Citation
The Mediation of the Relationship between Social Rhythmicity and Sleep by Light, Arousal, and Affect in Both Younger and Older Adults

Material Information

Title:
The Mediation of the Relationship between Social Rhythmicity and Sleep by Light, Arousal, and Affect in Both Younger and Older Adults
Creator:
Dautovich, Natalie
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (190 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Counseling Psychology
Psychology
Committee Chair:
McCrae, Christina S.
Committee Members:
Rice, Kenneth G.
West, Robin L.
Rowe, Meredeth A.
Graduation Date:
8/7/2010

Subjects

Subjects / Keywords:
Age groups ( jstor )
Biological rhythms ( jstor )
Circadian rhythm ( jstor )
Cognitive psychology ( jstor )
Insomnia ( jstor )
Multilevel models ( jstor )
Older adults ( jstor )
Sleep ( jstor )
Social psychology ( jstor )
Zeitgebers ( jstor )
Psychology -- Dissertations, Academic -- UF
adults, affect, arousal, elderly, geriatric, intraindividual, light, older, rhythmicity, sleep, social, younger
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Counseling Psychology thesis, Ph.D.

Notes

Abstract:
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE MEDIATION OF THE RELATIONSHIP BETWEEN SOCIAL RHYTHMICTY AND SLEEP BY LIGHT, AROUSAL, AND AFFECT IN BOTH YOUNGER AND OLDER ADULTS By Natalie Deidre Dautovich August 2010 Chair: Christina McCrae Major: Counseling Psychology Light currently reigns as the dominant factor influencing human circadian rhythms such as sleep. Recently, interest has shifted to other external factors which may influence circadian systems. Social activities (e.g., meal times, timing of exercise) are thought to exert effects on the timing of circadian rhythms. Social activities may entrain circadian rhythms either directly or indirectly by influencing exposure to light. Little research, however, has investigated mechanisms other than light that underlie the relationship between the rhythms of social activities and sleep. The present study examined the role of arousal and affect, in addition to light, as potential mediators of the relationship between social rhythms and sleep. The overarching goals were to 1) examine the amount of variability within social rhythmicity, light, arousal, affect and sleep; 2) examine how variability in social rhythmicity, light, arousal, and affect is related to sleep; 3) examine the day-to-day and overall relationships between social rhythmicity, light, arousal, affect, and sleep; and 4) examine the mediation of the relationship of social rhythmicity and sleep by light, arousal, and affect. Each of these goals was investigated in both younger and older adults. Younger (n = 50) and older (n = 50) community-dwelling individuals were recruited from the North Central Florida area and online. Participants completed daily online diaries for 14 consecutive days. Social activities were assessed using the Social Rhythm Metric (SRM-17), light exposure was evaluated with the Light Exposure Scale (LES), arousal was assessed using the Affect Grid, the Pre-Sleep Arousal Scale (PSAS), and the UWIST Mood Adjective Checklist Revised (UMACL-R), affect was evaluated with the Affect Grid and the Positive and Negative Affect Scale (PANAS), and sleep was assessed using a sleep diary. Results indicated that younger adults were highly variable within person for all variables. Younger adults were significantly more variable on all measures than older adults except for the PSAS somatic scale. Despite showing less variability compared to younger adults, older adults still demonstrated considerable variability. Age-related differences in mean-levels were observed for social rhythmicity, light exposure, arousal, affect, and sleep. Variability in social rhythmicity, light, arousal, and affect was associated with worse sleep for younger adults and better and worse sleep for older adults. Multi-level modeling analyses showed that social rhythmicity predicted sleep at the between person level and light, arousal, and affect predicted sleep at both within and between person levels for younger adults. For older adults, light, arousal, and affect predicted sleep at both between and within person levels. The potential mediating effects of light, arousal, and affect could not be tested as there were insufficient associations between the predictor (social rhythmicity), potential mediator (light, arousal, and affect), and outcome variables (sleep) for both age groups. Results provide support for the relationship between social rhythmicity, light, arousal, affect, and sleep for younger adults and for light, arousal, affect and sleep for older adults. Further research is needed to test the mediating effects of light, arousal, and affect. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2010.
Local:
Adviser: McCrae, Christina S.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31
Statement of Responsibility:
by Natalie Dautovich.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Dautovich, Natalie. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2011
Classification:
LD1780 2010 ( lcc )

Downloads

This item has the following downloads:


Full Text





5-16 Correlations among affect variables between persons (below diagonal) and within
person (above diagonal) for older adults ........................................ ....................... 98

5-17 Multilevel model with social rhythmicity variables predicting sleep in younger adults ...99

5-18 Multilevel model with social rhythmicity variables predicting sleep in older adults......100

5-19 Multilevel model with light variables predicting sleep in younger adults......................101

5-20 Multilevel model with light variables predicting sleep in older adults..........................103

5-21 Multilevel model with arousal variables predicting sleep in younger adults...................104

5-22 Multilevel model with arousal variables predicting sleep in older adults .......................106

5-23 Multilevel model with affect variables predicting sleep in younger adults................... 108

5-24 Multilevel model with affect variables predicting sleep in older adults.........................109









circadian system, and consequently impact sleep, it is warranted to examine the social stimuli

associated with circadian rhythm changes. One aspect of social stimuli, regularity of daily

activities, has been linked to sleep and age with greater rhythmicity in social activities associated

with better sleep and older age.

While the regularity of social stimuli has been linked to the sleep/wake cycle, the

mechanisms underlying this relationship have not been studied. Based on the negative

association between arousal and sleep, it was hypothesized that arousal may be a factor

mediating the relationship between social rhythmicity and sleep. To our knowledge, only one

study (Stetler et al., 2004) has examined the relationship between social rhythmicity and arousal.

Mandler's theory of mind and emotion (1975) states that interruption of one's plans and

organized behavior can generate arousal which can be interpreted as various emotions (Mandler,

1975). Consequently, it is hypothesized that greater social rhythmicity will be associated with

decreased arousal and better sleep. In terms of affect, we know that irregular social rhythms

have been associated with affective disorders such as depression (Brown et al., 1996; Prigerson

et al., 1994), mania (Ashman et al., 1999; Malkoff-Schwartz et al., 1998; Malkoff-Schwartz et

al., 2000) and anxiety (Shear et al., 1994). Additionally, literature on motivation indicates that

more regular, habitual behavior is associated with less emotional intensity. Wood, Quinn, &

Kashy (2002) found that individuals who engaged in habitual behaviors experienced less intense

emotions compared to when they engaged in nonhabitual behaviors (Wood, Quinn, & Kashy,

2002). Frijda's laws of emotion state that "continued pleasures wear off; continued hardships

lose their poignancy" (Frijda, 1988, p. 353). Consequently, it was hypothesized that higher

levels of social rhythm regularity will be associated with decreased affect and subsequently

better sleep. Finally, as light is the dominant zeitgeber affecting our circadian rhythms, it is































2010 NATALIE DEIDRE DAUTOVICH























Examples of types of light:


DARK/MOONLIGHT: no artificial light, very difficult to read in this light


DIM LIGHT very little light, light from television, a night-light, or candles, difficult to read in this light


INDOOR LIGHT: an overhead light, desk light, very easy to read in this light


CLOUDY you are outdoors in the daytime but there is not bright sunlight,there are clouds overhead, dusk, or dawn


BRIGHT SUNLIGHT you are outdoors in the daytime during bright sunlight, very few clouds overhead









specific events may have stronger entrainment potential for the different age groups. Arriving at

home may signal a 'winding down' of the day for older adults, whereas for younger adults, it

could signal the start of the next phase of activity.

Variability in the timing of exposure to light in the mornings was associated with worse

sleep for younger adults. Given that photic influences are the strongest at entraining the

circadian rhythm, it is not surprising that varying the timing of exposure to light could disrupt the

circadian rhythm. Variability in light exposure was not significantly associated with sleep

outcomes for older adults. Consistent with the hypothesis that it is the strength of the zeitgeber

that is influential rather than the presence/absence (Monk, 2010), it is possible that the variations

in light were not significant enough in magnitude to influence the sleep of older adults.

Additionally, due to a reduction in pupil diameter, as well as a yellowing of the lens with age, the

amount of light that reaches the retina in older adults may be attenuated to begin with (Hughes &

Neer, 1981; Teresi et al., 1994). As a result, the strength of variations in light exposure may

have to be even greater for older adults to compensate for age-related changes.

Fluctuations in arousal for all arousal measures for younger adults and for two measures

for older adults were associated with worse sleep. Previous research has shown that overall

higher levels of arousal can have a disrupting effect on sleep (Aschoff, 1965; Espie et al., 1989;

Nicassio et al., 1985). These results are interesting because they suggest that in addition to

higher levels of arousal being associated with worse sleep,fluctuations in the level of arousal are

also associated with poorer sleep. Consequently, it may be helpful to not only reduce the amount

of arousal, but also to aim for more stable levels of arousal.

Variability in affect was associated with worse sleep for younger but not older adults.

Given that older adults showed the least intraindividual variability for the affect variables, it is









Bouchareb, & Touitou, 2001). Cortisol and melatonin rhythms were measured in individuals

before and during the observation of the Muslim holiday of Ramadan. During Ramadan,

individuals refrain from eating, drinking, smoking, and sexual relations from sunrise to sunset

(Bogdan et al., 2001). The authors found that changes in the timing of meals during Ramadan

resulted in significant changes in cortisol and melatonin rhythms.

McClintock (1971) described the phenomenon of synchronized menstrual cycles among

women who live together. The author assessed the onset of the menstrual period among female

college residents residing in a dormitory (aged 17 22 years). Greater synchronicity in

menstrual cycles was observed among close friend groups and roommates compared to random

pairings in the dormitory. To examine whether the synchronicity was a response to common

exposure to light/dark periods, McClintock compared synchronicity of menstrual cycles between

close friends and roommates, assuming that roommates would share more similar light/dark

exposure patterns. The results (although not significant) showed a trend towards undermining a

photoperiodic effect with roommates sharing less synchronicity (p < 0.007) than close friends (p

< 0.003; McClintock, 1971). In summary, it appears that social stimuli can entrain circadian

rhythms both indirectly and directly as evidenced in both animal and human studies.

Regularity of Social Zeitgebers

In addition to studying the presence/absence of social cues and their effect on circadian

systems, research has begun to examine the regularity of social cues (Monk, Kupfer, Frank, &

Ritenour, 1991). Regularity or rhythmicity refers to an organism's tendency to respond to social

stimuli in an orderly manner at fixed or uniform intervals (Mish, 2000). Lifestyle regularity is

likely determined by a variety of factors including biological, genetic, individual preference, and

the circumstances in which the individuals are found (Monk, Houck, & Shear, 2006). The body

of research on social rhythms has shown that regularity of social rhythms is associated with two

















7 \--social
rhythmicity
6
-e-sleep
5

4

3

2
1 2 3 4 5 6 7 8 9 10 11 12 13 14

Days in Study


Figure 3-1. Example of hypothesized daily coupling of social rhythmicity and sleep. Y-axis
represents amount of social rhythmicity and sleep.









with worse TWT (r = 0.30,p < .05) and SQR (r = -0.38,p < .01). Higher levels of variability in

the UMACL-R tenseness scale was associated with worse SQR (r = -0.33, p < .05). Higher

levels of variability in the PSAS somatic and PSAS cognitive scales were associated with worse

SOL (r = 0.36,p < .05) and TST (r = -0.30,p < .05), respectively, for older adults (see Table 5-

9).

In terms of variability in affect, higher levels of within person variability were associated

with worse sleep overall for younger adults. More variability in the pleasant grid ratings was

associated with worse SQR (r = -0.31,p < .05). More variability in the PANAS positive scale

was associated with more NWAK (r = 0.29, p < .05) and more variability in the PANAS

negative scale was associated with worse SQR (r = -0.38,p < .01; see Table 5-6). There were no

significant associations between variability in affect and mean-level sleep in older adults.

Aim 3: Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light,
Arousal, and Affect are Coupled with Daily Fluctuations in Sleep

Prior to calculating the multilevel model (MLM), multicollinearity between the predictor

variables of the light, arousal, and affect groups was examined through the estimation of a

multivariate mixed-effects null model (see Tables 5-11 through 5-16). Multicollinearity is an

assumption that is not routinely examined in the MLM literature. In other parametric techniques,

collinearity between predictors becomes especially problemsome at levels exceeding 0.90 (Hair,

Anderson, Tatham, & Black, 1998). In MLM analyses, multicollinearity is of concern when

cross-level interactions are formed (Tabachnick & Fidell, 2007). Given the lack of consensus on

how best to handle Multicollinearity in MLM, and given the lack of cross-level interactions in

the present study, variables that were correlated at or above a level of 0.80 were residualized to

control for the covariance. There were two correlations that met or exceeded a correlation value

of .80 at the between person level (the arousal grid and UMACL-R energetic for younger adults r









of the analyses, examining day-to-day associations, was warranted. The importance of aim 1 is

the identification of microlevel changes in behavior. Historically, the study of aging has studied

changes at the macrolevel (Ram et al., 2009) occurring over years and decades. For example, the

study of fluid and crystallized intelligence over the lifetime. As a result, the fluctuations or

inconsistencies of short-term processes have been overlooked. As demonstrated by the

comparison of mean-level versus intraindividual level outcomes above, often times an age-

related difference at the mean level is qualified by significant intraindividual differences.

Examining behaviors at the microlevel captures the complexity of psychological variables.

Interestingly, older adults exhibited considerably less intraindividual variability compared

to younger adults on most measures other than sleep. Rocke and colleagues (2009) stressed the

importance of examining intraindividual variability in addition to mean-level relationships given

the assumption that aging is associated with increased variability. Previous research has

demonstrated increases in variability with aging for physiological (Lipsitz, 2002), cognitive

(Hultsch, Strauss, & Hunter, 2008), and sensorimotor functioning (Woollacott, 2000). In turn,

variability has been viewed as indicative of poorer functioning or a fault in the integrity of the

system (Nesselroade, 1991). For example, variability in cognitive functioning can be indicative

of neurological dysfunction. The view of increasing variability with age, and the view that the

increased variability is associated with negative outcomes, is consistent with a loss-oriented view

of aging. The results from the present study, however, present a more gain-oriented scenario

where older age is associated with less variability. The direction of the present results are

consistent with recent findings (Rocke et al., 2009) suggesting that some constructs become or

remain more stable with age. The majority of social rhythms, light, arousal, and affect variables










Table 5-5. Correlations of within person variability in arousal and mean-level sleep in younger adults. (N=50)
1 2 3 4 5 6 7 8 9 10 11
1. arousal grid 1
2. PSASsomatic 0.38** 1
3. PSAScognitive 5.11E+07 0.36** 1
4. UMACL-R energetic 4.87E+07 0.21 0.39** 1
5. UMACL-R tenseness 6.17E+07 0.37** 0.21 0.38** 1
6. SOL 0.25 0.15 0.36* 0.12 0.15 1
7. WASO 0.37** 0.32* 0.31* 0.24 0.05 0.32* 1
8. TWT 0.42** 0.24 0.27 0.30* 0.23 0.75*** 0.57*** 1
9. TST -0.003 -0.07 -0.01 -0.19 -0.14 -0.26 -0.06 -0.34* 1
10. NWAK 0.23 0.31* 0.24 0.22 0.20 0.27 0.44** 0.30* -0.13 1
11. SQR -.334* -0.32* -0.21 -0.38** -0.33* -0.13 -0.20 -0.23 0.43** -0.17 1
Note: *** Correlation is significant at the 0.001 level (2-tailed), ** correlation is significant at the 0.01 level (2-tailed), and *
Correlation is significant at the .05 level (2-tailed).
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating.









Reliability and validity of the UMACL-R was examined in a sample of undergraduate

students. Cronbach's alpha estimates ranged from 0.82 to 0.87 for the energetic arousal scale

and 0.78 to 0.83 for the tense arousal scale. Convergent validity estimates for the UMACL-R

and the Profile of Mood States for Adolescents (POMS-A) were 0.84 for the energetic arousal

scale, 0.68 for the tension arousal scale (Terry, Lane, Lane, & Keohane, 1999). Convergent

validity estimates for the UMACL-R and Thayer's Activation-Deactivation Adjective Checklist

(AD ACL) ranged from -0.79 to -0.85 for the energetic arousal scale and -0.60 to -0.69 for the

tense arousal scale (Thayer, 1986).

Affect

The Positive and Negative Affect Scale (PANAS; Appendix 2) was completed each day by

the participants. The PANAS is a measure of mood states and consists of two 10-item mood

scales asking participants to rate specific feelings and emotions associated with positive affect

(e.g., happy) and negative affect (e.g., distress). Individuals respond to each item using a likert

scale ranging from 1 (very slightly or not at all) to 5 (extremely) indicating how they feel at the

moment (Watson et al., 1988).

The PANAS has adequate reliability with internal consistency estimates for college

students ranging from Cronbach's alpha .85 (negative affect scale) to .89 (positive affect scale;

Watson et al., 1988). For community-dwelling adults, the internal consistency of the negative

affect scale was Cronbach's alpha of .87 and for the positive affect scale .86 (Watson et al.,

1988). Test-retest correlations of the positive affect scale within college samples ranged from

.47 to .68 while the test-retest correlation of the negative affect scale ranged from .39 to .71

(Watson et al., 1988).

The PANAS has good criterion validity for both college students and community-dwelling

adults with the negative affect scale correlating with other brief inventories of negative affect












15. Loss of Energy
S0 have as much energy as ever
O 1 I have less energy than I used to have
O 2 don't have enough energy to do very much
O 3 don t have enough energy to do anything
16. Changes in Sleeping Pattern
S 0 I have not experienced any change in my sleeping pattern
O la. I sleep somewhat more than usual.
O lb. I sleep somewhat less than usual.
O 2a. I sleep a lot more than usual.
O 2b. I sleep a lot less than usual.
O 3a. I sleep most of the day
O 3b I wake up 1-2 hours early ard can't get back to sleep.
17. Irritability
O 0 I am no more irritable than usual.
S1 I am more irrtable than usual.
S2 I am much more irritable than usual.
O 3 am irritable all the time
18. Changes in Appetite
O 0 I have not experienced any change in my appetite
O la. My appetite is somewhat less than usual.
O lb My appetite is somewhat greater than usual
O 2a. My appetite is much less than before.
O 2b My appetite is much greater than usual
O 3a I have no appetite at all.
O 3b. I crave food all the time,









quality. Also, overall more positive affect was associated with more total sleep time while

negative affect was associated with fewer awakenings.

Aim 4: Identify to What Extent Light, Arousal, and Affect Mediate the Relationship
Between Social Rhythmicity and Sleep

The goal of aim was to investigate the possible mediating effects of light, arousal, and

affect on the relationship between social rhythmicity and sleep. This analysis was to be

investigated using a four-step mediation model proposed by Baron and Kenny (1986). In order

to demonstrate mediation, both the predictor (SRM) and the mediators (light, arousal, and affect)

needed to be significantly associated with the outcome variables (sleep). While significantly

associations existed between variables for both younger and older adults, there were no

significant (p < .05) relationships that were the same for both SRM and sleep. Consequently, the

mediating effects of light, arousal, and affect could not be estimated as the first two steps of

Baron and Kenny's model could not be achieved.









(older adults rated their health status as poorer than younger adults 3.48 versus 3.84). For

mean-level differences, older adults reported significantly higher cognitive arousal than younger

adults. This finding is interesting given that the cognitive arousal occurred prior to bed and

involved endorsing items such as "worry about falling asleep", "review or ponder events of the

day", and "being distracted by noises in the environment". Higher levels of cognitive arousal for

older adults could reflect concern about sleep (60% of the older sample met criteria as "poor

sleepers") or, conversely, the poorer sleep of the older adult sample could be a reflection of the

higher level of cognitive arousal. Additionally, the endorsement of specific items such as "being

distracted by noises in the environment" could reflect genuine age-related changes in sleep such

as spending more time in the lighter stages of sleep.

