theoretical and correlational justification for the importance of adolescent sleep-onset problems in the development of neuroticism. Therefore, we conducted a fourth step to test a mediation model. Mediating variables account for the relation between other variables and explain how or why effects occur (Baron & Kenny, 1986). We controlled for sleep-onset problems in young adulthood by fixing the path between sleep-onset problems in young adulthood and neuroticism in midlife to zero (see path G). This was in order to see whether adolescent sleep-onset problems independently predict midlife neuroticism. Paths shown in the results are standardized.
2.5. Results Is neuroticism predictive of sleep-onset problems, or vice versa? Table 1 shows intercorrelations for variables in the model. Early neuroticism was not significantly related to later sleep-onset problems. Instead, sleep-onset problems in adolescence (r = .33, p < .001) and young adulthood (r = .24, p < .01) were associated with midlife neuroticism. Next, we tested the full model including concurrent and stability paths, and cross-lagged links. Table 2 shows fit indices and chisquare differences for consecutive stepwise procedures. First, we removed the path between adolescent neuroticism and sleep-onset problems in young adulthood (path D in Fig. 1, p = .26). This modification was acceptable in terms of model fit as indicated by the non-significant change in chi-square between steps 1 and 2 (see Table 2). In step 3, we removed the path between adolescent neuroticism and sleep-onset problems in midlife (path E in Fig. 1, p = .23). Again, the chi-square difference test indicated this was an acceptable modification (see Table 2). We made no further modifications, as all paths in this version of the model were statistically significant. The final model, had acceptable fit, v2 (3, N = 185) = 7.67, p = .05, RMSEA = .09, CFI = .95 (see Fig. 2). Sleep-onset problems appear stable from adolescence to young adulthood (b = .40, p < .001), and from young adulthood into midlife (b = .28, p < .001). Neuroticism showed stability from adolescence to midlife (b = .41, p < .001). Neuroticism and sleep-onset problems cooccurred in adolescence (b = .34, p < .001) and midlife (b = .25, p < .01). Theory suggests that neuroticism is important for the development of sleep-onset problems. Despite this, adolescent neuroticism failed to predict developing sleep-onset problems in either young adulthood or midlife. Instead sleep-onset problems in young adulthood were predictive of heightened neuroticism by midlife (b = .20, p < .01). The implication is that sleep-onset problems pose a risk for neuroticism in midlife, not vice versa. Table 1 Intercorrelations between sleep-onset problems and neuroticism in adolescence and adulthood (n = 185). Variable
1
2
3
1. 2. 3. 4. 5.
– .38*** .26** .33*** .33***
– .26** .03 .24**
– .07 .33
Sleep-onset problem ages 15–17 Sleep-onset problem age 25 Sleep-onset problem age 37 Neuroticism age 16 Neuroticism age 37
***
4
5
– .46***
–
Note: values in parentheses represent the significance level for non-significant correlations. p < .01. *** p < .001. **
Fig. 1. Cross-lagged path model for relations between sleep-onset problems and neuroticism from adolescence to midlife.
466
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
Table 2 Test of model fit and chi-square difference, removing least significant paths, and fixing the supposed mediating path in step 4 to zero.
v2
Model
df
v2/
RMSEA
CFI
Dv2
Ddf
.14 .10 .09 .11
.96 .96 .95 .89
1.26 1.44 6.64*
1 1 1
df Step Step Step Step
1. Full model 2. 3. 4. Mediation model
4.97* 6.23* 7.67 14.31**
1 2 3 4
4.86 3.12 2.56 3.58
Note: v2 = chi-square, df = degrees of freedom; CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation. * p < .05. ** p < .01.