The amount of intraindividual variability in affect was greater for younger (39 to 66%)

versus older adults (14 to 44%). Less variability in affect was observed for older adults than

found in previous studies (e.g., 27% for positive and 28% for negative affect from McCrae et al.,

2008). The age-related differences are consistent with previous research showing that older

adults have less intraindividual fluctuations in affect compared to younger adults (Rocke, Li, &

Smith, 2009). Possible explanations for the lower levels of variability in affect measures for

older adults are that 1) older adults have more stable external environments/lifestyles and are

therefore less vulnerable to negative experiences (as evidenced by findings on social rhythmicity

in this study and from Bouisson & Swendsen, 2003); 2) over time older adults have developed

skills to regulate emotions (Carstensen et al., 2000); 3) evidence for habituation processes with

age (Frijda, 1988); and 4) increasing use of life-management techniques with age to regulate

emotion (Carstensen, Isaacowitz, & Charles, 1999). For example, according to the

Socioemotional Selectivity Theory, when time is perceived to be limited (e.g., older age),










Table 5-20. Multilevel model with light variables predicting sleep in older adults
Fixed Effects
SOL WASO TWT TST NWAK SQR
Predictor Variable B (SE) t (d) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df)
Within person
time -0.21 -0.84 -0.46 -1.46 -0.39 -0.62 0.80 1.05 -0.03 -2.51 -0.01 -1.32
(0.25) (71.15) (0.31) (539.72) (0.63) (69.65) (0.76) (695.41) (0.01) (53.18)* (0.01) (66.71)
AMlightcentered 0.24 0.36 -1.70 -2.20 -1.55 -1.01 5.13 2.36 -0.04 -1.51 0.03 1.55
(0.67) (45.38) (0.78) (547.99)* (1.53) (40.36) (2.18) (1098.27) (0.03) (533.98) (0.02) (33.46)
PMlightoentered 0.49 0.83 0.78 0.79 0.46 0.26 -4.22 -1.89 0.0001 0.003 -0.005 -0.18
(0.59) (533.94) (0.99) (77.75) (1.81) (74.13) (2.23) (74.04) (0.04) (67.14) (0.03) (68.07)
light durationcentered -0.01 -0.61 -0.04 -2.21 -0.03 -0.96 0.16 3.69 -0.001 -1.89 0.0002 0.51
(0.01) (569.51) (0.02) (580.29)* (0.03) (570.53) (0.04) (969.90) (0.001) (548.88) (0.0004) (569.51)
light intensitycentered 0.004 1.02 -0.01 -1.20 -0.01 -1.07 -0.03 -2.29 -0.0001 -0.90 8.71E-5 0.67
(0.004) (39.77) (0.005) (550.42) (0.01) (52.44) (0.01) (638.82) (0.0002) (535.17) (0.0001) (496.89)


Between person
AMlightme n

PMlightme.n

light durationme

light intensitymean


0.33 0.17 2.94 1.18 6.62 1.34 7.43 1.14 0.18 1.22 -0.06 -0.92
(1.97) (51.44) (2.49) (55.24) (4.94) (50.66) (6.52) (52.46) (0.15) (48.04) (0.06) (54.82)
-0.87 -0.49 -2.34 -1.03 -5.22 -1.16 -8.91 -1.50 -0.13 -0.90 0.14 2.37
(1.80) (51.06) (2.27) (54.63) (4.51) (50.33) (5.95) (52.04) (0.14) (47.95) (0.06) (54.31)*
-0.4 -1.55 0.01 0.18 -0.03 -0.40 0.22 2.41 0.002 0.73 -3.61E-5 -0.04
(0.03) (48.76) (0.03) (52.21) (0.07) (48.18) (0.09) (49.40)* (0.002) (47.10) (0.001) (50.68)
0.03 2.29 0.008 0.55 0.04 1.38 -0.07 -1.82 -0.0003 -0.37 -0.001 -1.38
(0.01) (47.67)* (0.02) (50.34) (0.03) (47.12) (0.04) (48.69) (0.001) (46.74) (0.0004) (49.94)


Random Effects
SOL WASO TWT TST NWAK SQR
B (SE) Z B(SE) Z B(SE) Z B (SE) Z B (SE) Z B (SE) Z
Within person
time 0.57 1.19 0.00 0.00 3.76 1.27 0.00 0.00 0.002 1.26 0.001 1.58
(0.48) (2.95) (0.001) (0.001)
AMlightcentered 3.78 1.20 0.00 0.00 12.61 0.80 0.00 0.00 0.00 0.00 0.002 0.54
(3.15) (15.70) (0.003)
PMlightoentered 0.00s 0.00 11.97 2.01* 45.56 2.20* 47.75 1.59 0.02 2.34* 0.007 1.68
(5.96) (20.68) (30.11) (0.01) (0.004)
light duratiocentered 0.009 0.00 0.009 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
light intensitycentered 3.91E-5 0.39 0.00 0.00 0.001 1.11 0.00 0.00 0.00 0.00 0.00s 0.00
(0.001)
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep
quality rating. *Correlation is significant at the p < 0.05 (2-tailed), **correlation is significant at the p < 0.01 (2-tailed),***correlation is significant at the p <
0.001 (2-tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, although all convergence criteria were satisfied.









provide evidence for the association between variability and sleep. Examining variability in a

broader research context suggests that variability in physiological functioning is associated with

poorer overall functioning (Britton, 1997; Pagani, 1999). Additionally, individuals who are

diagnosed with insomnia typically have highly variable sleep patterns (Edinger, Marsh, McCall,

Erwin, & Lininger, 1991). Therefore, although this aim was exploratory to a degree, it was

hypothesized that in general, higher levels of variability in the predictor variables would be

associated with poorer sleep.

Aim 3: Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light,
Arousal, and Affect are Coupled with Daily Fluctuations in Sleep

Analysis

The goal of aim two was to examine the ability of social rhythmicity, light, arousal, and

affect to predict sleep on a daily level and mean level. A multilevel (MLM) approach was used

to examine how daily data for the predictor variables predicts sleep. MLM, also known as

hierarchical linear modeling or mixed effects modeling, is an alternative to univariate or

multivariate analysis of repeated measures. MLM allows for the analyses of change in

individual cases over an extended time period (Tabachnick & Fidell, 2007). MLM is an

extension of the general linear model and does not require observations to be independent.

Additionally, MLM allows all available data to be included in the analyses as it assumes random

missing data and consequently does not exclude a case due to one missing data point (Bryk &

Raudenbush, 1992). MLM is particularly suited to the analysis of daily data due to the

hierarchical structure of the data, with daily observations nested within the same participant

(Singer, Davidson, Graham, & Davidson, 1998; Singer, Fuller, Keiley, & Wolf, 1998; Singer &

Willett, 2003).









LIST OF FIGURES


Figure page

2-1 Mistleberger & Skene's (2004) theoretical model of temporal influences on
biological rhythm s. .........................................................................44

2-2 Social Zeitgeber Theory and Internal Trigger Hypothesis .............................................45

2-3 Mediated Model of Social Rhythmicity examined for both age groups in the present
study........................................................ 46

2-4 Cognitive M odel of Insomnia (Harvey, 2000) ............................................................ 47

3-1 Example of hypothesized daily coupling of social rhythmicity and sleep.........................56

5-1 Percentage of total variability in social rhythmicity that can be attributed to within
person variability. .......................................... ............................ 85

5-2 Percentage of total variability in light measures that can be attributed to within
person variability. .......................................... ............................ 86

5-3 Percentage of total variability in arousal measures that can be attributed to within
person variability. .......................................... ............................ 87

5-4 Percentage of total variability in affect measures that can be attributed to within
person variability. .......................................... ............................ 88

5-5 Percentage of total variability in sleep variables that can be attributed to within
person variability. .......................................... ............................ 89

6-1 Relationships in the mediated model of social rhythmicity that were observed in
younger adults................................................................................................. 136

6-2 Relationships in the mediated model of social rhythmicity that were observed in
o ld e r a d u lts ...................................... ................................................... 1 3 7









This result is consistent with previous research that has shown that regularity increases with age

(Monk et al., 1994; Monk et al., 1997).

The second estimate of variability was the intraindividual variability (the day-to-day)

variation in regularity scores. Both younger and older adults were highly variable in the

regularity of their daily activities. Of the total variability to be measured, 77% and 72% was due

to intraindividual variations in younger and older adults respectively (younger adults were

significantly more variable within person compared to older adults). These results indicate that

although older adults are more regular overall, there is still considerable variation day-to-day in

how regular they are in their daily activities. In fact, older adults vary more in their behaviors

within person than they do between person. The differences in the overall and day-to-day

measures of regularity are important as they provide different views of the level of routine in

older adults' lives. Over a two-week period, older adults generally appear more regular in the

timing of their daily activities compared to younger adults. On a day-to-day basis, however, an

older adult's daily routine is more likely to resemble the routine of another individual than to

resemble their own routine. This variation on a daily basis is considerably less than younger

adults but still suggests there is more variation in the timing of daily events for older adults than

would be seen simply by summing the two weeks of data.

Light in the present study was examined subjectively using the Light Exposure Scale. The

Light Exposure Scale was designed specifically for this study due to the lack of subjective

measures for assessing light. While the Light Exposure Scale has apparent face validity in terms

of the measures of timing, duration, and intensity of light exposure, there are no established

reliability or validity estimates for the measure. Consequently, the results must be interpreted

with caution and have limited generalizability. Intraindividual variability in light has not, to our












Please estimate the type of light you received for each hour of the previous day. For example, if from 8:20 am to 9 00 am
you were in bright .riiijlht ,;.1i woukild select 'bright sunlight for the response for'8 9 am'.
If you have difficulty remembering the type of light for part of the day, please make your best guess.


8- am
9- lam
10-11 am
11am- 12noon
12noon -1 pm
S-2pm
2-3pm
3-4pm
4- pm
5- 6pm
6-7pm
7 8pm
8- pm
9- l~r
9 10pm
10-11pm
11pm 12am
12-1 am
1-2am
2- am
3 -4 am
4-5 am
5-6am
6-7 am
7-8am
Comments'


darkAmoonlight
0
0
0
0
0
0
0
O
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0


dim light
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0


indoor eight
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O0


cloudy
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O


bright sunlight
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
O
O
O
O


ii
















In the past, has a mental health professional (psychiatrist, psychologist, or social
worker) ever treated you ?

Syes
Sno


If yes, explain:
If yes, explain:


Are you currently being treated by a mental health professional?
Syes
0no

OIf yes, explain:
If yes, explain:
II











Table 5-21. Multilevel model with arousal variables predicting sleep in younger adults
Fixed Effects


Predictor Variable
Within person


SOL WASO TWT TST NWAK SQR
B(SE) t(dfJ B(SE) t(dJ) B (SE) t (dJ) B (SE) t (d) B (SE) t (d) B (SE) t (d)


time -0.08 -0.85 0.08 0.67 0.12 0.49 -1.73 -1.65 -0.0004 -0.04 -0.01 -1.97
(0.10) (604.20) (0.12) (582.53) (0.26) (613.51) (1.05) (130.20) (0.01) (112.48) (0.01) (145.51)*
arousal gridcentered -0.56 0.72 0.73 1.38 1.10 1.15 -1.17 -0.31 0.003 0.08 -0.01 -0.41
(0.49) (19.85) (0.53) (65.40) (0.95) (50.05) (3.74) (50.85) (0.04) (64.61) (0.03) (55.64)
PSAS somatiCcentered 0.35 0.72 0.67 0.91 0.91 0.78 -0.49 -0.10 0.01 0.27 -0.09 -3.26
(0.49) (19.85) (0.74) (23.84) (1.17) (17.74) (4.94) (22.82) (0.04) (16.21) (0.03) (570.53)**
PSAS cognitivecentered 1.11 4.93 0.87 2.47 2.09 4.11 -4.47 -3.28 0.06 3.05 -0.05 -4.69
(0.22) (45.44)*** (0.35) (23.09)* (0.51) (35.37)*** (1.36) (549.78)** (0.02) (35.35)** (0.01) (33.80)***
UMACL-R energeticcentered -0.24 -1.66 0.15 1.04 -0.18 -0.47 -0.90 -0.73 -0.01 -0.62 0.03 2.89
(0.15) (71.22) (0.14) (559.82) (0.38) (68.08) (1.23) (54.44) (0.01) (52.41) (0.01) (62.42)**
UMACL-R tensecentered 0.05 0.33 0.23 1.03 0.34 0.91 -1.97 -1.50 0.02 1.44 -0.01 -1.22
(0.15) (56.36) (0.22) (39.15) (0.37) (44.56) (1.32) (487.05) (0.02) (49.37) (0.01) (58.80)


Between person
arousal gridme n


PSAS somaticme a

PSAS cognitiveme a

UMACL-R energeticme n

UMACL-R tensemem


5.61 2.43 0.15 0.09 7.69 1.59 -29.65 -1.95 -0.24 -1.33 -0.21 -1.78
(2.31) (49.92)* (1.73) (49.65) (4.84) (49.62) (15.17) (49.05) (0.18) (49.76) (0.12) (47.92)
0.02 0.02 1.40 1.29 3.42 1.12 17.50 -1.83 0.04 0.33 -0.07 -0.92
(1.46) (49.86) (1.09) (49.57) (3.05) (49.54) (9.57) (49.25) (0.11) (49.72) (0.07) (48.01)
1.74 3.47 0.64 1.71 2.37 2.26 0.008 0.003 0.06 1.44 -0.02 -0.86
(0.50) (50.14)** (0.38) (50.28) (1.05) (49.96)* (3.30) (49.98) (0.04) (50.28) (0.03) (48.38)
-2.68 -2.50 0.24 0.30 -2.81 -1.26 9.74 1.39 (49.23) 0.09 1.11 0.11 2.04
(1.07) (49.95)* (0.80) (49.75) (2.24) (49.67) (7.03) (0.08) (49.74) (0.05) (47.93)*
-1.21 -2.83 -0.30 -0.94 -1.72 -1.91 -1.11 -0.40 0.03 0.80 -0.03 -1.62
(0.43) (50.01)** (0.32) (49.76) (0.90) (49.72) (2.82) (49.19) (0.03) (50.003) (0.02) (48.39)









In summary, it appears that the circadian clock can be influenced by both photic (light)

and non-photic (behavioral state and social stimuli) sources. While light may directly influence

circadian rhythms, it is believed that social stimuli and behavioral state primarily influence

circadian rhythms through their respective impact on light (Mistlberger et al., 2000).

Considering that social stimuli can exert effects on the circadian system, and consequently

impact sleep, it is important to examine the particular social stimuli associated with circadian

rhythm changes.

Social Rhythms

Social Zeitgebers

Social nonphotic cues (zeitgebers other than light) have been shown to be important for

entrainment in both animal species and humans (Aschoff et al., 1971). Research indicates that

social entrainment begins soon after birth (Monk et al., 1990). Social stimuli can be important

zeitgebers for maintaining daily rhythms (e.g., appropriate peaks and troughs) which are

essential in a fluctuating daily rhythm (Monk et al., 1990). For example, the social stimuli of

leaving for work or school may help to energize or alert the individual to prepare them for

starting their day, while watching television or reading in the evening may help to relax the

individual in preparation for sleep. To elaborate on the earlier definition of social stimuli,

important social rhythm markers for human beings can include meals, starting of work, leaving

home, bedtime/waketime, exercise, watching television, and interacting with other individuals.

Evidence for Social Zeitgebers

As mentioned earlier, due to the pervasiveness of the rising and the setting of the sun, light

is the most powerful zeitgeber for most species. Recently, however, researchers have begun to

focus on another area of importance for regulating our lives the activity of other members of

our species. It is hypothesized that circadian rhythms in some species may be able to be reset not









explanations for the lack of association include the discrepancy between subjective and objective

measurement not unlike what is observed when assessing sleep using sleep diaries versus

Actiwatch-L. It is possible that the perception of light exposure is a different construct than the

objective assessment obtained through the Actiwatch-L. An additional explanation is that the

LES used a retrospective assessment format where individuals had to recall activities the day

earlier. This method could have resulted in a different memory of the experience of light

compared to what was captured using the Actiwatch-L device. Additionally, the geographical

location of the majority of study participants, southeastern United States, could have affected the

amount of light exposure participants received. Greater variability in light exposure could have

resulted from a different study location. Finally, the instructions on the LES asked participants

to record the brightest amount of sunlight they were exposed to in each hour. Consequently,

participants may have only been exposed to five minutes of bright sunlight but the LES measure

would reflect a full hour of bright sunlight exposure. Therefore, there could be overestimations

of light exposure with the LES compared to the Actiwatch-L device.

Arousal

The Affect Grid, a scale measuring pleasure and arousal was completed each day by the

participants (Russell, Weiss, & Mendelsohn, 1989; see Appendix 2). The Affect Grid is

designed to capture two dimensions of affect and arousal: pleasure-displeasure and arousal-

sleepiness. A variation of the grid was formatted for online use. For the two items, participants

were asked to indicate how they generally felt the day before ranging from high arousal to

sleepiness and unpleasant to pleasant feelings (9 point scale ranging from 1-9). The resulting

variables are the arousal grid variable which indicates level of arousal and the pleasant grid

which indicates the level of pleasant feelings.










Table 5-19. Continued
Random Effects
SOL WASO TWT TST NWAK SQR
B(SE) Z B (SE) Z B(SE) Z B(SE) Z B (SE) Z B (SE) Z
Within person
time 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0004 0.85
(0.001)
AMIightcentered 0.00s 0.00 0.29 0.43 1.16 0.69 0.00 0.00 0.001 0.20 0.002 (0.002) 0.80
(0.66) (1.68) (0.004)
PMlightoentered 0.00s 0.00 0.71 0.86 0.009 0.00 0.00 0.00 0.00 0.00 0.00 0.00
(0.82)
light duratiocentered 0.009 0.00 0.009 0.00 0.00 0.00 0.01 0.51 0.00 0.00 0.00 0.00
(0.02)
light 0.00 0.00 2.05E-5 0.93 0.00 0.00 0.00 0.00 4.20E-7 1.63 6.95E-8 0.81
intensitycentered (2.21E-5) (2.57E- (8.58E-8)
7)

Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating. *Correlation is significant at thep < 0.05 (2-tailed), **correlation is significant at thep <
0.01 (2-tailed),***correlation is significant at thep < 0.001 (2-tailed), Variance is too small to be estimated; the final Hessian matrix
was not positive definite, although all convergence criteria were satisfied.









Table 5-8. Correlations of within person variability in light and mean-level sleep in older adults. (N=50)
1 2 3 4 5 6 7 8 9 10
1. AMlight 1.00
2. PMlight 0.19 1.00
3. light duration 0.27 0.04 1.00
4. light intensity 0.12 -0.18 0.58*** 1.00
5. SOL -0.09 0.12 0.15 0.10 1.00
6. WASO -0.03 -0.17 0.02 0.15 0.17 1.00
7. TWT -0.09 -0.02 0.04 0.16 0.71*** 0.69*** 1.00
8. TST -0.17 -0.05 -0.15 -0.04 -0.28* -0.07 -0.31* 1.00
9. NWAK -0.26 -0.03 -0.22 -0.17 0.04 0.53*** 0.33* -0.04 1.00
10. SQR 0.06 -0.05 -0.13 -0.03 -0.29* -0.57*** -0.51*** 0.11 -0.50*** 1.00
Note: *** Correlation is significant at the 0.001 level (2-tailed), ** correlation is significant at the 0.01 level (2-tailed), and *
Correlation is significant at the .05 level (2-tailed).
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating.













A number of statements which people have used to describe themselves are given
below. Read each statement and then select the appropriate circle to indicate how you
have generally felt over the past two (2) weeks.

There are no right or wrong answers. Do not spend too much time on any one
statement but give the answer which seems to describe your present feelings best.
not at all somewhat moderately so very much so
I f- c 0 0 0 0
I feel secure 0 0 0 0
Sam tnse 0 0 0 0
O O O O
I feel strained 0 0 0 0
Sfeel strained 0 0 0 0
I feel at ease O 0 0
Sfeel upset 0 0 0 0
'- "O O O O
I am presently worrying O OC 0
over possible misfortunes
I feel satisfied 0 0 0 0
Sfeel frightened 0 0 0 0
I feel comfortable O O) 0
I feel self-confident ) 0 ) O
feelnevos 0 0 0 0
,,, jitte 0 0 0 0
I feel indecisive 0 0 0 0
,.oo, O O O O
Sa...elaed 0 0 0 0
I feel content 0 0 0 0
1- a worr 0 0 0 0
Sfee confused 0 0 0 0
Sfeel steady 0 0 0 0
Feel pleasant 0 0 0 0













Monk, T. H., Buysse, D. J., Hall, M., Nofzinger, E. A., Thompson, W. K., Mazumdar, S. A., et
al. (2006). Age-related differences in the lifestyle regularity of seniors experiencing
bereavement, care-giving, insomnia, and advancement into old-old age. Chronobiology
International, 23(4), 831-841.

Monk, T. H., Flaherty, J. F., Frank, E., Hoskinson, K., & Kupfer, D. J. (1990). The Social
Rhythm Metric An instrument to quantify the daily rhythms of life. Journal of Nervous
and Mental Disease, 178(2), 120-126.

Monk, T. H., Houck, P. R., & Shear, M. K. (2006). The daily life of complicated grief patients--
What gets missed, what gets added? Death Studies, 30(1), 77-85.

Monk, T. H., Kupfer, D. J., Frank, E., & Ritenour, A. (1991). The Social Rhythm Metric (SRM):
Measuring daily social rhythms over 12 weeks. Psychiatry Research, 36(2), 195-207.

Monk, T. H., Petrie, S. R., Hayes, A. J., & Kupfer, D. J. (1994). Regularity of daily-life in
relation to personality, age, gender, sleep quality and circadian-rhythms. Journal of Sleep
Research, 3(4), 196-205.

Monk, T. H., Reynolds, C. F., III, Buysse, D. J., DeGrazia, J. M., & Kupfer, D. J. (2003). The
relationship between lifestyle regularity and subjective sleep quality. Chronobiology
International, 20(1), 97-107.

Monk, T. H., Reynolds, C. F., Kupfer, D. J., Hoch, C. C., Carrier, J., & Houck, P. R. (1997).
Differences over the life span in daily life-style regularity. Chronobiology International,
14(3), 295-306.

Monk, T. H., Reynolds, C. F., Machen, M. A., & Kupfer, D. J. (1992). Daily social rhythms in
the elderly and their relation to objectively recorded sleep. Sleep, 15(4), 322-329.

Mroczek, D. K., & Kolarz, C. M. (1998). The effect of age on positive and negative affect: A
developmental perspective on happiness. Journal ofPersonality and Social Psychology,
75(5), 1333-1349.

Mrosovsky, N. (1996). Locomotor activity and nonphotic influences on circadian clocks.
Biological Reviews of the Cambridge Philosophical Society, 71(3), 23.

National Sleep Foundation (2008). Sleep in America Poll Summary ofFindings. Washington,
D.C.

Nemiah, J. C., & Sifneos, P. E. (1970). Psychosomatic Illness Problem in Communication.
P%)L /t 11/11htheq /'' and Psychosomatics, 18(1-6), 154-160.









49%. Younger adults were significantly more variable within persons compared to older adults

for AM light t(96) = 3.85,p <.001, PM light t(97) = 3.67,p <.001, light duration t(97) = 3.25, p

<.01, and light intensity t(98) = 4.12, p <.001. Older adults were exposed to significantly earlier

AM light t(97) = 2.94,p <.01 and more intense light t(98) = -2.89,p <.01 (see table 5-2) than

younger adults.

Arousal Variables

The ICC for the arousal variables ranged from 0.29 to 0.50 for younger adults and 0.47 to

0.78 for older adults (see Figure 5-3). For younger and older adults respectively, the amount of

within person variability in the arousal grid was 50% and 39%, for the PSAS somatic scale 64%

and 53%, for the PSAS cognitive scale 71% and 31%, for the UMACL-R energetic scale 64%

and 22%, and for the UMACL-R tenseness scale 54% and 22%. Younger adults were

significantly more variable within persons compared to older adults for the arousal grid t(97) =

4.18, p <.001, the PSAS cognitive scale t(97) = 3.02, p <.01, the UMACL-R tenseness scale

t(97) = 4.45,p <.001, and the UMACL-R energetic scale t(97) = 6.07,p <.001. In terms of

mean-level values, older adults reported significantly higher cognitive arousal according to the

PSAS cognitive scale t(98) = -3.27, p <.01 compared to younger adults (see table 5-2).

Affect Variables

The ICC for the affect variables ranged from 0.34 to 0.58 for the younger adults and 0.56

to 0.89 for the older adults (see Figure 5-4). The percentage of overall variability that could be

attributed to fluctuations within persons, respectively for younger and older adults, was 66% and

44% for the pleasant grid, 42% and 20% for the PANAS positive subscale, and 59% and 11% for

the PANAS negative subscale. Younger adults were significantly more variable within persons

compared to older adults for the pleasant grid t(95) = 2.90, p <.01, the PANAS positive scale

t(97) = 6.82,p <.001, and the PANAS negative scale t(94) = 4.08,p <.001. For mean-level









wake time during the night after falling asleep. Also, the number of awakenings older adults

reported decreased over the time of the study. At the between person level, PM light was

associated with SQR (P = 0.14, t(54.31) = 2.37, p <.05), light duration was associated with TST

(P = 0.22, t(49.40) = 2.41,p <.05), and light intensity was associated with SOL (P = 0.03,

t(47.67) = 2.29, p <.05). These results suggest that overall later exposure to light in evening was

associated with overall better sleep quality ratings, and higher levels of overall duration of light

exposure and light intensity were associated with more total time spent sleeping and longer time

to fall asleep, respectively.

For the relationship between the arousal variables and sleep in younger adults, there were

significant predictors at both level 1 and level 2 (see table 5-21). At level 1, the PSAS cognitive

subscale significantly predicted SOL (P = 1.11, t(45.44) = 4.93,p <.001), WASO (P = 0.87,

t(23.09) = 2.47, p <.05), TWT (P = 2.09, t(35.37) = 4.11, p <.001), TST (P = -4.47, t(549.78)= -

3.28, p <.01), NWAK (P = 0.06, t(35.35) = 3.05, p <.01), and SQR (P = -0.05, t(33.80) = -4.69, p

<.001). These results suggest that on days when younger adults have higher ratings of cognitive

arousal before sleep, they take longer to fall asleep, are awake for longer during the night, have

less total sleep time, more awakenings, and lower sleep quality ratings. Random effects were

significant for the PSAS cognitive scale (p < .05 top < .01) predicting SOL, WASO, and TWT

suggesting the level 1 effects mentioned above for SOL, WASO, and TWT vary in magnitude by

the individual. Time (P = -0.01, t(145.51) = -1.97,p <.05), the PSAS somatic (P = -0.09,

t(570.53) = -3.26,p <.01), and UMACL-R energetic scales (P = 0.03, t(62.42)= 2.89,p <.01)

also significantly predicted SQR suggesting that on days when younger adults experienced less

physical tension before bed and felt more alert during the day they reported better sleep quality

and that sleep quality ratings decreased over the study time.









Sleep


The Association Between Sleep and Light

Research suggests that light can be associated with sleep in a number of ways. The

timing of the sleep period can be affected by exposure to light. The sleep period was found to be

delayed when individuals were exposed to light in evening (Czeisler et al., 1989). Exposure to

light in the late night/early morning is associated with phase advances (the timing of the sleep

period begins earlier in the day; (Duffy & Wright, 2005). Additionally, it appears that the human

circadian system is responsive to light exposure throughout the day (Duffy & Wright, 2005).

Light has also been found to exert effects on sleep through exposure to varying intensities of

light. Initially, it was believed that bright light exposure (similar to outdoor natural light) was

necessary to entrain circadian rhythms (Wever, 1989). Recently, however, research suggests that

exposure to much less intense levels of light (i.e., indoor lighting) can result in a phase shift

(Duffy & Wright, 2005).

The overall amount of light an individual is exposed to has been associated with overall

time spent asleep during the night, as well as, the amount of time spent awake during the night

(Hood, Bruck, & Kennedy, 2004; Wallace-Guy et al., 2002). Additionally, individuals who are

exposed to insufficient light have been found to have poorer sleep (Campbell & Daston, 1991;

Campbell, Dawson, & Anderson, 1993; Kohsaka et al., 1999; Kohsaka et al., 2000; Lack &

Wright, 1993; Youngstedt, Kripke, Elliott, Baehr, & Sepulveda, 1999).

The Association between Sleep and Arousal

Arousal is investigated in the present study as one of the three mechanisms underlying the

relationship between social rhythmicity and sleep. Arousal is operationalized as referring to both

cognitive and somatic arousal. Arousal may be implicated in the relationship between social

rhythms and sleep, because in order for an individual to begin to fall asleep, the suppression of










Table 5-22. Multilevel model with arousal variables predicting sleep in older adults
Fixed Effects
SOL WASO TWT TST NWAK SQR
Predictor Variable B (SE) t (d) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (d)
Within person
time 0.07 0.31 -0.35 -1.15 -0.36 -0.59 0.34 0.46 -0.02 -1.66 -0.02 -2.25
(0.23) (83.17) (0.31) (71.62) (0.61) (65.04) (0.74) (62.98) (0.02) (49.67) (0.001) (62.28)*
arousal gridcentered -0.20 -0.24 1.00 0.77 0.77 0.43 -1.30 0.53 -0.06 -1.46 -0.04 -1.15
(0.85) (1125.78) (1.30) (27.23) (1.80) (551.30) (2.44) (532.21) (0.04) (26.59) (0.03) (33.39)
PSAS somaticcentered 0.69 0.57 1.27 0.85 -1.46 -0.63 0.89 0.25 -0.01 -0.18 -0.05 -1.29
(1.20) (3185.31) (1.50) (30.17) (2.33) (21.47) (3.52) (21.34) (0.05) (26.56) (0.04) (18.58)
PSAS cognitivecentered 0.94 2.61 2.11 3.04 3.55 3.56 -3.80 -2.48 0.08 4.81 -0.09 -5.25
(0.36) (33.19)* (0.70) (50.85)** (1.00) (40.00)* (1.53) (41.83)* (0.02) (33.81)*** (0.02) (35.01)***
UMACL-R energeticcentered 0.09 0.32 -0.82 -1.72 -0.14 -0.17 -0.91 -0.76 -0.01 -0.77 0.01 0.93 (38.88)
(.29) (590.36) (0.48) (46.12) (0.82) (66.57) (1.19) (49.75) (0.01) (317.77) (0.01)
UMACL-R tensecentered 0.12 0.41 -0.11 -0.27 0.87 1.15 0.35 -0.31 -0.03 -1.70 -0.001 -0.14
(0.29) (559.93) (0.41) (420.95) (0.76) (44.68) (1.13) (38.42) (0.01) (55.18) (0.01) (427.39)
Between person
arousal gridmean 0.53 0.30 1.64 0.72 4.06 0.91 5.22 0.86 -0.17 -1.19 -0.14 -2.96
(1.78) (46.43) (2.28) (47.60) (4.48) (47.34) (6.10) (47.27) (0.14) (47.15) (0.05) (46.45)**
PSAS somaticmean 1.18 0.58 -1.42 -0.55 -1.69 -0.33 1.51 0.22 0.08 0.49 -0.01 -0.18
(2.03) (46.48) (2.60) (48.04) (5.09) (47.13) (6.95) (47.76) (0.16) (47.01) (0.05) (45.25)
PSAS cognitivemean 1.78 2.38 2.58 2.68 5.07 2.69 -4.91 -1.91 0.09 1.40 -0.07 -3.41
(0.75) (46.49)* (0.96) (48.02)* (1.88) (47.15)* (2.58) (47.73) (0.06) (47.09) (0.02) (45.85)***
UMACL-R energeticmean -0.73 -0.43 -0.87 -0.40 -2.99 -0.71 7.90 1.37 0.01 0.07 -0.003 -0.07
(1.68) (46.44) (2.15) (47.85) (4.21) (47.12) (5.75) (47.54) (0.14) (47.03) (0.05) (45.07)
UMACL-R tensemean 0.10 .06 (46.46) -1.27 -0.58 -1.18 -0.28 -3.84 -0.65 -0.12 0.85 0.01 0.20 (45.11)
(1.72) (2.20) (47.89) (4.30) (47.13) (5.88) (47.59) (0.14) (47.07) (0.05)









Table 5-14. Correlations among arousal variables between persons (below diagonal) and within person (above diagonal) for older
adults
PSAS UMACL-R UMACL-R
arousal grid PSAS somatic cognitive energetic tense
arousal grid 0.06 0.02 -0.25 -0.05
PSAS
S0.48** 0.37 -0.11* 0.10*
somatic
PSAS
i 0.34* 0.56** -0.02 0.09*
cognitive
UMACL-R
-U 0.31* -0.07 -0.11 --0.58***
energetic
UMACL-R
S0.25 0.19 0.20
tense -0.90***
Note: *Correlation is significant at thep < 0.05 (2-tailed); **correlation is significant at thep < 0.01 (2-tailed); ***correlation is
significant at the p < 0.001 (2-tailed).

Table 5-15. Correlations among affect variables between persons (below diagonal) and within person (above diagonal) for younger
adults
pleasant grid PANAS positive PANAS negative
pleasant grid 0.32*** -0.31***
PANAS positive 0.62** -0.13**
PANAS negative -0.53** -0.16
Note: *Correlation is significant at thep < 0.05 (2-tailed); **correlation is significant at thep < 0.01 (2-tailed); ***correlation is
significant at the p < 0.001 (2-tailed).

Table 5-16. Correlations among affect variables between persons (below diagonal) and within person (above diagonal) for older adults
pleasant grid PANAS positive PANAS negative
pleasant grid 0.24*** -0.20***
PANAS positive 0.79*** -0.11*
PANAS negative -0.57** -0.36*
Note: *Correlation is significant at thep < 0.05 (2-tailed); **correlation is significant at thep < 0.01 (2-tailed); ***correlation is
significant at the p < 0.001 (2-tailed).













Kunzmann, U., Little, T. D., & Smith, J. (2000). Is age-related stability of subjective well-being
a paradox? Cross-sectional and longitudinal evidence from the Berlin Aging Study.
Psychology andAging,. 15(3), 511-526.

Kupfer, D. J., & Foster, F. G. (1972). Interval between onset of sleep and rapid-eye-movement
sleep as an indicator of depression. Lancet, 2(7779), 684-&.

Kvaal, K., Ulstein, I., Nordhus, I. H., & Engedal, K. (2005). The Spielberger State-Trait Anxiety
Inventory (STAI): The state scale in detecting mental disorders in geriatric patients.
International Journal of Geriatric Psychiatry, 20(7), 629-634.

Lack, L., & Wright, H. (1993). The effect of evening bright light in delaying the circadian-
rhythms and lengthening the sleep of early-morning awakening insomniacs. Sleep, 16(5),
436-443.

Lichstein, K. L., Durrence, H. H., Taylor, D. J., Bush, A. J., & Riedel, B. W. (2003).
Quantitative criteria for insomnia. Behaviour Research and Therapy, 41(4), 427-445.

Lichstein, K. L., & Morin, C. M. (2000). Treatment ofLate-life Insomnia. Thousand Oaks, CA:
Sage Publications Inc.

Lichstein, K. L., Riedel, B. W., & Means, M. K. (1999). Psychological treatment of late-life
insomnia. In R. Schulz, G. Maddox & P. Lawton (Eds.), AnuualReview of Gerontology
and Geriatrics: Vol. 18. Focus on Interventions Research iith Older Adults. (pp. 74-
110). New York: Springer.

Lichstein, K. L., & Rosenthal, T. L. (1980). Insomniacs perceptions of cognitive versus somatic
determinants of sleep disturbance. Journal ofAbnormal Psychology, 89(1), 105-107.

Lipsitz, L. A. (2002). Dynamics of stability: The physiologic basis of functional health and
frailty. Journals of Gerontology Series a-Biological Sciences and Medical Sciences,
57(3), B115-B125.

Lorr, M., Daston, P., & Smith, I. R. (1967). An analysis of mood states. Educational and
Psychological Measurement, 27(1), 89.

Malkoff-Schwartz, S., Frank, E., Anderson, B., Sherrill, J. T., Siegel, L., Patterson, D., et al.
(1998). Stressful life events and social rhythm disruption in the onset of manic and
depressive bipolar episodes A preliminary investigation. Archives of General
Psychiatry, 55(8), 702-707.














Do you currently have AIDS?
0ys
0.-
O no

If yes, explain:
I


If yes, does this interfere with your sleep?
0 yes
0-"
O no

If yes, explain how:
II

Do you currently have high blood pressure?
0 yes
0.-

If yes, explain:
II

If yes, does this interfere with your sleep?
0 yes
0Q


If yes, explain how:
O


If yes, explain how:


Do you currently have a neurological disease such as a seizure disorder or
Parkinson's?
0Qo
0 n.


I













10. Crying
0 I don't cry anymore than I used to
O 1 I cry more than I used to
S2 I cry over every little thing
O 3 I feel like crying, but I can't
11. Agitation
S 0 I am no more restless or wound up than usual.
S1 I feel more restless or wound up than usual.
O 2 I am so restless or agitated that it's hard to stay still.
S3 I am so restless or agitated that I have to keep moving or doing something.
12. Loss of Interest
O 0 I have not lost interest in other people or activities,
O 1 I am less interested in other people or things than before.
S2 I have lost most of my interest in other people or things
S 3 It's hard to get interested in anything.
13. Indecisiveness
O 0 make decisions about as well as ever
O 1 I find it more difficult to make decisions that usual
S 2 I have much greater difficulty in making decisions than used to
S3 I have trouble making any decisions.
14. Worthlessness
S0 do not feel I am worthless
O 1 Idon t consider myself as worthwhile and useful as I used to
S2 I feel more worthless as compared to other people
S3 feel utterly worthless















Do you currently have any other medical problems or handicaps?

0 yes
On-
O no


If yes, explain:
I


If yes, does this interfere with your sleep?

O yes
Sno


If yes, explain how:


List ALL medications (including vitamins and over-the-counter medications)taken within
the past month, how often you take them (e.g. daily, weekly), the time of day, and the
purpose of the medication.

For sleep medications please state how long you have been taking the medication.


Medication 1
Name of medication
Frequency
Time of Day
Purpose


Medication 2
Name of medication
Frequency
Time of Day
Purpose


j
j
j I
j



1 (
1 (
1 (
j












5. Guilty Feelings
0 I don't feel particularly guilty
O 1 I feel guilty over many things I have done or should have done
S2 I feel quite guilty most of the time
O 3 I feel guilty all of the time
6. Punishment Feelings

O 0 I don't feel I am being punished
S1 I feel I may be punished
S2 I expect to be punished.
S3 I feel I am being punished.
7. Self-Dislike
O 0 feel the same about myself as ever
O 1 I have lost confidence in myself
S2 I am disappointed in myself
Q 3 I dislike myself.
8. Self-Criticalness
S 0 I feel the same about myself as ever
O 1 I have lost confidence in myself
S2 I am disappointed in myself.
O 3 I blame myself for everything bad that happens.
9. Suicidal Thoughts or Wishes
O 0 don t have any thoughts of killing myself,
O 1 I have thoughts of killing myself, but I would not carry them out
S2 I would like to kill myself
S3 I would kill myself if I had the chance.









affect, older adults reported significantly more positive affect than younger adults t(98) = -3.87,

p <.001 (see table 5-2).

Sleep Variables

The ICC for the sleep variables ranged from 0.10 to 0.33 for the younger adults and 0.35 to

0.66 for the older adults (see figure 5-5). The amount of within person variability in sleep for

younger and older adults respectively was 67% and 59% for SOL, 90% and 65% for WASO,

78% and 58% for TWT, 84% and 60% for TST, 81% and 34% for NWAK, and 78% and 64%

for SQR. Younger adults were significantly more variable within persons compared to older

adults for SOL t(98) = -2.19, p <.05, WASO t(97) = -3.68,p <.001, TWT t(98) = -2.95, p <.01,

TST t(98)= 8.95,p <.001, NWAK t(96)= 2.11,p <.05, and SQR t(97)= 2.40,p <.05. Age-

related differences in the mean levels of sleep variables included greater SOL t(98) = -3.54, p

<.001, TWT t(98) = -2.10,p <.05, NWAK t(98) = -4.95,p <.001, and less TST t(98) = 2.34,p

<.05 for older adults (see table 5-2).

Aim 2: Identify How Variability in Social Rhythmicity, Light, Arousal, and Affect is
Related to Sleep

There were significant correlations of within person variability in the social rhythmicity

variable and sleep for younger adults (see Table 5-3). More variability in the timing of social

activities was associated with worse overall SOL (r = 0.75,p < .001), WASO (r = 0.38,p < .05),

and TWT (r = 0.33, p < .05). Younger adults who were more variable in the timing of their daily

activities tended to take longer to fall asleep, spend more time awake during the night before

awakening in the morning, and spend more time awake overall. There were no significant

correlations between variability in social rhythmicity and the sleep variables for older adults (see

Table 5-7).









levels have been associated with burnout symptomatology (Pruessner, Hellhammer, &

Kirschbaum, 1999) and heightened levels have been associated with neurotic (Portella, Harmer,

Flint, Cowen, & Goodwin, 2005) and depressive symptoms (Bhagwagar, Hafizi, & Cowen,

2003). The authors found that across ages, there was more variability within than between

persons in cortisol levels and that there was an increase in variability with age for men (Almeida,

Piazza, & Stawski, 2009). While the results from the present study use differing measures of

arousal, the high levels of variability within person for younger adults is consistent with these

findings. Less variability for older adults in the present study may reflect the differences in

measurement type (e.g., psychological versus physiological measures). Additionally, as cortisol

is a hormone tied to the circadian rhythms, it is possible that the age-related increased variability

in cortisol with age in the Almeida, Piazza, and Stawski (2009) study could reflect a weakening

of the circadian rhythm with age resulting in a less robust rhythm that is vulnerable to

fluctuations.

Possible explanations for the age differences in variability in arousal in the present study

are that younger adults may have more variability in their daily routines, for example, engaging

in school or work activities that are potentially arousing or anxiety-provoking. The variability

results from the section on social rhythmicity indicate that at the very least, younger adults are

more variable in the timing of their daily activities. The one variable that younger and older

adults did not differ on for variability was the Pre-sleep Arousal Somatic subscale. Older adults

were much more variable on this scale than on the other arousal measures. This scale includes

physical sensations of arousal including shortness of breath, cold feelings in the extremities, dry

mouth, and upset stomach. Older adults could be experiencing transient health-related

conditions with symptoms that overlap with some of the physical manifestations of anxiety













Have morning beverage

OQn
O n.
clock time
HH MM AMIPM
W Elm--


People involved (leave blank if you were alone)
just present
spouse/partner 0
children 0
other family members 0
other persons) 0
Have breakfast
O ys
O -
clock time
HH MM AMIPM


People involved (leave blank if you were alone)
just present
spouse/partner 0
children 0
other family members 0
other persons) 0
Go outside for the first time
Oes
0no
O -
clock time
HH MM AMIPM
Ol I --


actively involved
0
0
0
0
O
O
O
O


actively involved
O
0
0
0
O
O
O









parents, Mike and Angella Dautovich, the unfaltering encouragement of my sister Sonia

Dautovich, and the daily support of David Sams who cared enough to walk each step of this

journey with me.











100%

90%

80%

70%

60%

50%

40%

30%

20%

10%


PANAS positive


PANAS negative


Figure 5-4. Percentage of total variability in affect measures that can be attributed to within
person variability. Note: *** Differences are significant at the 0.001 level (2-tailed),
** differences are significant at the 0.01 level (2-tailed).
































88


* Old.ier adults
* Older adults


0%


grid pleasant









calculate a daily variable. It is possible that a daily value of social rhythmicity is less meaningful

than a summary variable. Conversely, it is possible that the lack of a day-to-day relationship

between social rhythmicity and sleep reflects the idea that the regularity or irregularity of

younger adults' activities is not as impactful on sleep on a day-to-day basis. For example, a

critical threshold may exist whereby irregularity in activities over a one-day period is not

dysregulating for sleep, but habitual irregularity disrupts sleep. Entrainment of circadian

rhythms by social cues may require prolonged/repeated exposure to the cue. Therefore, higher

levels of variability over an extended period (e.g., two weeks) may sufficiently disrupt an

individual's circadian rhythms whereas irregularity on a daily basis is not sufficient to disrupt

already entrained rhythms.

The lack of association between social rhythmicity and sleep for older adults is puzzling

given previous research showing greater regularity is associated with better sleep in older adults

(Brown et al., 1996; Zisberg, Gur-Yaish, & Shochat, 2010). Possible explanations for the lack of

findings are 1) the lower levels of regularity in the older sample compared to previous studies

and 2) omission of aspects of regularity relevant for regulating sleep. First, the social

rhythmicity score of this sample of older adults (4.10 out of 7) was lower than the previously

reported mean scores for older adults of 4.75 (Zisberg et al., 2010) and 4.40 (Monk et al., 1992).

Given age-related declines in endogenous circadian entrainment mechanisms, the lower levels of

regularity in the older adult sample may not have been sufficient to entrain the circadian system

and consequently see an association with sleep. Monk (2010) made an interesting

recommendation regarding the need to study zeitgeber enhancement. He suggests that it may not

be the regularity/irregularity or presence/absence of social cues that serve to entrain circadian

systems, rather the strength of the entraining abilities of the zeitgeber that is of importance









over the study period for older adults (see table 5-18). Although none of the SRM variables

significantly predicted sleep for older adults, higher overall levels of irregularity approached

significance for predicting NWAK (P = -0.45, t(46.68) = -1.90, p =.06).