In step 4, we tested a mediation model by controlling for sleeponset problems in young adulthood. Adolescent sleep-onset problems independently predicted midlife neuroticism (b = .20, p < .05). While a less satisfactory fit, this step suggests sleep-onset problems in young adulthood mediate the relation between adolescent sleep-onset problems and midlife neuroticism (see Table 2). 3. Discussion Our findings suggest that neuroticism is unrelated to developing sleep-onset problems. Instead, sleep-onset problems in adolescence and young adulthood pose a risk for developing neuroticism by midlife. Adolescent sleep-onset problems appear to influence midlife neuroticism directly and indirectly through continued sleep-onset problems into young adulthood. The relation seems to go in the opposite direction than might be expected from previous assumptions. Adolescent sleep-onset problems may aggravate neuroticism through changes brought on by sleep deprivation. Adolescence is a critical period for physiological and neurological changes associated with health (e.g., stress, diet, and sleep; Hyde, Mezulis, & Abramson, 2008) and gender aspects (e.g., Susman, Reiter, Ford, & Dorn, 2002). The kindling model suggests that stress sensitization occurs at the gene level (e.g., Monroe & Harkness, 2005; Post, 1992). Predisposition may lead to disorder when problems occur in timing and magnitude of gene expression (e.g., Paus, Keshavan, & Giedd, 2008). Over time, disorders become less dependent on life-stressors for recurrence. Perhaps sleep-onset problems influence neuroticism through sleep related worry. Sleep problems are associated with worry (Harvey, 2003; Harvey & Greenall, 2003). Mutual processes, such as worry, may perpetuate sleep-onset problems and increase neuroticism. Worry might elicit physiological arousal and emotional
distress contributing to both problems. Many people with sleeponset problems worry about sleep at bedtime and during the day. Thus, sleep-onset problems are a potential stressor and source of anxiety. Racing thoughts and ensuing emotions may elicit bodily tension or cognitive alertness or both. People may worry about bodily and cognitive sensations of arousal. The harder people try to sleep, the more elusive sleep may become (Ascher & Efran, 1978). Getting enough sleep may become an added concern. How people with sleep-onset problems tend to feel, think, and behave is a potential catalyst for neuroticism. Persistent sleep-onset problems may increase reactivity and reduce stress thresholds through physiological changes brought on by sleep deprivation (e.g., Meerlo, Sgoifo, & Suchecki, 2008) making it difficult to regulate emotions. Sleep deprivation is implicated in amplification of hyper-limbic response by the amygdala to negative emotional stimuli. Sleep-deprived people experienced a 60+% greater magnitude of amygdala reactivity relative to controls (Yoo, Gujar, Hu, Jolesz, & Walker, 2007). Sleep deprivation may even prime people to react to negative stimuli (Franzen, Siegle, & Buysse, 2008). In another study, sleep loss amplified negative experience of daytime disruptions and reduced positive benefits of rewarding activities (Zohar, Tzischinsky, Epstein, & Lavie, 2005). Sleep problems appear to leave people vulnerable to experiencing heightened neuroticism. Because of increased emotional arousal (amygdale, etc.) it may become difficult to continually regulate emotions. Thus worry predominates, that is, cannot be controlled or regulated. Sleep related emotional memory encoding and consolidation also appear to influence neuroticism. Sleep-deprived people retain more negative memories relative to controls (e.g., Walker & van der Helm, 2009). Modulation and processing of emotion during sleep may ‘‘reset” reactivity to emotional challenges the next day (Yoo et al., 2007). With repeated sleep cycles, especially rapid eye movement (REM), memories strengthen and associated affective tone decreases. Without decreases in affective tone, chronic anxiety may result (e.g., Walker & van der Helm, 2009). Sleep appears crucial for emotionality. We are aware of study limitations. One-item measures may result in underestimation of associations between variables. However, one-item sleep-onset measures are accepted in epidemiological sleep research (e.g., Laberge, Petit, Simard, Vitaro, & Tremblay, 2001; Morphy, Dunn, Lewis, Boardman, & Croft, 2007; Roberts, Roberts, & Chen, 2002; Roberts, Roberts, & Duong, 2008; Roberts, Shema, Kaplan, & Strawbridge, 2000). Stricter criteria for sleep-onset problems, as well as, mental and physical health measures would likely result in stronger connections between neuroticism and sleep-onset problems. Examples are sleep apnoea, restless legs syndrome, depression, and hypomania. Also, gender differences
Fig. 2. Final model testing cross-lagged path relations between sleep-onset problems and neuroticism from adolescence to midlife.
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
were unexamined, leaving the question of gender effects to later studies. Limits of this study represent areas for future research which would provide a broader picture of sleep and emotionality. Inclusion of process measures such as catastrophizing and worry would allow for examination of interaction effects. Also, person-oriented approaches examining differential pathways for developing sleep-onset problems may guide development of prevention and intervention programs. We believe that the strengths of this study outweigh the limitations. The longitudinal design provides a developmental look at neuroticism and sleep-onset problems. The sample is representative of the Swedish population in general. Impressively, 91% of the cohort still alive at the last data collection participated. This is remarkable given the 20-year time-span. Also, our neuroticism and sleep-onset measures show long-term stability. Final strengths are the high validity and reliability of the neuroticism measures, as well as having three time-points for sleep-onset. Our findings suggest that adolescent sleep-onset problems have a long-term negative impact on neuroticism. Adolescent sleep-onset problems likely influence neuroticism through emotional dysregulation as well as maladaptive cognitive and behavioural patterns that develop. These processes may be especially vulnerable to change in adolescence. Discovering when and if sleeping problems stabilize, and whether adolescence is a critical period for developing sleep-onset problems, has implications for prevention and intervention. Comorbidity with other disorders suggests sleep-onset problems pose risk for psychopathology (Roberts et al., 2002). Possibly early detection and treatment of those at risk for sleep-onset problems would have a positive impact on adolescent emotional development.