For the prediction of sleep by light, there were significant predictors at levels 1 and 2 for

the younger adults (see table 5-19). At the within person level (level 1) time (P = -0.30,

t(472.46) = -2.35,p <.05) and AM light (3 = 0.63, t(476.38) = 2.04,p <.05) significantly

predicted SOL. Younger adults' SOL decreased over the 14 days of the study and on days when

younger adults were exposed to light later in the day, they took longer to fall asleep. AM light

also significantly predicted TWT (3 = 2.06, t(119.74) = 2.74, p <.01) indicating that days with

later exposure to light during the morning were associated with more total time spent awake that

night. Days with more light duration (3 = 0.14, t(140.03) = 2.06, p <.05) and less light intensity

(P = -0.05, t(413.80) = -2.67,p <.01) were associated with less TST for that night. Days with

longer exposure to light but with less light intensity were associated with more total time spent

sleeping. At the between person level, AM light (P = 16.58, t(53.73) = 2.99,p <.01), light

duration (P = 0.30, t(54.41) = 3.48,p <.01), and light intensity (P = -0.09, t(53.88) = -3.34,p

<.01) were associated with TST. Younger adults who overall received light exposure later in the

day, higher levels of duration overall, and lower levels of light intensity reported spending more

total sleeping overall.

For older adults, there were significant predictors at the within and between person levels

(see table 5-20). AM light (P = -1.70, t(547.99) = -2.20,p <.05) and light duration (P = -0.04,

t(580.29) = -2.21,p <.05) significantly predicted WASO and time (P = -0.03, t(53.18) = -2.51,p

<.05) significantly predicted NWAK. These results suggest that days with exposure to light later

in the morning and days with exposure to light for a longer duration was associated with less









to age-related changes in sleep per se, but rather due to a combination of factors associated with

aging (Vaz Fragoso & Gill, 2007; Vitiello, Moe, & Prinz, 2002). For example, a meta-analysis

of studies examining objectively-measured sleep (PSG or actigraphy) in healthy community-

dwelling individuals across the lifespan found that the majority of changes seen in sleep patterns

across the life-span occur between early adulthood and age 60, with minimal declines from age

60 to age 102 (Ohayon et al., 2004).

It has been argued that sleep problems in non-healthy older adults are the result of factors

associated with aging such as loss of physical function, reduction in health status, and higher

levels of primary sleep disorders (e.g., sleep-disordered breathing; Vaz Fragoso & Gill, 2007).

Consequently, sleep disruption in older adulthood that is likely multifactorial in origin (Vaz

Fragoso & Gill, 2007).

In summary, it appears that difficulty sleeping remains a problem for the majority of both

younger and older adults, but the origin of sleeping difficulties may differ by age groups.

Although there are are-related changes in the physiology underlying sleep, it is likely that sleep

disruption in older age is due to multiple factors associated with aging. Due to differences in

perhaps the origin and expression of sleeping difficulties across age groups, the proposed model

was examined separately for each age group.

Age in Relation to Mediators of Light, Arousal, and Affect

It was necessary to examine the possible mediating effects of light, arousal, and affect

within two different age groups as age could interact with the variables of interest in numerous

ways. For example, we know that older adults typically are more regular in their social rhythms

but experience changes in their sleep (due to a combination of factors). Consequently, while

increased regularity could serve as an adaptive function for older adults, we may not see better

sleep outcomes in the older adults compared to the younger adults due underlying age-related














clock time
HH MM AMIPM
E:1 1 I

People involved (leave blank if you were alone)
just present actively involved
spouselpartner 0 0
children 0
other family members 0 0
other persons) O 0

Have dinner
0 es
Sno

clock time
HH MM AMIPM
7:O I -- I

People involved (leave blank if you were alone)
just present actively involved
spouse/partner 0 0
children 0 0
other family members O 0
other person() 0 0

Physical exercise

Syes
O -

clock time
HH MM AMIPM


People involved (leave blank if you were alone)
just present actively involved
spouselpartner O 0
children O O
other family members 0 0
other persons) O 0















Have an evening snack or drink

O p
O no

clock time
HH MM AMIPM


People involved (leave blank if you were alone)
just present actively involved
spouselpartner O O
children O O
other family members O O
other persons) O O

Watch evening TV news program

Syes
On-

clock time
HH MM AMIPM


People involved (leave blank if you were alone)
just present actively involved
spouselprtner 0 0
children 0 0
other family members O O
other persons) 0 0

Watch another TV program

OQys
0Qo
O

clock time
HH MM AMIPM
OL I--













Vaz Fragoso, C. A., & Gill, T. M. (2007). Sleep complaints in community-living older persons:
A multifactorial geriatric syndrome. Journal of the American Geriatrics Society, 55(11),
1853-1866.

Vgontzas, A. N., Bixler, E. O., Lin, H. M., Prolo, P., Mastorakos, G., Vela-Bueno, A., et al.
(2001). Chronic insomnia is associated with nyctohemeral activation of the
hypothalamic-pituitary-adrenal axis: Clinical implications. Journal of Clinical
Endocrinology and Metabolism, 86(8), 3787-3794.

Vitaterna, M. H., Takahashi, J. S., & Turek, F. W. (2001). Overview of circadian rhythms.
Alcohol Research & Health, 25(2), 85.

Vitiello, M. V., Moe, K. E., & Prinz, P. N. (2002). Sleep complaints cosegregate with illness in
older adults Clinical research informed by and informing epidemiological studies of
sleep. Journal ofPsychosomatic Research, 53(1), 555-559.

Wallace-Guy, G. M., Kripke, D. F., Jean-Louis, G., Langer, R. D., Elliott, J. A., & Tuunainen, A.
(2002). Evening light exposure: Implications for sleep and depression. Journal of the
American Geriatrics Society, 50(4), 738-739.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures
of positive and negative affect: the PANAS scales. Journal of Personality and Social
Psychology, 54(6), 1063-1070.

Watts, A. G., Swanson, L. W., & Sanchezwatts, G. (1987). Efferent projections of the
Suprachiasmatic nucleus: Studies using retrograde transport of fluorescent dyes and
simultaneous peptide immunohistochemistry in the rat. Journal of Comparative
Neurology, 258(2), 204-229.

Weaver, D. R., & Reppert, S. M. (1989). Direct inutero perception of light by the mammalian
fetus. Developmental Brain Research, 47(1), 151-155.

West, B. (1985). An essay on the importance of being nonlinear. New York, NY: Springer-
Verlag.

Wever, R. A. (1989). Light effects on human circadian-rhythms-A review of recent Andechs
experiments. Journal ofBiological Rhythms, 4(2), 161-185.

Wicklow, A., & Espie, C. A. (2000). Intrusive thoughts and their relationship to actigraphic
measurement of sleep: towards a cognitive model of insomnia. Behaviour Research and
Therapy, 38(7), 679-693.
















Baj-selineMasures .60


Medication 3
Name of medication
Frequency
Time of Day
Purpose


Medication 4
Name of medication
Frequency
Time of Day
Purpose


Medication 5
Name of medication
Frequency
Time of Day
Purpose


Medication 6
Name of medication
Frequency
Time of Day
Purpose


Medication 7
Name of medication
Frequency
Time of Day
Purpose


Medication 8
Name of medication
Frequency
Time of Day
Purpose


I I
I I
I I
I I




I I
I I
I I
I I




I I
I I
I I
I I




I I
I I
I I
I I




I I
I I
I I
I I




I I
I I
I I
I I









morning by the participants. Previous research examining the relationship between light and

sleep has assessed the intensity, duration, and time of day of exposure to light. The LES

measures these variables by asking participants to record the intensity of the light exposure

(ranging from darkness/moonlight to daylight with sun shining, scored as 1-5), the duration of

exposure (participants will estimate their exposure in one-hour intervals), and the time of day of

exposure (ratings using a 24-hour clock). Consequently, the LES has face validity in terms of

the typical components of light commonly assessed. Four variables are generated from the LES:

AM light (the first hour of the day exposed to cloudy or bright light [> 1000 lux]), PM light (the

last hour of the day exposed to cloudy or bright light [> 1000 lux]), light duration (the number of

hours exposed to cloudy or bright light over a 24-hour period), and light intensity (the total

amount of light [ranging from 1-5] multiplied by the number of hours).

In order to determine the criterion validity of the LES, a sub-group of randomly selected

participants were supplied with the Actiwatch-L (Mini Mitter Co. Inc., 2001) in order to

provide an objective assessment of light for comparison. Power analyses estimates using

G*Power (Faul, Buchner, Erdfelder, & Albert-Georg, 2006) suggested that a proposed sample

size of 12 individuals will provide sufficient power (beta > .80) to detect a large effect (d = .80)

at statistical probability values of .05. Sample sizes of 27 and 156 participants will provide

sufficient power to detect medium (d = .50) and small (d = 0.20) effects at probability values of

.05. Consequently, in order to detect a medium effect, a sub-sample of 15 participants from each

age group was selected to provide validation of the LES using the Actiwatch-L. Correlations

were run to compare the scores received for the LES variables of light duration and light

intensity to the comparable Actiware variables of TALT and Total Exposure. There were no

significant correlations between the objective and subjective measures of light. Possible











100% -

90%

80%

70%

60%

50%

40%

30%

20%

10%


WASO TWT


TST NWAK SQR


Figure 5-5. Percentage of total variability in sleep variables that can be attributed to within
person variability. Note: *** Differences are significant at the 0.001 level (2-tailed),
** differences are significant at the 0.01 level (2-tailed), and differences are
significant at the .05 level (2-tailed).


*'I Oldr Tad ults

* Older adults


0% +









by illumination, necessarily, but rather in response to interactions with other members of the

species and to other social stimuli (Mistlberger & Skene, 2004). Evidence from animal studies

suggests that animals do tend to coordinate their behavior to synchronize with other members of

their social group (Mistlberger & Skene, 2004). Evidence for social zeitgebers can be found in

studies of both animals and humans.

Animal studies

Animal studies have demonstrated the existence of social zeitgebers for animals in utero,

postnatally, and in the timing of daily activities between and within species. It is suggested that

the circadian timing of behavior may be particularly crucial when animals are young and are

vulnerable to predation and adverse environmental conditions (Mistlberger & Skene, 2004).

Studies of mammals have shown fetal circadian rhythms to be entrained to the rhythms of

the mother. While research has shown that light can penetrate the womb (Weaver & Reppert,

1989), it was shown that the entrainment of the fetus to maternal circadian rhythms occurs

independent of environmental light-dark cycles. The SCN of the fetus has been found to be

synchronized to maternal timing both in utero and postnatally (Reppert, 1995). Additionally,

some rodent species entrain to the timing of the maternal daily activities to the point where the

clock genes in the SCN are affected (Ohta, Honma, Abe, & Honma, 2002). The circadian clock

genes that are involved in circadian rhythm oscillations were found to be entrained to the

mother's rhythms. This phenomenon has been observed up until the weaning of the rodent

(Shimoda, Hanada, Yamada, Takahashi, & Takahashi, 1986).

The social zeitgeber of meal times has been demonstrated in the feeding activity of rabbits

where circadian timing is entrained to the social zeitgeber of maternal behavior. Rabbit pups

were able to anticipate the arrival of the doe for nursing (Jilge, 1995). Due to an inability to

maintain warmth during the first days of life, rabbit pups tend to huddle in a group deep within









(Himmelfarb & Murrell, 1983). Additionally, the scale is a sensitive indicator of transitory

anxiety and was selected for its long-standing reliability and validity of state-anxiety in

Caucasian, African American, and Hispanic populations (Novy, Nelson, Goddwin, & Rowzee,

1993).

Social Rhythm Metric

The Social Rhythm Metric (SRM-17; Appendix 2) was completed each morning by the

participants. The SRM-17 is a scale designed to assess the extent to which an individual is

regular or irregular on a daily basis in terms of the timing of daily events (Monk et al., 2003).

The scale was originally developed by Monk and colleagues (1990). Events were selected for

inclusion in the original measure if they were involved in the expression of a biological or

emotional drive and consequently likely to be tied to circadian rhythms (Monk et al., 1990). As

a result, the scale contains information about the timing of meals, exercise, going outside,

napping, and bed and wake times. There are two spaces (Activity A and Activity B) where the

individual can write in an activity that is specific to them. Additionally, as interpersonal

interactions have been found to have strong entrainment potential (Klein & Wegmann, 1975), the

scale also assesses whether other individuals were present for each activity. Once a day, in the

morning, individuals complete the scale retrospectively for the previous day. Individuals record

the clock time that the activity was completed and then indicate if they were alone, and if they

were not alone, who else was present.

Based on the daily rhythm scores, a total score can be calculated. First, a habitual time is

calculated for each activity using an outlier elimination algorithm Next, the number of times in



1 Since an event could occur at 9:00, 9:10, 9:30 and 12:02, to calculate a mean habitual time would not be an
accurate representation of the actual time of the event. Consequently, Monk et al., (1990), developed an algorithm
for eliminating outliers that could unduly influence the calculation of a habitual time. First the clock time is
decimalized (e.g. 9:45 becomes 9.75). Then 12 hours are added to all p.m. times and 24 hours are added to all times









CHAPTER 5
DISCUSSION

Review of the Findings

While prior research has shown light to be the dominant zeitgeber influencing circadian

rhythms, recent research has examined the influence of social activities. The regularity or

irregularity of social rhythms has been linked to sleep in both younger and older adults. Despite

the interest in the association between social rhythms and sleep, few studies have investigated

the role of specific mechanisms in mediating this relationship. This study expanded on prior

research by assessing the role of light, arousal, and affect in mediating the relationship between

social rhythmicity and sleep. Given the circadian, or 24-hour rhythm of sleep patterns, this study

investigated sleep on a daily basis. In order to justify examining the relationship between the

predictor variables and sleep on a daily basis within individuals, the constructs had to possess

sufficient within person variability. Once the amount of within person variability was

established, the relationship between variability and sleep outcomes was analyzed. Next, the

daily coupling of the predictor variables and sleep was evaluated. Finally, the role of light,

arousal, and affect as potential mediators of the relationship between social rhythmicity and sleep

was investigated. Each of these goals was examined in both younger and older adults. The

results of each specific aim will be reviewed and interpreted separately, followed by

consideration of the strengths and major limitations of the study, discussion of the implications

of the study (both theoretical and applied), and discussion of future directions.

Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep
Fluctuate Daily both Within and Between Persons

Within person variability or intraindividual variability has become increasingly of interest

in the social sciences. In fact, a recent issue of Psychology andAging devoted a special section

to the study of Intraindividual Variability and Aging. Intraindividual variability warrants









of the proposed mediators investigated in the present study (arousal and affect) and the outcome

variable of sleep. No studies, to our knowledge, have examined the relationship between social

rhythmicity and light.

Regularity and arousal

Few studies have examined the association between regularity of social rhythms and

arousal. One interesting investigation by Stetler, Dickerson, and Miller (2004) examined the

relationship between social zeitgebers and cortisol levels in healthy controls and depressed

individuals. The authors found that for the control group, the number of activities an individual

engaged in predicted their cortisol levels with more activities associated with a decline in cortisol

across the day. A similar relationship was not observed for depressed individuals. In terms of

the regularity of daily activities, among the control group, more regularity in social activities was

associated with a greater decline in cortisol throughout the day. A similar trend was observed in

depressed individuals but was not found to be significant. Interestingly, when the authors

combined total activities, regularity of activities, and the presence of another individual during

the activity in the model, only regular activities conducted when in the presence of another

individual were found to be associated with a decline in cortisol activity throughout the day.

When all three variables were entered into the model for the depressed individuals, there were no

significant associations between the variables and cortisol levels. The authors concluded that

individuals who are clinically depressed may not be as readily influenced by social activities

(i.e., social activities may not be interpreted the same way or have as great an impact as for

healthy controls; (Stetler, Dickerson, & Miller, 2004).

Regularity and affect Depression

By examining the association between social rhythmicity and affective disorders, it is

possible to develop an understanding of how rhythmicity is related to affect. Research has linked









social rhythms to depression, mania, and anxiety. In terms of depression, individuals diagnosed

with depression have been found to exhibit irregular biological rhythms (e.g., irregular sleep-

wake cycles, melatonin, temperature, and cortisol rhythms; Howland & Thase, 1999).

Additionally, circadian rhythms of those diagnosed with depression were found to be different in

timing and amplitude than those not diagnosed with depression (Kupfer & Foster, 1972).

In addition to irregular biological rhythms, research indicates that individuals diagnosed

with depression may also experience irregular social rhythms. Brown and colleagues (1996)

examined the stability of social rhythms following spousal bereavement among elderly

participants. Individuals who were experiencing a major depression episode exhibited

significantly lower levels of social rhythm regularity compared to non-depressed controls

(Brown et al., 1996). Similar findings emerged from the Prigerson and colleagues (1994) study

examining the relationship between stressful life events, social rhythms, and levels of depressive

symptomatology among elderly participants (Prigerson et al., 1994).

Contrasting with findings illustrating a significant relationship between social rhythms and

depression, Monk and colleagues (1990) found that there were no differences in social

rhythmicity scores between individuals with unipolar depression and normal controls. A possible

explanation for these findings, however, is that the study of social rhythms among patient

populations may be limited by unstandardized medication regimes of those diagnosed with mood

disorders (Monk et al., 1990). Considering that medications have been shown to affect circadian

rhythms, it is possible that the relationship between social rhythms and depressive episodes may

be moderated by psychotropic medication use (Bendetti et al., 2001).

Regularity and affect Manic episodes

Research has also examined the relationship between social rhythmicity and bipolar

disorder. Twice as many individuals who developed a manic episode experienced a social









The social zeitgeber theory proposes that changes in mood result from disrupted social

rhythms caused by life events (see Figure 2-2). Life events refer to a stressful occurrence that

takes place close in time to the onset of the affective disorder (Grandin et al., 2006). Life events,

in turn, can affect social zeitgebers. For instance, the individual can decide to stay home and not

leave the house. Changes in social zeitgebers, in turn, can result in a disruption of social

rhythms. A disruption in social rhythms could impact biological rhythms (e.g., sleep-wake

cycle, hormonal levels) which in turn could be manifested in somatic symptoms (e.g., decreased

appetite, poor sleep). Finally, affective episodes are thought to result from the disruption of

biological rhythms, resulting in somatic symptoms.

The internal trigger hypothesis (see Figure 2-2) expands on the social zeitgeber theory by

proposing that social and biological rhythm disruptions in individuals are influenced by an

underlying abnormality in the circadian pacemaker brought about through a physiological

process such as a genetic mutation (Grandin et al., 2006). The authors propose that irregularity

in the rhythms of the individual (either social or biological rhythms) is a result of a genetic

mechanism, perhaps resulting from a genetic mutation. For example, an individual who is

experiencing manic symptoms may have an underlying genetic mutation affecting the regularity

of his/her circadian pacemaker which is manifested in irregular mood. The internal trigger

hypothesis is an improvement over the social zeitgeber theory in that it proposes a mechanism

underlying disruptions in social rhythms and biological rhythms. Unfortunately, there are two

limitations of this model. First, the model does not propose a mechanism by which social

rhythms influence biological rhythms. The authors suggest that abnormality in the circadian

pacemaker concurrently influences both social and biological rhythms. While the authors













Ram, N., Lindenberger, U., & Blanchard-Fields, F. (2009). Introduction to the special section on
intraindividual variability and aging. Psychology andAging. 24(4), 775-777.

Ready, R. E., Marquez, D. X., & Akerstedt, A. (2009). Emotion in Younger and Older Adults:
Retrospective and Prospective Associations with Sleep and Physical Activity. [Article].
Experimental Aging Research, 35(3), 348-368.

Reppert, S. M. (1995). Interaction between the circadian clocks of mother and fetus Circadian
Clocks and Their Adjustment (Vol. 183, pp. 198-207).

Rocke, C., Li, S., & Smith, J. (2009). Intraindividual variability in positive and negative affect
over 45 days: Do older adults fluctuate less than young adults? Psychology andAging.
24(4), 863-878.

Rodenbeck, A., Huether, G., Ruther, E., & Hajak, G. (2002). Interactions between evening and
nocturnal cortisol secretion and sleep parameters in patients with severe chronic primary
insomnia. Neuroscience Letters, 324(2), 159-163.

Russell, J. A., & Bullock, M. (1985). Multidimensional-Scaling of Emotional Facial Expressions
Similarity from Preschoolers to Adults. Journal ofPersonality and Social Psychology,
48(5), 1290-1298.

Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect Grid: A single-item scale of
pleasure and arousal. Journal ofPersonality and Social Psychology, 57(3), 493-502.

Scheuermaier, K., Laffan, A. M., & Duffy, J. F. (2010). Light Exposure Patterns in Healthy
Older and Young Adults. Journal of Biological Rhythms, 25(2), 113-122.

Shaffery, J. P., Sinton, C. M., Bissette, G., Roffwarg, H. P., & Marks, G. A. (2002). Rapid eye
movement sleep deprivation modifies expression of long-term potentiation in visual
cortex of immature rats. Neuroscience, 110(3), 431-443.

Shear, M. K., Randall, J., Monk, T. H., Ritenour, A., Tu, X., Frank, E., et al. (1994). Social
rhythm in anxiety disorder patients. Anxiety, 1(2), 90-95.

Shimoda, K., Hanada, K., Yamada, N., Takahashi, K., & Takahashi, S. (1986). Periodic exposure
to mother is potent zeitgeber of rat pups' rhythm. Physiology and Behavior, 36, 7.

Singer, J. D., Davidson, S. M., Graham, S., & Davidson, H. S. (1998). Physician retention in
community and migrant health centers Who stays and for how long? Medical Care,
36(8), 1198-1213.









exhibited less variability for older adults compared to younger adults suggesting a stabilizing

rather than destabilizing function of age.

The higher levels of variability for sleep compared to the other constructs for older adults

fits with previous findings indicating greater variability in cognitive, sensorimotor, and

physiological functioning. It is possible that sleep represents a more physiological process

(albeit one that is affected by psychological processes) than social rhythmicity, light exposure,

affect, or arousal which are more behavioral or psychological in nature. Consequently, it fits that

more variability would be seen for sleep given that sleep also has a strong physiological

component.

Aim 2: Identify How Variability in Social Rhythmicity, Light, Arousal, and Affect is
Related to Sleep

While aim 1 of the study established the presence of within person variability, aim 2

investigated the positive and negative associations of variability. For example, do individuals

who are more variability in their social rhythmicity, light exposure, arousal levels, and affect

experience better or worse sleep? In general, within person variability was typically associated

with negative outcomes for younger adults and was less related to positive or negative outcomes

for older adults.

In particular, greater variability in social rhythms was associated with worse sleep for

younger adults. This result replicates previous findings showing that greater irregularity in social

rhythms was associated with poor sleep for college students (Carney et al., 2006). In particular,

greater fluctuations in the timing of daily activities were associated with a longer time to fall

asleep and more time spent awake during the night. It is possible that varying the timing of daily

activities could 1) affect the timing of the sleep period by undermining the circadian rhythm

(hence longer time to fall asleep) and/or 2) increase levels of arousal/negative affect which in









The category of'social stimuli' refers to the sights, sounds, smells, and other stimuli that

an individual may encounter during the day. The area of 'social stimuli' was under investigation

in the present study. Daily patterns of behavior originally were based on availability of light.

We have since become increasingly detached from the regularizing power of light (e.g., since the

advent of electricity and technology). In place of the absolute regularizing powers of light, daily

life has increasingly been governed by exigencies of daily social life (Monk, Flaherty, Frank,

Hoskinson, & Kupfer, 1990). For example, individuals will stay active long after daylight ends

in order to engage in any number of social activities. A student may remain awake for a portion

of the night in order to finish an assignment or a parent may spend the night awake tending to a

newborn baby. Through the advancement of technology, we have increasingly turned to

artificial light in order to extend the time of day we can engage in activities. Although light is

typically present during these activities, it is proposed that we cannot deny the influence of the

activity itself in terms of influencing our biological rhythms. For example, a student who delays

the onset of their sleep period in order to finish an assignment may be motivated to stay awake

by both their desire to finish the assignment as well as by the influence of light. In fact, it could

be argued that light alone would not be sufficient to maintain the alertness of the student without

the accompanying motivation/arousal/anxiety to complete the assignment. Consequently, this

study aimed to investigate the association between social stimuli and biological rhythms while

still accounting for the influence of light.

Social stimuli can influence the circadian system via several pathways, including direct

pathways and indirect pathways. Indirect pathways include affecting the behavioral state of the

individual (e.g., exerting an autonomic affect on pupil size) which would affect the amount of

light influencing the SCN and impact the circadian system. Social stimuli could also directly












100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%


PM i'ht


' ht duration


Iight intensity


Figure 5-2. Percentage of total variability in light measures that can be attributed to within
person variability. Note: *** Differences are significant at the 0.001 level (2-tailed),
** differences are significant at the 0.01 level (2-tailed).


* iOLJudi-r adults

* Older adults


AM iht


I


I









turn results in more wake time during the night. Research on motivation has shown that more

regular, habitual behavior is associated with less emotional intensity (Wood et al., 2002).

Therefore, younger adults who are more irregular in their daily activities may experience greater

emotional intensity (arousal or affect) which in turn disrupts sleep. The lack of association of

variability in social rhythmicity with sleep for older adults was surprising. As discussed for aim

3 below, it may be the strength rather than the presence of the zeitgeber that is important for

entraining daily rhythms. Monk (2010) commented on the increasing difficulty of entraining

circadian rhythms with age. Therefore, some activities may be more salient for older adults in

terms of their ability to entrain daily rhythms. The calculation of social rhythms involves an

average of all completed daily activities. It may be that for older adults, certain activities have

stronger entrainment ability than others. The examination of variability in the timing of specific

activities (see below) provides support for this explanation given that irregularity in specific

activities was associated with sleep for older adults.

The association between variability in specific daily events and sleep was also examined.

Greater fluctuation in the time younger adults got out of bed was associated with worse sleep.

This result is consistent with models of insomnia that advocate for regular wake times in order to

entrain the circadian rhythm to in turn promote better sleep (Bootzin, 1972; Spielman, Saskin, &

Thorpy, 1987). Additionally, one of the few studies that examined the timing of daily events in

relation to sleep found that variability in the out of bed time was associated with worse sleep for

younger adults (Carney et al., 2006). Meal times were also predictive for younger adults with

varying breakfast and lunch times associated with worse sleep.

For older adults, variability in nap times, time returning home, and time for the evening

snack was associated with worse sleep. Sleep hygiene treatment recommendations often









Table 5-12. Correlations among light variables between persons (below diagonal) and within person (above diagonal) for older adults
Amlight Pmlight light duration light intensity


Amlight 0.05 -0.55*** -0.45***
Pmlight -0.12 0.53*** 0.31***
light duration -0.74*** 0.49** 0.67***
light intensity -0.64*** 0.56** 0.79***-
Note: *Correlation is significant at thep < 0.05 (2-tailed); **correlation is significant at thep < 0.01 (2-tailed); ***correlation is
significant at the p < 0.001 (2-tailed).


Table 5-13. Correlations among arousal variables between persons (below diagonal) and within person (above diagonal) for younger
adults
PSAS UMACL-R UMACL-R
arousal grid PSAS somatic cognitive energetic tense
arousal grid 0.02 0.16*** -0.57*** 0.17***
PSAS 0.20
PSAS 00 0.29*** -0.13** 0.22***
somatic
PSAS 0.27
0.52** -0.13** 0.25***
cognitive
UMACL-R -0.82***
Uener c -0.26 -0.39* -0.36***
energetic
UMACL-R
UMACL-R 0.60** 0.39* 0.52* -0.72***
tense
Note: *Correlation is significant at thep < 0.05 (2-tailed); **correlation is significant at thep < 0.01 (2-tailed); ***correlation is
significant at the p < 0.001 (2-tailed).









Procedure

Participants were recruited from the North Central Florida area and online. Recruitment

materials consisted of flyers posted around the University of Florida and the community of

Gainesville, FL, online advertisements in community classified such as craigslist,

undergraduates were recruited from undergraduate psychology courses at the University of

Florida and from the undergraduate psychology research pool, and older adults were recruited

using the University of Florida's Age Network Participant Registry. For the younger adults, 2%

of the participants were recruited by word of mouth (e.g., from a friend participating in the

study), 4% were recruited using an online description of the study, 12% were recruited using

craigslist, 38% were recruited from advanced psychology classes, and 44% were recruited from

the undergraduate psychology research pool. For older adults, 6.1% were recruited using online

descriptions of the study, 24.5% were recruited using the Age Network Participant Registry, and

34.6% were recruited both by word of mouth and from craigslist.

The study materials were completed online. After a brief introduction to the study,

individuals completed the electronic informed consent approved by the University of Florida

Institutional Review Board. Once consent was given, participants were screened for the

exclusionary criteria consisting of 1) currently working shift work; 2) unable to complete study

materials using a computer; 3) visual deficits that preclude participation (e.g., have severe self-

reported difficulty reading the newspaper); 4) are currently on vacation or plan to take a vacation

during the next two weeks; 5) have been diagnosed with a dementia disorder; 6) sleep disorders

other than insomnia (e.g., periodic limb movement disorder, sleep apnea); 7) are currently

pregnant. If participants met the criteria for participating in the study, they completed the

baseline surveys online (i.e., the demographic and health questionnaire, the BDI-II, and the

STAI-Y1). Participants were given the option to wear the Actiwatch device if they were within









CHAPTER 4
RESULTS

The overall aim of the proposed study was to investigate the role of light, arousal, and

affect in mediating the relationship between social rhythmicity and sleep. The overall aim of the

study was examined by investigating a number of specific aims (aim 1 through aim 4). Each of

the following three aims was analyzed separately for the younger and older adult samples. See

Table 5-1 for a summary and description of the variables used in the analyses.

Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep
Fluctuate Daily both Within and Between Persons

The purpose of aim one was to establish the extent to which individuals vary both within

and between persons in their social rhythmicity, light, arousal, affect, and sleep.

Social Rhythmicity

The intraclass correlation coefficient (ICC) for the social rhythmicity variable for

younger adults was 0.23 indicating that 77% of the variability in social rhythmicity is a within

person phenomenon (see Figure 5-1). For older adults, the ICC was 0.28 indicating that 72% of

the variability in social rhythmicity occurs within persons. Younger adults were significantly

more variable within persons in SRM compared to older adults t(98) = 6.42, p <.001. The mean

social rhythmicity value (average variability during the study) was 2.36 (SD = 0.81) for younger

adults and 4.10 (SD = 0.82) for older adults. The difference in mean variability was significant

t(98)= -10.71,p <.001 (see table 5-2).

Light Variables

The ICC for the light variables ranged from 0.25 to 0.47 for the younger adults and 0.43 to

0.52 for the older adults (see Figure 5-2). For younger and older adults respectively, the amount

of within person variability in the AM light scale was 74% and 55%, for the PM light scale 74%

and 57%, for the light duration scale 57% and 48%, and for the light intensity scale 53% and













Malkoff-Schwartz, S., Frank, E., Anderson, B. P., Hlastala, S. A., Luther, J. F., Sherrill, J. T., et
al. (2000). Social rhythm disruption and stressful life events in the onset of bipolar and
unipolar episodes. Psychological Medicine, 30(5), 1005-1016.

Mandler, G. (1975). Mind and emotion. New York: Wiley.

McCarley, R. W., & Massaquoi, S. G. (1992). Neurobiological structure of the revised limit
cycle reciprocal interaction model of REM cycle control. The Journal of Sleep Research,
1, 132-137.

McClintock, M. K. (1971). Menstrual synchrony and suppression. Nature, 229(5282), 244.

McCrae, C. S., McNamara, J. P. H., Rowe, M. A., Dzierzewski, J. M., Dirk, J., Marsiske, M., et
al. (2008). Sleep and affect in older adults: Using multilevel modeling to examine daily
associations. Journal of Sleep Research, 17(1), 42-53.

McCrae, C. S., Rowe, M.A., Dautovich, N.D., Lichstein, K.L., Durrence, H.H., Riedel,, & B.W.,
T., D.J., & Bush, A.J. (2006). Sleep hygiene practices in two community dwelling
samples of older adults. Sleep, 29(12), 9.

Mehrabian, A., & Russell, J. A. (1974). An approach to environmental psychology. Cambridge,
MA: MIT Press.

Mini Mitter Co. Inc. (2001). Actiwatch 16/Actiwatch 64/Actiwatch-L/Actiwatch-Score
Instruction Manual: Mini Mitter Co., Inc.

Mish, F. C. E. (2000). In F. C. Mish (Ed.), Merriam-Webster's Collegiate Dictionary (10 ed.).
Springfield, Massachusetts: Merriam-Webster, Incorporated.

Mishima, K., Okawa, M., Shimizu, T., & Hishikawa, Y. (2001). Diminished melatonin secretion
in the elderly caused by insufficient environmental illumination. Journal of Clinical
Endocrinology & Metabolism, 86(1), 129-134.

Mistlberger, R. E., Antle, M. C., Glass, J. D., & Miller, J. D. (2000). Behavioral and
serotongergic regulation of circadian rhythms. Biological Rhythm Research, 31, 43.

Mistlberger, R. E., & Skene, D. J. (2004). Social influences on mammalian circadian rhythms:
animal and human studies. Biological Reviews, 79(3), 533-556.

Monk, T. H. (2005). Aging Human Circadian Rhythms: Conventional Wisdom May Not Always
Be Right. Journal of Biological Rhythms, 20(4), 366-374.

Monk, T. H. (2010). Enhancing Circadian Zeitgebers. [Editorial Material]. Sleep, 33(4), 421-422.
































Figure 6-2. Relationships in the mediated model of social rhythmicity that were observed in
older adults. The dashed lines indicate associations that were not observed. The solid
lines indicate associations that were observed. The solid lines indicate the presence
of mean-level relationships. The fluctuating black solid lines indicate the presence of
day-to-day relationships.









essential to examine the role of light as mediating the relationship between social rhythmicity

and sleep. Based on the possible differential effects of age on the relationships of interest, it was

warranted to examine whether the mediation of the social rhythmicity and sleep relationship is

similar for both younger and older adults.









Due to the hierarchical nature of the data (14 consecutive days nested within 50

participants), it was possible to examine to what extent the predictor variables were related to

sleep at both level 1 and level 2. Level 1 analyses examined the day-to-day relationship between

the variables for each individual. For example, "On days when an individual is more or less

regular in their social activities, do they have more or less SOL." (see Figure 3-1 for example).

Level 1 analyses were computed using person-centered variables which are the intraindividual

fluctuations around the individual's own mean. Level 2 analyses examined the overall

relationship between mean-level variables. For example, "Are higher or lower levels of overall

regularity associated with more or less SOL overall." In order to compute level 2 analyses,

mean-levels were calculated for the variables of interest.

Fixed and random effects were also estimated in addition to examining level 1 and level 2

relationships. Fixed effects are the effects that hold for all persons. Fixed effects can be

measured at both level 1 and level 2. For example, fixed effects at level 1 refer to the

relationship between the predictors and sleep, on a day-to-day level, that hold for all persons, on

all occasions. Fixed effects at level 2 refer to the overall, mean-level relationships, between the

predictors and sleep for all persons on all occasions. Random effects, on the other hand, test

whether there are significant individual differences in the observed fixed effects. A random

effect of the fixed effect of social rhythmicity predicting sleep at level 1 would indicate that

although social rhythmicity predicts sleep on a day-to-day basis, there is significant within-

person variation in the magnitude of this relationship.

Hypothesis

The hypothesis for the level 1 analysis was exploratory in that few studies have examined

variability in the predictor variables within the individual and no studies (to our knowledge) have

examined the daily relationships between the predictor variables and sleep. Consequently, the









recommend restricting nap behavior and, if it is necessary to nap, selecting a consistent nap

schedule. Napping can meet some of an individual's sleep need and consequently, disrupt

nighttime sleep. By napping at inconsistent times, older adults may be preventing their daily

rhythms from adjusting to the disruption in sleep drive caused by the nap. Varying the return to

home time could have an arousing effect by disrupting the daily schedule and shortening the time

to wind down before bed. Fluctuations in the timing of evening snacks could negatively affect

sleep by causing the digestive system to metabolize an unaccustomed late snack or by causing an

alerting effect by engaging in a social interaction during the snack.

Variability in the timing of lunch and work was associated with better sleep for older

adults. Both these activities occur towards the beginning to middle of the day. Possibly varying

the timing of these events has a stimulating or arousing effect that activates older adults (thereby

increasing activity levels during the day and interfering with napping behavior). Given the

timing of these events, older adults would have enough time to 'wind down' during the day and

the extra stimulation/activation during the day could increase feelings of tiredness at night (hence

less time to fall asleep).

Interestingly, other than the timing of lunch, differing daily activities were associated with

sleep for younger and older adults. The age discrepancies in results suggest that there may be

age-related differences in the ability of different daily events to entrain rhythms in younger and

older adults. It may be that older and younger adults differ in the variability of timing of

different events. For example, meals may occur at more regular times for older adults (perhaps

due to the timing of medication or medical conditions requiring regular meals) while work may

occur at more regular times for younger adults. The differing levels of variability for discrete

events may affect the relationship between variability in these events and sleep. Additionally,









CHAPTER 3
METHODS

The present study examined the relationships between social rhythmicity, light, arousal,

affect, and sleep in older and younger adults. Study materials were completed via the internet.

Prior to beginning participation in the study, participants completed the informed consent,

demographics and health questionnaire, the BDI-II, and the STAI-Y1. After completing the pre-

study assessment, participants completed the social rhythmicity, light, arousal, affect, and sleep

questionnaires on a daily basis for fourteen days.

Materials

Demographics and Health Survey

This questionnaire (see Appendix 1), collects information on demographics, sleep disorder

symptoms, physical health, and mental health (Lichstein, Durrence, Taylor, Bush, & Riedel,

2003). The Demographics and Health Survey was administered prior to the start of the 14 days

of data collection. Health conditions that are assessed include: heart attack, cancer, AIDS,

hypertension, neurological disorder (e.g., Parkinson's, seizures), breathing disorder (asthma,

emphysema), urinary problems (kidney disease, prostate problems), diabetes, pain (arthritis, back

pain, migraines), and gastrointestinal disorders (stomach, irritable bowels, ulcers, and gastric

reflux).

Self-report sleep questions on the survey ask whether the participant had a sleep problem

and if they or a bed partner noticed heavy snoring, difficulty breathing or gasping for breath,

frequent leg jerks, restlessness before sleep onset, sleep attacks during the day, or paralysis at

sleep onset. If they answer yes to any of these problems they are asked to describe the problem

and indicate how often and how long the symptoms have occurred. Additionally, the participants

are asked for the names of all medications they are currently taking and about any mental health









Table 5-6. Correlations of within person variability in affect and mean-level sleep in younger adults. (N=50)
1 2 3 4 5 6 7 8 9
1. pleasant grid 1
2. PANASpositive 0.52*** 1
3. PANASnegative 0.44** 0.36** 1
4. SOL 0.06 0.06 0.17 1
5. WASO 0.17 0.25 -0.04 0.32* 1
6. TWT 0.18 0.13 0.12 0.75*** 0.57*** 1
7. TST -0.14 -0.07 -0.17 -0.26 -0.06 -0.34* 1
8. NWAK 0.15 0.29* 0.03 0.27 0.44** 0.30* -0.13 1
9. SQR -0.31* -0.23 -0.38** -0.13 -0.20 -0.23 0.43** -0.17 1
Note: *** Correlation is significant at the 0.001 level (2-tailed), ** correlation is significant at the 0.01 level (2-tailed), and *
Correlation is significant at the .05 level (2-tailed).
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating.

Table 5-7. Correlations of within person variability in social rhythmicity and mean-level sleep in older adults. (N=50)


1 2 3 4 5 6 7


1. SRM 1
2. SOL 0.11 1
3. WASO 0.22 0.17 1
4. TWT 0.26 0.71*** 0.69*** 1
5. TST -0.11 -0.28* -0.07 -0.31* 1
6. NWAK -0.02 0.04 0.53*** 0.33* -0.04 1
7. SQR -0.17 -0.29* -0.57** -0.51*** 0.11 -0.50** 1
Note: *** Correlation is significant at the 0.001 level (2-tailed), ** correlation is significant at the 0.01 level (2-tailed), and *
Correlation is significant at the .05 level (2-tailed).
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating.















Bsu-s 60


Medication 9
Name of medication
Frequency
Time of Day
Purpose


Medication 10
Name of medication
Frequency
Time of Day
Purpose


Medication 11
Name of medication
Frequency
Time of Day
Purpose


Medication 12
Name of medication
Frequency
Time of Day
Purpose


Medication 13
Name of medication
Frequency
Time of Day
Purpose


Have you experienced a stressful or disruptive life event (e.g. move, birth, death, illness,
marriage, divorce) over the past (select all that apply)

1 year

| month

Sdaweek

E day


j
j I
j
j




j
1 (
j
j




1 (
1 (
j
j




j
j
j
1 (




j
j
j
j I










Table 5-23. Multilevel model with affect variables predicting sleep in younger adults
Fixed Effects
SOL WASO TWT TST NWAK SQR
Predictor Variable B (SE) t (d) B (SE) t (dJ) B (SE) t (d) B (SE) t (d) B (SE) t (d) B (SE) t (d)
Within person
time -0.25 -2.26 -0.02 -0.12 -0.22 -0.82 -1.67 -1.63 -0.01 -0.67 -0.003 -0.40
(0.11) (93.38)* (0.13) (614.36) (0.27) (617.79 (1.02) (610.63) (0.01) (112.85) (0.008) (134.46)
pleasant gridcentered -0.58 -1.67 -0.58 -1.15 -1.37 -1.70 2.53 0.64 -0.04 -1.07 0.13 4.84
(0.35) (44.16) (0.50) (45.60) (0.81) (604.22) (3.93) (49.71) (0.04) (33.60) (0.03) (53.26)***
PANAS positivecentered -0.16 -1.76 0.02 0.18 -0.24 -1.10 -0.31 -0.40 -0.01 -1.19 0.02 2.62
(0.09) (54.74) (0.11) (46.33) (0.22) (60.13) (0.76) (42.45) (0.01) (44.29) (0.01) (54.13)*
PANAS negativecentered 0.15 0.80 0.66 2.32 0.87 2.07 -2.47 -1.68 0.03 1.25 -0.01 -0.94
(0.19) (44.69) (0.28) (32.19) (0.42) (22.59)* (1.47) (610.66) (0.02) (45.26) (0.01) (41.53)

Between person
pleasant gridme. -1.42 -0.95 -0.47 -0.46 -4.82 -1.68 28.14 3.61 0.03 0.31 0.14 2.15
(1.49) (48.87) (1.01) (50.75) (2.86) (49.84) (7.79) (49.58)** (0.10) (50.52) (0.07) (49.75)*
PANAS positivemea 0.13 0.65 0.18 1.31 0.65 1.70 -1.29 -1.25 0.004 0.28 0.003 0.34
(0.20) (48.17) (0.13) (49.58) (0.38) (49.04) (1.04) (48.32) (0.01) (49.99) (0.01) (48.94)
PANAS negativemen 0.24 0.47 0.07 0.19 -0.08 -0.08 -1.92 -0.70 0.04 0.99 -0.05 -2.22
(0.52) (49.68) (0.36) (52.29) (1.01) (50.75) (2.75) (51.19) (0.04) (50.89) (0.02) (50.79)*

Random Effects
SOL WASO TWT TST NWAK SQR
B (SE) Z B(SE) Z B(SE) Z B (SE) Z B (SE) Z B (SE) Z
Within person
time 0.01 0.15 0.00 0.00 0.00 0.00 0.00 0.00 0.001 1.06 0.003 0.93
(0.08) (0.001) (0.003)
pleasant gridcentered 0.52 0.56 3.95 1.78 0.00 0.00 224.17 1.74 0.01 1.04 0.01 1.73
(0.93) (2.23) (128.72) (0.01) (0.01)
PANAS positivemea 0.08 1.38 0.10 1.09 0.42 1.27 1.78 0.39 0.001 1.21 0.0003 1.29
(0.06) (0.09) (0.33) (4.54) (0.001) (0.0003)
PANAS negativemea 0.35 1.42 1.69 2.09* 1.01 0.61 0.00 0.00 0.01 2.30 0.0002 0.28
(0.25) (0.81) (1.65) (0.004) (0.0007)
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating. *Correlation is significant at thep < 0.05 (2-tailed), **correlation is significant at thep <
0.01 (2-tailed),***correlation is significant at thep < 0.001 (2-tailed), Variance is too small to be estimated; the final Hessian matrix
was not positive definite, although all convergence criteria were satisfied.