References Arbuckle, J. L. (1996). Full information estimation in the presence of incomplete data. In G. A. Marcoulides & R. E. Schumaker (Eds.), Advanced structural equation modeling: Issues and techniques (pp. 243–277). Mahwah, NJ: Lawrence Erlbaum. Ascher, L. M., & Efran, J. S. (1978). Use of paradoxical intention in a behavioral program for sleep onset insomnia. Journal of Consulting and Clinical Psychology, 46, 547–550. Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246. Browne, T. D., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury Park, CA: Sage. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Lawrence Erlbaum Associates. Cattell, R. B. (1962). Handbook for the Jr.–Sr. High School Personality Questionnaire: ‘‘The HSPQ”. Savoy, IL: Institute for Personality and Ability Testing. Dahl, R. E., & Lewin, D. S. (2002). Pathways to adolescent health: Sleep regulation and behavior. Journal of Adolescent Health, 31(Suppl.), 175–184. Dorsey, C. M., & Bootzin, R. R. (1997). Subjective and psychophysiologic insomnia: An examination of sleep tendency and personality. Biological Psychiatry, 41, 209–216. Drake, C. L., & Roth, T. (2006). Predisposition in the evolution of insomnia: Evidence, potential mechanisms, and future directions. Sleep Medicine Clinics, 1, 333–349. Espie, C. A. (2002). Insomnia: Conceptual issues in the development, persistence, and treatment of sleep disorder in adults. Annual Review of Psychology, 53, 215–243. Eysenck, H. J., & Eysenck, S. B. G. (1975). Manual for the Eysenck Personality Questionnaire. San Diego, CA: Educational and Industrial Testing Service. Fan, X., Thompson, B., & Wang, L. (1999). Effects of sample size, estimation methods, and model specification on structural equation modeling fit indexes. Structural Equation Modeling, 6, 56–83. Franzen, P. L., Siegle, G. J., & Buysse, D. J. (2008). Relationships between affect, vigilance, and sleepiness following sleep deprivation. Journal of Sleep Research, 17, 34–41. Freedman, R. R., & Sattler, H. L. (1982). Physiological and psychological factors in sleep-onset insomnia. Journal of Abnormal Psychology, 91, 380–389. Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40, 869–893.
467
Harvey, A. G. (2003). Beliefs about the utility of presleep worry: An investigation of individuals with insomnia and good sleepers. Cognitive Therapy and Research, 27, 403–414. Harvey, A. G., & Greenall, E. (2003). Catastrophic worry in primary insomnia. Journal of Behavior Therapy and Experimental Psychiatry, 34, 11–23. Hyde, J. S., Mezulis, A. H., & Abramson, L. Y. (2008). The ABCs of depression: Integrating affective, biological, and cognitive models to explain the emergence of the gender difference in depression. Psychological Review, 115, 291–313. Jang, K. L., McCrae, R. R., Angleitner, A., Riemann, R., & Livesley, W. J. (1998). Heritability of facet-level traits in a cross-cultural twin sample: Support for a hierarchical model of personality. Journal of Personality and Social Psychology, 74, 1556–1565. Karlberg, P., Klackenberg, G., Engström, L., Klackenberg-Larsson, L., Lichtenstein, H., Stensson, J., et al. (1968). The development of children in a Swedish urban community: A prospective, longitudinal study: Parts I–VI. Acta Paediatrica Scandinavica, 187(Suppl.), 9–93. Karlberg, P., Taranger, J., Engström, I., Lichtenstein, H., Svennberg-Redegren, I., Bruning, B., et al. (1976). The somatic development of children in a Swedish urban community: A prospective longitudinal study. Acta Paediatrica Scandinavica, 258(Suppl.), 1–70. Laberge, L., Petit, D., Simard, C., Vitaro, F., & Tremblay, R. E. (2001). Development of sleep patterns in early adolescence. Journal of Sleep Research, 10, 59–67. MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130–149. McCrae, R. R., Jang, K. L., Livesley, W. J., Riemann, R., & Angleitner, A. (2001). Sources of structure: Genetic, environmental, and artifactual influences on the covariation of personality traits. Journal of Personality, 69, 511–535. Meerlo, P., Sgoifo, A., & Suchecki, D. (2008). Restricted and disrupted sleep: Effects on autonomic function, neuroendocrine stress systems and stress responsivity. Sleep Medicine Reviews, 12, 197–210. Monroe, S. M., & Harkness, K. L. (2005). Life stress, ‘‘the