Figure 6-1. Relationships in the mediated model of social rhythmicity that were observed in
younger adults. The dashed lines indicate associations that were not observed. The
solid lines indicate associations that were observed. The solid lines indicate the
presence of mean-level relationships. The fluctuating black solid lines indicate the
presence of day-to-day relationships.













What time do you usually go to bed?
HH MM AMIPM
time E r7 I

What time do you usually get up in the morning?
HH MM AMIPM
time |! I

If you nap during the day is it because: (check all that apply)
- you fall asleep accidentally
D you decideto have a nap

Do you think a nap during the day (check all that apply)
D helps you sleep better at night
] makes your sleep at night worse
D does not affect your sleep
D helps you to feel less tired dunng the day
] helps you to perform better during the day









Table 5-10. Correlations of within person variability in affect and mean-level sleep in older adults. (N=50)
1 2 3 4 5 6 7 8 9
1. pleasant grid 1.00
2. PANASpositive 0.36* 1.00
3. PANASnegative 0.07 0.45** 1.00
4. SOL 0.13 0.03 -0.01 1.00
5. WASO 0.06 -0.12 -0.08 0.17 1.00
6. TWT 0.15 0.00 -0.07 0.71*** 0.69*** 1.00
7. TST -0.05 -0.03 -0.04 -0.28* -0.07 -0.31* 1.00
8. NWAK -0.08 -0.03 -0.04 0.04 0.53*** 0.33* -0.04 1.00
9. SQR -0.14 0.05 -0.16 -0.29* -0.57** -0.51** 0.11 -0.50** 1.00
Note: *** Correlation is significant at the 0.001 level (2-tailed), ** correlation is significant at the 0.01 level (2-tailed), and *
Correlation is significant at the .05 level (2-tailed).
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating.

Table 5-11. Correlations among light variables between persons (below diagonal) and within person (above diagonal) for younger
adults
Amlight Pmlight light duration light intensity
Amlight -0.50*** -0.40*** 0.20***
Pmlight -0.57** 0.69*** 0.50***
light duration -0.54** 0.67*** 0.32***
light intensity 0.06 0.57** 0.31
Note: *Correlation is significant at thep < 0.05 (2-tailed); **correlation is significant at thep < 0.01 (2-tailed); ***correlation is
significant at the p < 0.001 (2-tailed).





















Please respond to the following items. All information will remain confidential.
VWhat is the county that you live in (e.g., Gainesville, FL is located in Alachua
County)'
Mat is the state/province that you live in? I
hat is your sex (e g female)?
How would you identify your cultural background? (e.g.,El Salvadorian, white, black,
Chinese Amencan...)
VWat is your age? I
hat is your height?
Mat is your weight?
What is your primary language?
What is your highest level of education?
If you are in a relationship, what is your spouselpartners highest level of education


What is your date of birth
MM DD YYYY
date I IF1 II


Please respond yes or no


Are you presently
employed?
Does anyone live with you?


What is your marital status?


Q Marned

i] Single
D Common-law

D Widowed

D Divorced

] Separated









Age and Sleep

Difficulty sleeping is a problem faced by the majority of Americans across different age

groups. The National Sleep Foundation's 2008 Sleep in America poll found that 65% of working

Americans, age 18 and above, report experiencing a sleep problem at least a few nights a week

(National Sleep Foundation, 2008). Similarly, an estimated 57% of adults age 60 and older

complain of difficulty sleeping (Kryger, Monjan, Bliwise, & Ancoli-Israel, 2004). There are

significant negative outcomes associated with difficulty sleeping for older adults including

excessive daytime sleepiness, attention and memory problems, depressed mood, falls, overuse of

hypnotic drugs, and lower quality of life (Kryger et al., 2004). One type of sleeping disorder,

insomnia, is characterized by difficulty falling asleep, staying asleep, waking early, and

experiencing nonrestorative sleep (American Psychiatric Association, 2000).

Sleep undergoes various physiological changes across the lifespan. The most prominent

changes are reductions in slow wave sleep (SWS) and Rapid Eye Movement (REM) sleep. Slow

wave sleep consists of stage three and four non-rapid eye movement (non-REM) sleep. SWS is

described as restorative sleep in that the individual experiences a slowing of brain activity;

enabling repairs and rest for brain areas (Home, 1988). REM sleep involves low-voltage brain

waves and dreaming in addition to the movement of the eye muscles and involuntary muscle

jerks. REM sleep is important for functioning in that it helps to develop synaptic connections

during early life and is needed to maintain these connections in late life (Shaffery, Sinton,

Bissette, Roffwarg, & Marks, 2002). Additionally, older adults can experience a phase advance

in their sleep wake rhythm whereupon they become sleepier earlier in the day and have earlier

wake times. Despite underlying physiological changes, however, recent research has suggested

that age-related differences in sleep are not as drastic as previously believed (Ohayon,

Carskadon, Guilleminault, & Vitiello, 2004). Sleeping problems in older adults may not be due













If so, please briefly describe this event


Has someone you know experienced a stressful or disruptive life event (e.g. move, birth,
death, illness, marriage, divorce) over the past (select all that apply)
D year
| month


R day

If so, please briefly describe this event


Do you frequently feel nervous or depressed?
Q yes
0-"
O no

If yes, explain:
II

If yes, does this interfere with your sleep?
0Qyes
O P
O.-

If yes, explain:
I I

Please list the number of mental disorders you have and the number of years you have
had the disorders)
I I

Does this problems) interfere with your sleep?
0 yes
0no
O P
O









THE MEDIATION OF THE RELATIONSHIP BETWEEN SOCIAL RHYTHMICTY AND
SLEEP BY LIGHT, AROUSAL, AND AFFECT IN
BOTH YOUNGER AND OLDER ADULTS




















By

NATALIE DEIDRE DAUTOVICH


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2010









Louis et al., 2000; Kawinska et al., 2005; Mishima, Okawa, Shimizu, & Hishikawa, 2001) which

found older adults had less or equal levels of light exposure compared to younger adults.

Conversely, a minority of studies have shown older adults to have higher levels of light exposure

compared to younger adults (Campbell, Kripke, Gillin, & Hrubovcak, 1988; Scheuermaier et al.,

2010). It is possible that the use of a healthy, community-dwelling sample of older adults

resulted in a sample that was more active than previous studies with less healthy, residential

older adults. Although, the 'overall' age-related differences in the light variables are important

for understanding the relationship between age and light exposure they do not tell the complete

story. For example, the within person variability in morning light exposure and intensity for

younger and older adults was 53% and 49%. Therefore, older adults vary almost as much within

person as they do between persons. The addition of intraindividual variability data sheds new

light on age-related differences in light exposure and adds an important caveat to assumptions

made about age-differences based on mean-level data. The mean-level differences tell us that

younger and older adults differ in the timing of their light exposure and the intensity of light

received, overall. The intraindividual variability data tells us that these age-related differences

do not hold across all older/younger adults and that there is significant variability in the timing

and intensity of light received within person for younger and older adults.

Overall, younger adults showed as much or more variability in arousal within than they did

between persons (50% to 71%) while older adults showed more between person variability and

less variability day-to-day for four out of the five variables (22% to 53%). Few studies have

examined variations in arousal on a daily basis within individuals. One recent study examined

age-related differences in variability in the cortisol awakening response (CAR; Almeida, Piazza,

& Stawski, 2009). While the implications of the CAR are not completely understood, lower









Multi-level modeling analyses showed that social rhythmicity predicted sleep at the between

person level and light, arousal, and affect predicted sleep at both within and between person

levels for younger adults. For older adults, light, arousal, and affect predicted sleep at both

between and within person levels. The potential mediating effects of light, arousal, and affect

could not be tested as there were insufficient associations between the predictor, potential

mediators, and outcome variables for both age groups. Results provide support for the

relationship between social rhythmicity, light, arousal, affect, and sleep for younger adults and

for light, arousal, affect and sleep for older adults. Further research is needed to test the

mediating effects of light, arousal, and affect.









attention given that our primary object of study humans are dynamic with changing and

sometimes unpredictable behaviors (Ram, Lindenberger, & Blanchard-Fields, 2009).

Consequently, the use of relatively static representations of behaviors (e.g., one time of

measurement) may not capture the full phenomena of human activity (West, 1985). Aim 1 of the

study extracted the amount of between and within person variability in the constructs of interest.

The importance of aim 1 is two-fold. First, a minimum amount of within person variability must

be present to warrant conducting the analyses for aim 2 (examining fluctuations over time).

Second, intraindividual variability in general, and age-related differences in intraindividual

variability in particular, have not been investigated for many of the variables studied.

Documenting the presence of intraindividual variability is an important first step in validating the

use of intraindividual variability analyses in the social sciences (Boker, Molenaar, &

Nesselroade, 2009). Boker, Nesselroade, and Molenaar (2010) commented on the need to focus

on the time dynamics of variability once the amount of variability has been established.

Consequently, the analyses of aim 1 contribute to the foundational literature in intraindividual

variability by documenting the amount of variability present in social rhythmicity, light, arousal,

affect, and sleep in younger and older adults.

For social rhythmicity, two estimates of variability were generated. The first traditional

estimate is an overall estimation of the variability of the individual that is based on two weeks of

data. Younger adults, on average, had a social rhythmicity score of 2.36 while older adults on

average scored 4.10. The social rhythm scale ranges from 0 to 7 with higher scores indicating

more regularity. Therefore, older adults were, overall, more regular than irregular in their daily

activities while younger adults were more irregular than regular in the timing of daily activities.









In terms of research examining the association between affect and sleep, a small number

of studies have examined the association between daily levels of affect and sleep. Berry and

Webb (1985) examined the association between mood (measured by the Lorr Mood Test; Lorr,

Daston, & Smith, 1967) and sleep (EEG data) in women aged 56 66 years. Affect was found to

be significantly associated with sleep, such that positive affect (cheerful and energetic moods)

was associated with decreased sleep onset latency and negative affect (anxiety) was associated

with decreased sleep efficiency and greater sleep onset latency (Berry & Webb, 1985a).

Another study by the same authors found similar results with positive affect associated with

increased sleep efficiency and total sleep time and negative affect associated with increased

wake time after sleep onset (Berry & Webb, 1985b). Affect was also found to be associated with

sleep in a study by Totterdell and colleagues (1994). Individuals that were higher in positive

affect experienced an earlier onset of the sleep period (Totterdell, Reynolds, Parkinson, &

Briner, 1994).

Finally, a recent study by McCrae and colleagues (2008) demonstrated an association

between affect and sleep over an extended data collection period (14 days). On days when

individuals reported higher levels of positive affect they also reported higher levels of sleep

quality. In terms of the overall sample, individuals who reported lower positive affect reported

greater overall total wake time. For negative affect, on days when individuals reported higher

levels of negative affect they were more likely to report below average sleep quality. In terms of

the overall sample, between groups analyses showed that on average, higher negative affect was

associated with greater total wake time and lower sleep quality (McCrae et al., 2008). In

summary, both affective disorders and negative affect appear to be associated with impaired

sleep while positive affect is associated with better sleep outcomes.










Table 5-24. Multilevel model with affect variables predicting sleep in older adults
Fixed Effects
SOL WASO TWT TST NWAK SQR
Predictor Variable B (SE) t (df) B (SE) t (d) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df)
Within person
time -0.19 -0.87 -0.40 -1.24 -0.40 -0.69 0.50 0.67 -0.03 -2.28 -0.01 -1.39
(0.22) (56.97) (0.33) (79.80) (0.58) (61.09) (0.75) (70.38) (0.01) (51.96)* (0.01) (80.36)
pleasant gridcentered -0.88 -1.11 -2.71 -2.29 -3.81 -2.41 5.50 2.55 -0.06 -1.96 0.13 4.08
(0.79) (21.51) (1.19) (19.56)* (1.58) (16.36)* (2.16) (7.49)* (0.03) (23.35) (0.03) (25.04)***
PANAS positivecentered 0.11 0.44 -0.20 -0.55 0.02 0.04 -0.95 -1.09 -0.02 -1.55 0.02 1.95
(0.24) (17.70) (0.37) (13.86) (0.58) (26.79) (0.87) (28.50) (0.01) (27.54) (0.008) (25.43)
PANAS negativeentered -0.09 -0.23 -0.63 -1.15 -1.04 -1.11 1.88 1.43 -0.0004 -0.02 0.004 0.30
(0.38) (561.03) (0.55) (557.18) (0.94) (541.61) (1.31) (22.59) (0.02) (558.45) (0.01) (579.75)

Between person
pleasant gridme. -2.12 -0.84 -7.07 -2.25 -9.79 -1.62 20.14 -2.44 -0.28 -1.44 0.17 2.44
(2.52) (49.17) (3.14) (50.40)* (6.04) (49.72) (8.25) (47.91)* (0.20) (47.69) (0.07) (49.83)*
PANAS positiveme. -0.01 -0.02 0.24 0.47 -0.53 0.54 2.94 2.17 0.03 1.03 0.01 0.56
(0.41) (48.54) (0.52) (49.70) (0.99) (49.37) (1.35) (47.05)* (0.03) (47.63) (0.01) (49.54)
PANAS negativeme. 0.16 0.37 -0.56 -1.02 -1.11 -1.04 -2.51 -1.72 -0.07 -2.06 0.01 0.75
(0.45) (50.23) (0.55) (51.52) (1.06) (50.15) (1.46) (49.42) (0.03) (47.71)* (0.01) (49.91)

Random Effects
SOL WASO TWT TST NWAK SQR
B (SE) Z B (SE) Z B(SE) Z B (SE) Z B (SE) Z B (SE) Z
Within person
time 0.11 0.26 0.26 0.36 6.68 0.92 0.57 0.12 0.002 1.61 0.001 0.12
(0.44) (0.73) (2.93) (4.60) (0.001) (0.001)
pleasant gridcentered 6.46 1.12 16.99 1.20 3.13 0.16 6.18 0.11 0.01 0.74 0.02 1.65
(5.79) (14.13) (19.65) (54.71) (0.01) (0.01)
PANAS positiveme. 0.80 1.04 2.28 1.08 4.10 1.17 13.10 1.59 0.001 1.03 0.001 0.96
(0.77) (2.11) (3.51) (8.24) (0.001) (0.001)
PANAS negativemea 0.00 0.00 0.00 0.00 0.00 0.00 1.47 0.16 0.00 0.00 0.00 0.00
(9.50)
Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number
of awakenings, SQR=sleep quality rating. *Correlation is significant at thep < 0.05 (2-tailed), **correlation is significant at thep <
0.01 (2-tailed),***correlation is significant at thep < 0.001 (2-tailed), Variance is too small to be estimated; the final Hessian matrix
was not positive definite, although all convergence criteria were satisfied.









LIST OF ABBREVIATIONS


NWAK Number of Awakenings: the number of awakenings during the night.

SOL Sleep Onset Latency: the amount of time it takes to fall asleep.

SQR Sleep Quality Rating: the rating of the quality of sleep ranging from poor
to excellent on a 5-point scale.

SRM Social Rhythm Metric: a weekly or daily score ranging from 0-7 with
higher scores indicating greater regularity.

TST Total Sleep Time: the total amount of time spent sleeping during the night.

TWT Total Wake Time: the total amount of time spent awake during the night
including the time to fall asleep (SOL), the time spent awake during the
night (WASO), and the time spent awake in the morning before getting
out of bed.

WASO Wake Time after Sleep Onset: the amount of time spent awake during the
night after falling asleep and before awakening in the morning.











Daily Measure-
Pre-Sle Scal


Please indicate how much you experienced each of the following as
you fell asleep last night

not at all slightly moderately a lot extremely
Heart racing, pounding or beating irregularly 0 0 0C
A jittery, nervous feeling in your body 0 0 0 0 0
Shortness of breath or labored breathing 0 0 0 0 0
A tight, tense feeling in your muscles 0 0 0 0 0
Cold feeling in your hands, feet or your body in general 0 0 0 0)
Have stomach upset (knot or nervous feeling in stomach, O C C) COO
heartburn, nausea, gas etc)
Perspiration in palms of your hands or other parts of 0 0 0 0 0
your body
Dry feeling in mouth or throat 0 0 0 0 0
Worry about falling asleep 0 0 0 0 0
Review or ponder events of the day 0 0 0 0 0
Depressingor anxious thoughts 0 0 0 0 0
Worry about problems other than sleep 0 0 0 0 0
Being mentally alert, active 0 0 0 0 0
Can't shut off your thoughts 0 0 0 0 0
Thoughts keep running through your head 0 0 0 0 0
Being distracted by sounds, noise in the environment 0 0 O ) 0
(e.g ticking of dock, house noises, traffic 0o 0 0 0









For older adults, arousal predicted sleep at both level 1 and level 2 (see table 5-22). For

the within person level (level 1), the PSAS cognitive subscale significantly predicted SOL (p =

0.94, t(33.19) = 2.61, p <.05), WASO (P = 2.11, t(50.85) = 3.04, p <.01), TWT (P = 3.55,

t(40.00) = 3.56,p <.05), TST (3 = -3.80, t(41.83) = -2.48,p <.05), NWAK (3 = 0.08, t(33.81)=

4.81,p <.001), and SQR (P = -0.09, t(35.01) = -5.25,p <.001). These results suggest that on

days when older adults experience more cognitive arousal before bed, they take longer to fall

asleep, have more awake time during the night, have less total sleep time, more awakenings and

lower sleep quality ratings. Random effects were significant for the PSAS cognitive scale

predicting WASO and the SQR suggesting that the effects mentioned above vary in magnitude

between individuals. There was also a significant within person effect for time (P = -0.02,

t(62.28) = -2.25, p <.05) indicating that older adults' sleep quality ratings decreased over the

course of the study. This effect is qualified by a significant random effect for time indicating

that the magnitude of this effect varies significantly within individuals. For level 2 (between

person effects), the PSAS cognitive subscale significantly predicts SOL (P = 1.78, t(46.49)=

2.38, p <.05), WASO (P = 2.58, t(48.02) = 2.68, p <.05), TWT (P = 5.07, t(47.15) = 2.69, p

<.05), and SQR (P = -0.07, t(45.85) = -3.41,p <.001). Overall higher levels of cognitive arousal

were associated with overall longer times to fall asleep, more wake time during the night, and

poorer sleep quality ratings in older adults. Additionally, at level 2, the arousal grid significantly

predicted SQR (P = -0.14, t(46.45) = -2.96, p <.05). Older adults who had overall higher scores

on the arousal grid (indicating sleepiness) reported poorer sleep quality.

For the affect variables, there were significant predictions of sleep at level 1 and 2 for

younger adults (see table 5-23). At the within persons level, time significantly predicted SOL (p

= -0.25, t(93.38) = -2.26, p <.05), the pleasant grid predicted SQR (P = 0.13, t(53.26) = 4.84, p









(insertion of an indwelling intravenous catheter placement), individuals with higher levels of free

cortisol had impaired sleep as represented by lower sleep efficiency, more time spent in sleep

stage one, and greater EEG beta activity during NREM sleep (Prinz, Bailey, & Woods, 2000).

In summary, while numerous studies have supported the relationship between cognitive

arousal and poor sleep, it is difficult to separate the effects of both cognitive and somatic arousal.

For example, while an individual is cognitively aroused, they may simultaneously be

physiologically aroused and vice versa. Additionally, each component of arousal may work to

increase the other. In order to encompass both of these possibly interacting constructs, the

present study will assess both components of arousal (cognitive and somatic).

Theoretical model Linking sleep and arousal. Several models of insomnia have been

proposed which implicate the role of arousal in maintaining poor sleep. The model that is most

relevant to the present study is the cognitive model of insomnia proposed by Harvey (2000).

While labeled the 'cognitive' model of insomnia, Harvey's model addresses both physiological

and cognitive aspects of arousal. Harvey proposes that pre-sleep cognitive arousal (occurring

both during the day and just before bed) will result in impaired sleep (Harvey, 2002). The

specific steps of the model are illustrated in Figure 2-4. First, the individual experiences

negatively toned cognitive activity (e.g., worry). Second, the negatively toned cognitive activity

triggers autonomic arousal resulting from activation of the sympathetic nervous system. Third,

based on the increased arousal, the individual will experience selective attention and monitoring

of threat-related cues. These cues consist of external and internal cues that indicate that the

individual is not sleeping. For example, the individual may become aware of external noises in

the environment or become aware of muscle tension. The threat-related cues then feed back into

the negatively toned cognitive activity. The forth stage of the model consists of a distorted













If yes, explain: 6
If yes, explain:


If yes, does this interfere with your sleep?

O s
0 -


If yes, explain how:


Do you currently have breathing problems such as asthma or emphysema?
Q yes
O es
Qon

If yes, explain:
If yes, explain:


If yes, does this interfere with your sleep?

Syes
0 n


If yes, explain how:


Do you currently have diabetes?
O es
0Q


If yes, explain:
If yes, explain:


I










Daily Measures
Atitd Scl


This scale consists of a number of words that describe different
feelings and emotions.

Read each item and then pick the appropriate answer next to that
word. Indicate to what extent you have felt this way yesterday.
Very slightly or A little Moderately Quite a bit Extremely
not at all
interested 0 0 0 0 0
Disessed O O O
Exted 0 0 0 0 0
Upset 0 0 0 0 0
Strong Q Q Q Q Q
Guilty 0 0 0 0 0
Scared 0 0 0 0 0
Hoste 0 0 0 0 0
Enthusiastic 0 0 0 0 0
Proud 0 0 0 0 0
Irritable 0 0 0 0 Q
,Ale O O O O
ooAd 0 0 0 0o

Inspired O O O O O
Ne vus Q 0 0 0 0
Determined O O O O
DO O O O O
A^ ve O O O O O
Jittery O O 0 0
A,,ve 0O 0 0 0 0
Afraid 0 0 0 0 0













Bendetti, F., Barbini, B., Campori, E., Fulgosi, M. C., Pontiggia, A., & Colombo, C. (2001).
Sleep phase advance and lithium to sustain the antidepressant effect if total sleep
deprivation in bipolar depression: New findings supporting the internal coincidence
model? Journal of Psychiatric Research, 35, 323-329.

Berry, D. T. R., & Webb, W. B. (1985a). Mood and sleep in aging women. Journal of
Personality and Social Psychology, 49(6), 1724-1727.

Berry, D. T. R., & Webb, W. B. (1985b). Sleep and cognitive functions in normal older adults.
Journals of Gerontology, 40(3), 331-335.

Bhagwagar, Z., Hafizi, S., & Cowen, P. J. (2003). Increase in concentration of waking salivary
cortisol in recovered patients with depression. American Journal ofPsychiatry, 160(10),
1890-1891.

Bogdan, A., Bouchareb, B., & Touitou, Y. (2001). Ramadan fasting alters endocrine and
neuroendocrine circadian patterns. Meal-time as a synchronizer in humans? Life Sciences,
68(14), 1607-1615.

Boker, S. M., Molenaar, P. C. M., & Nesselroade, J. R. (2009). Issues in intraindividual
variability: Individual differences in equilibria and dynamics over multiple time scales.
Psychology andAging. 24(4), 858-862.

Bonnet, M. H., & Arand, D. L. (1995). 24-hour metabolic-rate in insomniacs and matched
normal sleepers. Sleep, 18(7), 581-588.

Bonnet, M. H., & Arand, D. L. (1999). Level of arousal and the ability to maintain wakefulness.
Journal of Sleep Research, 8(4), 247-254.

Bootzin, R. R. (1972). A stimulus control treatment for insomnia. Proceedings of the 80th
annual convention of the american psychological association, 395-396.

Bouisson, J., & Swendsen, J. (2003). Routinization and emotional well-being: An experience
sampling investigation in an elderly French sample. Journals of Gerontology Series B-
Psychological Sciences and Social Sciences, 58(5), P280-P282.

Britton, J. (1997). Measurement of peak flow variability in community populations:
Methodology. European Journal ofRespiration, 10, 42-44.

Brown, L. F., Reynolds, C. F., Monk, T. H., Prigerson, H. G., Dew, M. A., Houck, P. R., et al.
(1996). Social rhythm stability following late-life spousal bereavement: associations with
depression and sleep impairment. Psychiatry Research, 62(2), 161-169.









rhythm disruption of events in their life during an eight week period prior to the manic episode

(Malkoff-Schwartz et al., 1998). Interestingly, individuals who experienced a manic episode

were more likely to have experienced a social rhythm disruption in the pre-onset period while

individuals who experienced a depressive episode were not (Malkoff-Schwartz et al., 1998).

Researchers hypothesized that the eight-week study period may have been long enough to

witness a manic episode triggered by the disruptive event but not long enough to observe the

development of a depressive episode. A follow-up study examined the relationship between

events disrupting social rhythms and manic and depressive episodes over 8 and 20 week study

periods. The results replicated the previous findings in that participants experiencing bipolar

manic episodes were more susceptible to the influence of social rhythm disruptive events

compared to individuals diagnosed with depressive episodes (Malkoff-Schwartz et al., 2000).

Another study found that there was a significant difference in social rhythm regularity

between individuals diagnosed with bipolar disorder and controls (Ashman et al., 1999).

Individuals ranging in age from 29 61 who met the criteria for rapid cycling bipolar disorder

were compared to matched control participants in terms of social rhythms (Ashman et al., 1999).

Additionally, within-subjects analyses were performed to evaluate whether social rhythms varied

with mood state (hypomania, depression, and euthymia). Social rhythm regularity was not found

to vary systematically with mood state although power analyses indicated a larger sample size

was needed to detect an effect. Interestingly, however, individuals diagnosed with bipolar

disorder had significantly lower social rhythmicity scores compared to matched controls.

Compared to control participants, individuals diagnosed with bipolar disorder less regularly had

breakfast, started work, had lunch, exercised, had an evening snack, watched an evening news

program, watched another television program, or took a nap.





















Thank you for all of your effort completing the measures today.

Your time is very much appreciated!









14 days of repeated measurement for each individual. The ISD describes the extent to which an

individual's scores tend to vary around their mean score. The data was detrended in order to

control for the effects of time (e.g., time was 'residualized' out of the variables of interest).

Bivariate correlations were then conducted between the ISD of the predictor variables (social

rhythmicity, light, arousal, and affect) and sleep. Given the exploratory nature of this aim, all

predictor and sleep variables were included in the analyses. Significant correlations between the

ISDs and sleep suggested that within person variability in social rhythmicity, light, arousal and

affect is associated with sleep.

Aim 2 also allowed for examination of habitual social behaviors within the individual in

relation to sleep. For example, a person may be regular or irregular in their social activities but

that pattern may be habitual for them (see Figure 5). For example, an individual could habitually

be regular in social rhythms and days of irregular social rhythms would be atypical or 'variable'

for them. Conversely, someone could habitually be irregular in their social rhythms and days

where they are more regular in their activities would be considered atypical or 'variable' for

them. Consequently, the level 1 or within person analysis examined the extent to which

variability within the individual's social rhythmicity was related to sleep. In addition to social

rhythmicity, aim 2 also provided information on fluctuations in the remaining predictor variables

(light, arousal, affect) in relation to sleep.

Hypothesis

In general, it was proposed that higher levels of variability in the predictor variables

would be associated with poorer sleep. In terms of social rhythmicity, one study (Monk et al.,

1991) found that individuals diagnosed with depression were more likely to have greater

intraindividual variability on the SRM-17 compared to healthy controls. This study provides

support for the association of variability in social rhythmicity with depressed mood but does not









= -0.82, p < 0.001 and UMACL-R energetic and UMACL-R tenseness for older adults, r -

0.90, p < 0.001). The residualized variables of the arousal grid and UMACL-R energetic (for

younger adults), and UMACL-R tenseness and UMACL-R energetic (for older adults) were used

in the following MLM analyses.

In the first step of the MLM, a null model was estimated for each of the six sleep outcomes

(SOL, WASO, TWT, TST, NWAK, and SQR) that included only fixed and random intercepts

for these variables. This model served as a comparison to later models that included the

predictor variables and was used to derive the intraclass correlation coefficient used in aim 1. In

step two of the MLM, the linear effects of time were added to the model to control for any

systematic change in the data resulting from the effects of 14 days of measurement. In step three

of the MLM, the estimates of the fixed and random intercepts and fixed and random linear slopes

were added to the model.

In the models with SRM predicting sleep, there were significant predictors at the within

and between person levels for younger adults (see Table 5-17). At the within persons level, time

significantly predicted SOL (P = -0.26, t(90.64) = -2.46, p <.05) indicating that the amount of

time it took younger adults to fall asleep decreased over the 14 study days. At the between

person level, more irregularity overall was associated with overall longer SOL (p = -3.40,

t(48.70) = -2.51,p <.05), more TWT (P = -5.44, t(49.97) = -2.03,p <.05), and poorer SQR (P =

0.16, t(48.58) = -2.03,p <.05). These results suggest that the more irregular younger adults were

in their daily activities, overall, the longer it took to fall asleep, the more time they spent awake

in total during the night, and the lower they rated their sleep quality. For older adults, there were

significant predictors at the within person level (level 1) with time significantly predicting

NWAK (P = -0.03, t(50.77) = -2.30, p <.05) suggesting that the number of awakenings decreased



































Figure 2-2. Social Zeitgeber Theory and Internal Trigger Hypothesis. Blue indicates addition of
Internal Trigger hypothesis to the Social Zeitgeber Theory. Red denotes area of
investigation in present study.













Bryk, D. T. R., & Raudenbush, S. W. (1992). Hierarchial Linear Models for Social and
Behavioral Research: Applications and Data Analysis Methods. Newbury Park, CA:
Sage.

Campbell, S. S., & Daston, P. (1991). Bright light treatment of sleep disturbance in older
subjects. Sleep Research, 20, 448.

Campbell, S. S., Dawson, D., & Anderson, M. W. (1993). Alleviation of sleep maintenance
insomnia with timed exposure to bright light. Journal of the American Geriatrics Society,
41(8), 829-836.

Campbell, S. S., Kripke, D. F., Gillin, J. C., & Hrubovcak, J. C. (1988). Exposure to Light in
Healthy Elderly Subjects and Alzheimers Patients. Physiology & Behavior, 42(2), 141-
144.

Card, J. P. (Ed.). (1999). Anatomy of the mammalian circadian time-keeping system. Boca Raton:
CRC Press.

Carney, C., Edinger, J., Meyer, B., Lindman, L., & Istre, T. (2006). Daily activities and sleep
quality in college students. Chronobiology International: The Journal of Biological &
Medical Rhythm Research, 23(3), 623-637.

Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory of
socioemotional selectivity. American Psychologist, 54(3), 165-181.

Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000). Emotional experience in
everyday life across the adult life span. Journal ofPersonality and Social Psychology,
79(4), 644-655.

Charles, S. T., Reynolds, C. A., & Gatz, M. (2001). Age-related differences and change in
positive and negative affect over 23 years. Journal of Personality and Social Psychology,
80(1), 136-151.

Chuo, T. C., Scammell, T. E., Gooley, J. J., Gaus, S. E., Saper, C. B., & LU, J. (2003). Critical
role of dorsomedial hypothalamic nucleus in a wide range of behavioral circadian
rhythms. The Journal ofNeuroscience, 23, 10691-10702.

Cortoos, A., Verstraeten, E., & Cluydts, R. (2006). Neurophysiological aspects of primary
insomnia: Implications for its treatment. Sleep Medicine Reviews, 10(4), 255-266.

Czeisler, C. A., Kronauer, R. E., Allan, J. S., Duffy, J. F., Jewett, M. E., Brown, E. N., et al.
(1989). Bright light induction of strong (Type-0) resetting of the human circadian
pacemaker. Science, 244(4910), 1328-1333.




Full Text

PAGE 1

1 THE MEDIATION OF THE RELATIONSHIP BETWEEN SOCIAL RHYTHMICTY AND SLEEP BY LIGHT, AROUSAL, AND AFFECT IN BOTH YOUNGER AND OLDER ADULTS By NATALIE DEIDRE DAUTOVICH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVE RSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 NATALIE DEIDRE DAUTOVICH

PAGE 3

3 To Pokey, Tika, and Pip for helping me to keep everything in perspective

PAGE 4

4 A CKNOWLEDGMENTS I would like to express my appreciation to my dissertation chair, Christina McCrae. She has encouraged me to believe in myself, to strive for greatness, and fostered my interest in the field of behavioral sleep medicine. As a result, she h as enabled my grow th as a researcher. I would also like to acknowledge my Masters and Dissertation committee members Meredeth Rowe, Ken Rice, and Robin West. I thank Meredeth Rowe for providing helpful feedback and being kind enough to share valuable equ ipment I thank Ken Rice for his insightful feedback and challenging questions as I have progressed through the counseling program. I acknowledge Robin West for reinforcing the broader lifespan perspective in relation to my education and research. I wou ld like to thank Amber Gum and Larry Schonfeld for their encouragement, support, and patience and Deborah Hedgecock and Sarah Cobb for their statistical insights. I could not have accomplished my objective data collection without the donation of Actiwatches by Christina McCrae or the unwavering, cheerful support and assistance of Christine Towler. I would like to thank Joseph Dzierzewski for his enthusiastic support for my ideas and willingness to discuss analytic issues. Ashley Stripling requires specia l acknowledgement for her emotional support her practical support with recruitment and data collection, and for sharing a passion for older adults. I would like to thank Timothy Monk for the use of the Social Rhythm Metric (SRM 17) and for his assistance in developing a daily rhythmicity score. I especially would like to acknowledge the efforts of both the younger and older adult participants who gave their time to this project. I greatly appreciated the prayers of Margaret Sams as I moved through this process I thank my 90 year old grandmother, Helen Dautovich, for showing me the beauty of aging. My work was supported by funding by the American Psychological Association Science Directorate, Dissertation Research Award. Finally, none of this project could have been accomplished without the unconditional support and confidence in my abilities of my wonderful

PAGE 5

5 parents, Mike and Angella Dautovich, the unfaltering encouragement of my sister Sonia Dautovich, and the daily support of David Sams who cared enough to walk each step of this journey with me.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................9 LIST OF FIGURES .......................................................................................................................11 LIST OF ABBREVIATIONS ........................................................................................................12 ABSTRACT ...................................................................................................................................13 CHAPTER 1 INTRODUCTION ..................................................................................................................15 Background: Influences on Biological Rhythms ....................................................................15 Circadian Rhythms Defined ............................................................................................15 Theoretical Model Influences on Biological Rhythms .................................................17 Social Rhythms .......................................................................................................................20 Soc ial Zeitgebers .............................................................................................................20 Evidence for Social Zeitgebers ........................................................................................20 Animal studies ..........................................................................................................21 Human studies ..........................................................................................................22 Regularity of Social Zeitgebers .......................................................................................23 Regularity and arousal ..............................................................................................24 Regularity and affect Depression ..........................................................................24 Regularity and affect Manic episodes ...................................................................25 Regularity and affect Anxi ety ...............................................................................27 Regularity of social rhythms and sleep ....................................................................27 Theoretical Models Relationship between Social Rhythms, Light, Affect, Arousa l, and Sleep ......................................................................................................................28 Sleep .......................................................................................................................................31 The Association Between Sleep and Light ......................................................................31 The Association between Sleep and Arousal ..................................................................31 The Association between Sleep and Affect .....................................................................36 Age ..........................................................................................................................................38 Age and Social Rhythms .................................................................................................38 Age and Sleep ..................................................................................................................39 Age in Relation to Mediators of Light, Arousal, a nd Affect ...........................................40 Summary .................................................................................................................................41 2 STATEMENT OF THE PROBLEM ......................................................................................48 Purpose of the Study ...............................................................................................................48

PAGE 7

7 Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep Fluctuate Daily both Within and Between Persons ...................................................49 Analysis ...........................................................................................................................49 Hypothesis .......................................................................................................................50 Aim 2: Identify how Variability in Social Rhythmicity, Light, Arousal, and Affect is Related to Sleep ...................................................................................................................50 Analysis ...........................................................................................................................50 Hypothesis .......................................................................................................................51 Aim 3: I dentify to What Extent the Daily Fluctuations in Social Rhythmicity, Light, Arousal, and Affect are Coupled with Daily Fluctuations in Sleep ....................................52 Analysis ...........................................................................................................................52 Hypothesis .......................................................................................................................53 Aim 4: Identify to What Extent Light, Arousal, and Affect Mediate the Relationship Between Social Rhythmicity and Sleep ..............................................................................54 Analysis ...........................................................................................................................54 Hypothesis .......................................................................................................................55 3 METHODS .............................................................................................................................57 Materials .................................................................................................................................57 Demographics and Health Survey ...................................................................................57 Beck Depression Inventory Second Edition ...................................................................58 State Trait Anxiety Inventory, Form Y1 (STAI Y1) ......................................................58 Social Rhythm Metric ......................................................................................................59 Light Exposur e ................................................................................................................60 Arousal ............................................................................................................................62 Affect ...............................................................................................................................65 Sleep Diary ......................................................................................................................66 Procedure ................................................................................................................................67 Participants .............................................................................................................................68 4 RESULTS ...............................................................................................................................73 Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep Fluctuate Daily both Within and Between Persons ...................................................73 Social Rhythmicity ..........................................................................................................73 Light Variables ................................................................................................................73 Arousal Variables ............................................................................................................74 Affect Variables ...............................................................................................................74 Sleep Variables ................................................................................................................75 Aim 2: Identify How Variability in Social Rhythmicity, Light, Arousal, and Affect is Related to Sleep ...................................................................................................................75 Aim 3: Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light, Arousal, and Affect are Coupled with Daily Fluctuations in Sleep ....................................77 Aim 4: Identify to What Extent Light, Arousal, and Affect Mediate the Relationship Between Social Rhythmicity and Sleep ..............................................................................83

PAGE 8

8 5 DISCUSSION .......................................................................................................................110 Review of the Findings .........................................................................................................110 Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep Fluctuate Daily both Within and Between Persons ...................................110 Aim 2: Identify How Variability in Social Rhythmicity, Light, Arousal, and Affect is Related to Sleep ......................................................................................................121 Aim 3: Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light, Arousal, and Affect are Coupled with Daily Fluctuations in Sleep ................125 Aim 4: Identify to What Extent Light, Arousal, and Affect Mediate the Relationship Between Social Rhythmicity and Sleep ................................................131 Strengths and Weaknesses of the Present Study ...................................................................131 Implications ..........................................................................................................................132 Future Directions ..................................................................................................................133 Conclusion ............................................................................................................................134 APPENDIX A BASELINE SURVEY ..........................................................................................................138 B DAILY SURVEY .................................................................................................................160 LIST OF REFERENCES .............................................................................................................176 BIOGRAPHICAL SKETCH .......................................................................................................190

PAGE 9

9 LIST OF TABLES Table page 41 Participant descriptive statistics for younger adults (N=50) ..............................................69 51 Summary of the variables used in the analyses. ................................................................84 52 Age related differences in mean level values for social rhythmicity, light, arousal, affect, and sleep variables ..................................................................................................90 53 Correlations of within person variability in social rhythmicity and mean level sleep in younger adults. (N=50) ..................................................................................................91 54 Correlati ons of within person variability in light and meanlevel sleep in younger adults. (N=50) ....................................................................................................................91 55 Correlations of within person variability in arousal and meanlevel sleep in younger adults. (N= 50) ....................................................................................................................92 56 Correlations of within person variability in affect and meanlevel sleep in younger adults. (N=50) ....................................................................................................................93 57 Correlation s of within person variability in social rhythmicity and mean level sleep in older adults. (N=50) .......................................................................................................93 58 Correlations of within person variability in light and meanlevel sleep in older adults (N=50) ................................................................................................................................94 59 Correlations of within person variability in arousal and meanlevel sleep in older adults. (N=50) ....................................................................................................................95 510 Correl ations of within person variability in affect and meanlevel sleep in older adults. (N=50) ....................................................................................................................96 511 Correlations among light variables between persons (below diagonal) and within person (above diagonal) for younger adults .......................................................................96 512 Correlations among light variables between persons (below diagonal) and within person (above diagonal) for older adults ...........................................................................97 513 Correlations among arousal variables between persons (below diagonal) and within person (above diagonal) for younger adults .......................................................................97 514 Correlations among arou sal variables between persons (below diagonal) and within person (above diagonal) for older adults ...........................................................................98 515 Correlations among affect variables between persons (below diagonal) and within perso n (above diagonal) for younger adults .......................................................................98

PAGE 10

10 516 Correlations among affect variables between persons (below diagonal) and within person (above diagonal) for older adults ...........................................................................98 517 Multilevel model with social rhythmicity variables predicting sleep in younger adults ...99 518 Multilevel model with social rhythmicity variables predicting sleep in older adults ......100 519 Multilevel model with light variables predicting sleep in younger adults .......................101 520 Mult ilevel model with light variables predicting sleep in older adults ............................103 521 Multilevel model with arousal variables predicting sleep in younger adults ...................104 522 Multilevel model with arousal variables predicting sleep in older adults .......................106 523 Multilevel model with affect variables predicting sleep in younger adults .....................108 524 Multilevel model with affect variables predicting sleep in older adults ..........................109

PAGE 11

11 LIST OF FIGURES Figure page 21 Mistleberger & Skenes (2004) theoretical model of temporal influences on biological rhythms. ............................................................................................................44 22 Social Zeitgeber Theory and Internal Trigger Hypothesis ................................................45 23 Mediated Model of Social Rhythmicity examined for both age groups in the present study. ..................................................................................................................................46 24 Cognitive Model of Insomnia (Harvey, 2000). .................................................................47 31 Example of hypothesized daily coupling of social rhythmicity and sleep ........................56 51 Percentage of total var iability in social rhythmicity that can be attributed to within person variability. ..............................................................................................................85 52 Percentage of total variability in light measures that can be attributed to within person variability ..............................................................................................................86 53 Percentage of total variability in arousal measures that can be attributed to within person variability. ..............................................................................................................87 54 Percentage of total variability in affect measures that can be attributed to within person variability. ..............................................................................................................88 55 Percentage of total variability in sleep variables that can be attributed to within person variability. ..............................................................................................................89 61 Relationships in the mediated model of social rhythmicity that were observed in younger adults. .................................................................................................................136 62 Relationships in the mediated model of social rhythmicity that were observed in older adults. ......................................................................................................................137

PAGE 12

12 LIST OF ABBREVIATION S NWAK Number of Awakenings: the number of awakenings during the night. SOL Sleep Onset Latency: th e amount of time it takes to fall asleep. SQR Sleep Quality Rating: the rating of the quality of sleep ranging from poor to excellent on a 5 point scale. SRM Social Rhythm Metric: a weekly or daily score ranging from 0 7 with higher scores indicating great er regularity. TST Total Sleep Time: the total amount of time spent sleeping during the night. TWT Total Wake Time: the total amount of time spent awake during the night including the time to fall asleep (SOL), the time spent awake during the night (WASO), and the time spent awake in the morning before getting out of bed. WASO Wake Time after Sleep Onset: the amount of time spent awake during the night after falling asleep and before awakening in the morning.