kindling” hypothesis, and the recurrence of depression: Considerations from a life stress perspective. Psychological Review, 112, 417–445. Morphy, H., Dunn, K. M., Lewis, M., Boardman, H. F., & Croft, P. R. (2007). Epidemiology of insomnia: A longitudinal study in a UK population. Sleep, 30, 274–280. Mulaik, S. A., James, L. R., Alstine, J. V., Bennett, N., Lind, S., & Stilwell, C. D. (1989). Evaluation of goodness of fit indices for structural equation models. Psychological Bulletin, 105, 430–445. Muthén, B. O., Kaplan, D., & Hollis, M. (1987). On structural equation modeling with data that are not missing completely at random. Psychometrika, 52, 431– 462. Muthén, L. K., & Muthén, B. O. (2007). Mplus user’s guide (version 5). Los Angeles: Muthén and Muthén. Möller, K. (2004). The longitudinal and concurrent role of neuroticism for partner relationships. Scandinavian Journal of Psychology, 45, 79–83. Ormerod, M., & Billing, K. (1982). A six orthogonal factor model of adolescent personality derived from the HSPQ. Personality and Individual Differences, 3, 107–117. Paus, T., Keshavan, M., & Giedd, J. N. (2008). Why do many psychiatric disorders emerge during adolescence? Nature Reviews Neuroscience, 9, 947–957. Perlis, M. L., Giles, D. E., Mendelson, W. B., Bootzin, R. R., & Wyatt, J. K. (1997). Psychophysiological insomnia: The behavioural model and a neurocognitive perspective. Journal of Sleep Research, 6, 179–188. Post, R. M. (1992). Transduction of psychosocial stress into the neurobiology of recurrent affective disorder. American Journal of Psychiatry, 149, 999–1010. Roberts, R. E., Roberts, C. R., & Chen, I. G. (2002). Impact of insomnia on future functioning of adolescents. Journal of Psychosomatic Research, 53, 561–569. Roberts, R. E., Roberts, C. R., & Duong, H. T. (2008). Chronic Insomnia and its negative consequences for health and functioning of adolescents: A12-month prospective study. Journal of Adolescent Health, 42, 294–302. Roberts, R. E., Shema, S. J., Kaplan, G. A., & Strawbridge, W. J. (2000). Sleep complaints and depression in an aging cohort: A prospective perspective. American Journal of Psychiatry, 157, 81–88. Roth, T. (2007). Insomnia: Definition, prevalence, etiology, and consequences. Journal of Clinical Sleep Medicine, 3(Suppl. 5), S7–10. Shealy, R. C., Lowe, J. D., & Ritzler, B. A. (1980). Sleep onset insomnia: Personality characteristics and treatment outcome. Journal of Consulting and Clinical Psychology, 48, 659–661. Stattin, H., & Klackenberg-Larsson, I. (1990). The relationship between maternal attributes in the early life of the child and the child’s future criminal behavior. Development and Psychopathology, 2, 99–111. Susman, E. J., Reiter, E. O., Ford, C., & Dorn, L. D. (2002). Work group I: Developing models of healthy adolescent physical development. Journal of Adolescent Health, 31(Suppl. 6), 171–174. Vahtera, J., Kivimäki, M., Hublin, C., Korkeila, K., Suominen, S., Paunio, T., et al. (2007). Liability to anxiety and severe life events as predictors of new-onset sleep disturbances. Sleep, 30, 1537–1546. Vernon, P. A., Jang, K. L., Harris, J. A., & McCarthy, J. M. (1997). Environmental predictors of personality differences: A twin and sibling study. Journal of Personality and Social Psychology, 72, 177–183. Walker, M. P., & van der Helm, E. (2009). Overnight therapy? The role of sleep in emotional brain processing. Psychological Bulletin, 135, 731–748.
468
N.S. Danielsson et al. / Personality and Individual Differences 48 (2010) 463–468
Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465–490. Watson, D., Clark, L. A., & Harkness, A. R. (1994). Structures of personality and their relevance to psychopathology. Journal of Abnormal Psychology, 103, 18–31. Watson, N. F., & Goldberg, J. (2006). Genetic and environmental influences on insomnia, daytime sleepiness, and obesity in twins. Sleep, 29, 645–649.
Yoo, S.-S., Gujar, N., Hu, P., Jolesz, F. A., & Walker, M. P. (2007). The human emotional brain without sleep – A prefrontal amygdala disconnect? Current Biology, 17, R877–878. Zohar, D., Tzischinsky, O., Epstein, R., & Lavie, P. (2005). The effects of sleep loss on medical residents’ emotional reactions to work events: A cognitive-energy model. Sleep, 28, 47–54.