PAGE 13

13 Abstract of Dissertation Presented to the G raduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE MEDIATION OF THE RELATIONSHIP BETWEEN SOCIAL RHYTHMICTY AND SLEEP BY LIGHT, AROUSAL, AND AFFECT IN BOTH YOUNGER AND OLDER ADULTS By Natalie Deidre Dautovich August 2010 Chair: Christina McCrae Major: Counseling Psychology Light currently reigns as the dominant factor influencing human circadian rhythms such as sleep. Recently, interest has shifted to other external factors which may influence circadian systems. Social activities (e.g., meal times, timing of exercise) are thought to exert effects on the timing of circadian rhythms. S ocial activities may entrain circadian rhythms either directly or indirectl y by influencing exposure to light. Little research, however, has investigated mechanisms other than light that underlie the relationship between the rhythms of social activities and sleep. The present study examined the role of arousal and affect in addition to light, as potential mediators of the relationship between social rhythms and sleep. The overarching goals were to 1) examine the amount of variability within social rhythmicity, light, arousal, affect and sleep; 2) examine how variability in soc ial rhythmicity, light, arousal, and affect is related to sleep; 3) examine the day to day and overall relationships between social rhythmicity, light, arousal, affect, and sleep; and 4) examine the mediation of the relationship of social rhythmicity and s leep by light, arousal, and affect. Each of these goals was investigated in both younger and older adults.

PAGE 14

14 Younger ( n = 50) and older ( n = 50) community dwelling individuals were recruited from the North Central Florida area and online. Participants completed daily online diaries for 14 consecutive days. Social activities were assessed using the Social Rhythm Metric (SRM 17), light exposure was evaluated with the Light Exposure Scale (LES), arousal was assessed using the Affect Grid, the Pre Sleep Arousa l Scale (PSAS), and the UWIST Mood Adjective Checklist Revised (UMACL R), affect was evaluated with the Affect Grid and the Positive and Negative Affect Scale (PANAS), and sleep was assessed using a sleep diary. Results indicated that younger adults we re highly variable within person for all variables. Younger adults were significantly more variable on all measures than older adults except for the PSAS somatic scale. Despite showing less variability compared to younger adults, older adults still demon strated considerable variability. Agerelated differences in mean levels were observed for social rhythmicity, light exposure, arousal, affect, and sleep. Variability in social rhythmicity, light, arousal, and affect was associated with worse sleep for y ounger adults and better and worse sleep for older adults. Multi level modeling analyses showed that social rhythmicity predicted sleep at the between person level and light, arousal, and affect predicted sleep at both within and between person levels for younger adults. For older adults, light, arousal, and affect predicted sleep at both between and within person levels. The potential mediating effects of light, arousal, and affect could not be tested as there were insufficient associations between the predictor (social rhythmicity), potential mediator (light, arousal, and affect), and outcome variables (sleep) for both age groups. Results provide support for the relationship between social rhythmicity, light, arousal, affect, and sleep for younger adul ts and for light, arousal, affect and sleep for older adults. Further research is needed to test the mediating effects of light, arousal, and affect.

PAGE 15

15 CHAPTER 1 INTRODUCTION Light currently reigns as the dominant factor influencing human circadian r hythms such as sleep. Recently, however, interest has shifted to other external factors that may influence circadian systems. Social influences (e.g., meal times, timing of exercise) are thought to exert effects on the timing of biological rhythms. Rese arch suggests that social influences may influence our biological rhythms either directly or indirectly, by influencing exposure to light. Additionally, constructs such as arousal and affect may mediate the relationship between social rhythms and sleep. A review of the literature is presented below. First, the background on the influences on biological rhythms will be discussed. Second, research related to social rhythms will be examined, and social rhythms will be linked to the possible mediators of li ght, arousal, and affect as well as to the outcome variable of sleep. Third, a theoretical model will be proposed to link social rhythms, the mediators of light, arousal, affect, and the outcome variable of sleep. Fourth, the relationship between the pot ential mediators and sleep will be explored. Finally, the impact of age on the variables of interest will be discussed. Background: Influences on Biological Rhythms Circadian Rhythms Defined In order to examine the influence of social rhythms on sleeping behavior, it is necessary to first describe the system underlying the timing of the sleep/wake cycle the circadian system. Circadian rhythms refer to a cycle within an individual that is self sustained and fluctuates roughly on a daily/24 hour cycle ( circa meaning around and diem meaning day). In order for a rhythm to be classified as 'circadian', the rhythm must persist in absence of environmental stimuli. For example, free running studies will place an organism in a light deprived setting in

PAGE 16

16 order to determine if a rhythm (e.g., hormonal fluctuation) will persist without the environmental cue of light. Circadian rhythms are essential and provide an adaptive function in that they are a time keeping mechanism within the individual that enabl es the individual to prepare for changes in the physical environment that are associated with day and night (Vitaterna, Takahashi, & Turek, 2001) Consequently, the individual not only responds to changes in light in the environment, but also has an intern al pacemaker that prepares the individual for environmental challenges. While circadian rhythms are self sustaining, they can be influenced by external cues. The ability of an external cue (e.g., the light dark cycle) to influence a circadian rhythm by c ausing it to align to a new schedule is referred to as 'entrainment'. During daylight savings time, we are able to adjust our sleep/wake cycles based on the fact that the external light cues have changed and the circadian rhythm can become entrained by th ese cues. An environmental cue that can entrain circadian rhythms is referred to as a zeitgeber, meaning time giver. The biological mechanism underlying the circadian system in humans is the suprachiasmatic nucleus (SCN). The SCN is located in the anterior hypothalamus. In a review of studies on the SCN, Fuller and colleagues (2006) described the influence of the SCN on the timing of the circadian clock as occurring through many different mechanisms (Fuller, Gooley, & Saper, 2006. ) For example, t he SCN may exert its influence on the circadian rhythm through projections to the mesopontine tegmental nuclei (McCarley & Massaquoi, 1992) with the greatest dendritic branching of the SCN occurring in the subparaventricular zone (SPZ; Watts, Swanson, & Sa nchezwatts, 1987) Fuller and colleagues (2006) identified the primary SCN pathway that exerts control on the circadian clock as involving the SCN, the SPZ, and the dorsomedial hypothalamic nucleus (DMH). The DMH in turn sends projections to the sleep

PAGE 17

17 (v entrolateral preoptic nucleus) and wake (lateral hypothalamus) centers of the brain (Chuo et al., 2003) In summary, the circadian system plays a vital role in regulating many human functions including the sleep/wake cycle. The area of interest in the pre sent study is the environmental factors that may influence the circadian clock and consequently impact sleep. Of particular interest are social stimuli and the regularity of social stimuli within an individuals daily life. Mistleberger and Skene (2004) proposed a theoretical model illustrating the influence of environmental factors on the circadian clock. This model provides a framework for the current investigation which examines the social components of the model. Theoretical Model Influences on Bio logical Rhythms Mistleberger and Skene (2004) propose that three main forces influence biological rhythms: 1) light or photic sources, 2) behavioral state, and 3) social stimuli (see Figure 2 1 ). The three influences on the circadian system will be discuss ed below. Light is the primary, most dominant zeitgeber influencing circadian rhythms (Mistlberger & Skene, 2004) Light influences the circadian pacemaker, located in the suprachiasmic nucleus, by entering the retina and traveling along the retino hypo thalamic tract (RHT; Klein, Moore, & Reppert, 1991) Light can be derived from direct sunlight or from another light source (e.g., indoor fluorescent lighting). The behavioral state is defined as behaviors such as opening and closing eyes, or entering or exiting the house that influence exposure to light (Mistlberger & Skene, 2004) Behavioral state can be directly interpreted as the behavioral responses or actions in which an individual engages. Behavioral states influence the circadian rhythm by exerting an influence over the timing and exposure to light.

PAGE 18

18 The category of social stimuli refers to the sights, sounds, smells, and other stimuli that an individual may encounter during the day. T he area of social stimuli was under investigation i n the present study. Daily patterns of behavior originally were based on availability of light. We have since become increasingly detached from the regularizing power of light (e.g., since the advent of electricity and technology). In place of the absolute regularizing powers of light, daily life has increasingly been governed by exigencies of daily social life (Monk, Flaherty, Frank, Hoskinson, & Kupfer, 1990) For example, individuals will stay active long after daylight ends in order to engage in an y number of social activities. A student may remain awake for a portion of the night in order to finish an assignment or a parent may spend the night awake tending to a newborn baby. Through the advancement of technology, we have increasingly turned to a rtificial light in order to extend the time of day we can engage in activities. Although light is typically present during these activities, it is proposed that we cannot deny the influence of the activity itself in terms of influencing our biological rhy thms. For example, a student who delays the onset of their sleep period in order to finish an assignment may be motivated to stay awake by both their desire to finish the assignment as well as by the influence of light. In fact, it could be argued that l ight alone would not be sufficient to maintain the alertness of the student without the accompanying motivation/arousal/anxiety to complete the assignment. Consequently, this study aimed to investigate the association between social stimuli and biological rhythms while still accounting for the influence of light. Social stimuli can influence the circadian system via several pathways, including direct pathways and indirect pathways. Indirect pathways include affecting the behavioral state of the individua l (e.g., exerting an autonomic affect on pupil size) which would affect the amount of light influencing the SCN and impact the circadian system. Social stimuli could also directly

PAGE 19

19 affect the circadian system without needing to indirectly provide access to light. Mistlberger and Skene (2004) suggest a number of pathways through which social stimuli could directly access circadian pacemakers, including sensory specific pathways (e.g., smells, sights, sounds, and tactile stimuli) or a more general input path way (e.g., a pathway encoding arousal and/or stress). The specific biological mechanisms through which social stimuli could access the circadian system include neural afferents from the median raphe, thalamic intergeniculate leaflet, or cholinergic cell g roups (Card, 1999; Mistlberger, Antle, Glass, & Miller, 2000; Mrosovsky, 1996) An additional explanation for the direct effects of social stimuli on the circadian system involves classical conditioning (Mistlberger & Skene, 2004) When a non photic socia l stimulus, such as an alarm clock, is repeatedly paired with a photic stimulus, such as opening the blinds or turning on lights, the social stimulus could begin to influence the circadian rhythm in a manner similar to that of the photic stimulus (Amir, Be aule, Arvanitogiannis, & Stewart, 2002; Amir & Stewart, 1996) A final consideration is that it is possible the social stimuli could exert effects on behavior without impacting the circadian system or timing of daily events. Social stimuli would be s aid to be masking the true circadian phase if they were to affect the timing of behaviors without altering the circadian rhythm. For example, an individual may choose to wake up earlier based on social stimuli (e.g., need to leave for work for an early meeting) which would temporarily change the timing of the sleep/wake cycle but may not necessarily alter the underlying circadian rhythm. The next day, when the individual no longer has to wake up early, they would revert to their previous sleep/wake cycl e.

PAGE 20

20 In summary, it appears that the circadian clock can be influenced by both photic (light) and nonphotic (behavioral state and social stimuli) sources. While light may directly influence circadian rhythms, it is believed that social stimuli and behavio ral state primarily influence circadian rhythms through their respective impact on light (Mistlberger et al., 2000) Considering that social stimuli can exert effects on the circadian system, and consequently impact sleep, it is important to examine the p articular social stimuli associated with circadian rhythm changes. Social Rhythms Social Zeitgebers Social nonphotic cues (zeitgebers other than light) have been shown to be important for entrainment in both animal species and humans (Aschoff et al., 1971) Research indicates that social entrainment begins soon after birth (Monk et al., 1990) Social stimuli can be important zeitgebers for maintaining daily rhythms (e.g., appropriate peaks and troughs) which are essential in a fluctuating daily rhythm ( Monk et al., 1990) For example, the social stimuli of leaving for work or school may help to energize or alert the individual to prepare them for starting their day, while watching television or reading in the evening may help to relax the individual in preparation for sleep. To elaborate on the earlier definition of social stimuli, important social rhythm markers for human beings can include meals, starting of work, leaving home, bedtime/waketime, exercise, watching television, and interacting with other individuals. Evidence for Social Zeitgebers As mentioned earlier, due to the pervasiveness of the rising and the setting of the sun, light is the most powerful zeitgeber for most species. Recently, however, researchers have begun to focus on another area of importance for regulating our lives the activity of other members of our species. It is hypothesized that circadian rhythms in some species may be able to be reset not

PAGE 21

21 by illumination, necessarily, but rather in response to interactions with other m embers of the species and to other social stimuli (Mistlberger & Skene, 2004) Evidence from animal studies suggests that animals do tend to coordinate their behavior to synchronize with other members of their social g roup (Mistlberger & Skene, 2004) Evidence for social zeitgebers can be found in studies of both animals and humans. Animal s tudies Animal studies have demonstrated the existence of social zeitgebers for animals in utero, postnatally, and in the timing of daily activities between and within species. It is suggested that the circadian timing of behavior may be particularly crucial when animals are young and are vulnerable to predation and adverse environmental conditions (Mistlberger & Skene, 2004) Studies of mammals have shown fetal circ adian rhythms to be entrained to the rhythms of the mother. While research has shown that light can penetrate the womb (Weaver & Reppert, 1989) it was shown that the entrainment of the fetus to maternal circadian rhythms occurs independent of environment al lightdark cycles. The SCN of the fetus has been found to be synchronized to maternal timing both in utero and postnatally (Reppert, 1995) Additionally, some rodent species entrain to the timing of the maternal daily activities to the point where the clock genes in the SCN are affected (Ohta, Honma, Abe, & Honma, 2002) The circadian clock genes that are involved in circadian rhythm oscillations were found to be entrained to the mothers rhythms. This phenomenon has been observed up until the weaning of the rodent (Shimoda, Hanada, Yamada, Takahashi, & Takahashi, 1986) The social zeitgeber of meal times has been demonstrated in the feeding activity of rabbits where circadian timing is entrained to the social zeitgeber of maternal behavior. Rabbit pups were able to anticipate the arrival of the doe for nursing (Jilge, 1995) Due to an inability to maintain warmth during the first days of life, rabbit pups tend to huddle in a group deep within

PAGE 22

22 the nesting material. After experiencing four to five r ounds of nursing, however, rabbit pups would crawl to the surface of the nest during a 60 minute period prior to the does arrival in order to prepare for the feeding (Jilge, 1995) To eliminate the confounding effects of light zeitgebers, the animals were kept in continuous light conditions during the study (Jilge, 1995) Finally, animal studies supporting the existence of social zeitgebers include studies of species which engage in daily coordinated activity and species that engage in daily desynchroni zed activity. For example, some species engage in activities in a coordinated manner (eating, drinking, and moving at the same time) in order to increase survival advantage due to protection from group size or to protect territory (Halle & Stenseth, 2000) Interestingly, desynchronized activity among species also points to a possible influence of social zeitgebers. For example, segregating activities of different group members to different times of the day can help to preserve resources. The existence of diurnal and nocturnal species is an example of temporal segregation which can provide a survival advantage in terms of conserving resources and sharing territory (Kronfeld Schor et al., 2001) In summary, animal studies on a variety of species suggest th at the activity of organisms can be entrained to social stimuli beyond the influence of photic stimuli. Human studies Several studies have examined the relationship between social zeitgebers and circadian rhythms in human beings. The timing of meals wa s found to be associated with the length of the subjective day (Aschoff, 1986) The gap between breakfast and lunch was associated with bedtime and wake time in free running conditions (Aschoff, 1986) Additionally, exercise has been associated with phas e shifts in circadian rhythm (Barger, Wright, Hughes, & Czeisler, 2004) The influence of meal times on circadian rhythms was demonstrated in a study examining the difference in hormonal levels of Muslim men during Ramadan fasting (Bogdan,

PAGE 23

23 Bouchareb, & Touitou, 2001) Cortisol and melatonin rhythms were measured in individuals before and during the observation of the Muslim holiday of Ramadan. During Ramadan, individuals refrain from eating, drinking, smoking, and sexual relations from sunrise to sunset (Bogdan et al., 2001) The authors found that changes in the timing of meals during Ramadan resulted in significant changes in cortisol and melatonin rhythms. McClintock (1971) described the phenomenon of synchronized menstrual cycles among women who li ve together. The author assessed the onset of the menstrual period among female college residents residing in a dormitory (aged 17 22 years). Greater synchronicity in menstrual cycles was observed among close friend groups and roommates compared to random pairings in the dormitory. To examine whether the synchronicity was a response to common exposure to light/dark periods, McClintock compared synchronicity of menstrual cycles between close friends and roommates, assuming that roommates would share mor e similar light/dark exposure patterns. The results (although not significant) showed a trend towards undermining a photoperiodic effect with roommates sharing less synchronicity ( p < 0.007) than close friends ( p < 0.003; McClintock, 1971) In summary, i t appears that social stimuli can entrain circadian rhythms both indirectly and directly as evidenced in both animal and human studies. Regularity of Social Zeitgebers In addition to studying the presence/absence of social cues and their effect on circadian systems, research has begun to examine the regularity of social cues (Monk, Kupfer, Frank, & Ritenour, 1991) Regularity or rhythmicity refers to an organisms tendency to respond to social stimuli in an orderly manner at fixed or uniform intervals (Mish 2000) Lifestyle regularity is likely determined by a variety of factors including biological, genetic, individual preference, and the circumstances in which the individuals are found (Monk, Houck, & Shear, 2006) The body of research on social rhythms has shown that regularity of social rhythms is associated with two

PAGE 24

24 of the proposed mediators investigated in the present study (arousal and affect) and the outcome variable of sleep. No studies, to our knowledge, have examined the relationship between social rhythmicity and light. Regularity and arousal Few studies have examined the association between regularity of social rhythms and arousal. One interesting investigation by Stetler, Dickerson, and Miller (2004) examined the relationship between social zeitgebers and cortisol levels in healthy controls and depressed individuals. The authors found that for the control group, the number of activities an individual engaged in predicted their cortisol levels with more activities associated with a decline in cortisol across the day. A similar relationship was not observed for depressed individuals. In terms of the regularity of daily activities, among the control group, more regularity in social activities was associated with a greater decline in cortisol throughout the day. A similar trend was observed in depressed individuals but was not found to be significant. Interestingly, when the authors combined total activities, regularity of activities, and the presence of another individual during the activit y in the model, only regular activities conducted when in the presence of another individual were found to be associated with a decline in cortisol activity throughout the day. When all three variables were entered into the model for the depressed individuals, there were no significant associations between the variables and cortisol levels. The authors concluded that individuals who are clinically depressed may not be as readily influenced by social activities (i.e., social activities may not be interpreted the same way or have as great an impact as for healthy controls; (Stetler, Dickerson, & Miller, 2004) Regularity and affect Depression By examining the association between social rhythmicity and affective disorders, it is possible to develop an understanding of how rhythmicity is related to affect. Research has linked

PAGE 25

25 social rhythms to depression, mania, and anxiety. In terms of depression, individuals diagnosed with depression have been found to exhibit irregular biological rhythms (e.g., irregular sleep wake cycles, melatonin, temperature, and cortisol rhythms; Howland & Thase, 1999) Additionally, circadian rhythms of those diagnosed with depression were found to be different in timing and amplitude than those not diagnosed with depression (Kup fer & Foster, 1972) In addition to irregular biological rhythms, research indicates that individuals diagnosed with depression may also experience irregular social rhythms. Brown and colleagues (1996) examined the stability of social rhythms following sp ousal bereavement among elderly participants. Individuals who were experiencing a major depression episode exhibited significantly lower levels of social rhythm regularity compared to non depressed controls (Brown et al., 1996) Similar findings emerged from the Prigerson and colleagues (1994) study examining the relationship between stressful life events, social rhythms, and levels of depressive symptomatology among elderly participants (Prigerson et al., 1994) Contrasting with findings illustrating a significant relationship between social rhythms and depression, Monk and colleagues (1990) found that there were no differences in social rhythmicity scores between individuals with unipolar depression and normal controls. A possible explanation for these findings, however, is that the study of social rhythms among patient populations may be limited by unstandardized medication regimes of those diagnosed with mood disorders (Monk et al., 1990) Considering that medications have been shown to affect circadi an rhythms, it is possible that the relationship between social rhythms and depressive episodes may be moderated by psychotropic medication use (Bendetti et al., 2001) Regularity and affect Manic episodes Research has also examined the relationship between social rhythmicity and bipolar disorder. Twice as many individuals who developed a manic episode experienced a social

PAGE 26

26 rhythm disruption of events in their life during an eight week period prior to the manic episode (MalkoffSchwartz et al., 1998) Interestingly, individuals who experienced a manic episode were more likely to have experienced a social rhythm disruption in the pre onset period while individuals who experienced a depressive episode were not (MalkoffSchwartz et al., 1998) Researchers hypothesized that the eight week study period may have been long enough to witness a manic episode triggered by the disruptive event but not long enough to observe the development of a depressive episode. A follow up study examined the relationship bet ween events disrupting social rhythms and manic and depressive episodes over 8 and 20 week study periods. The results replicated the previous findings in that participants experiencing bipolar manic episodes were more susceptible to the influence of socia l rhythm disruptive events compared to individuals diagnosed with depressive episodes (MalkoffSchwartz et al., 2000) Another study found that there was a significant difference in social rhythm regularity between individuals diagnosed with bipolar disorder and controls (Ashman et al., 1999) Individuals ranging in age from 29 61 who met the criteria for rapid cycling bipolar disorder were compared to matched control participants in terms of social rhythms (Ashman et al., 1999) Additionally, within su bjects analyses were performed to evaluate whether social rhythms varied with mood state (hypomania, depression, and euthymia). Social rhythm regularity was not found to vary systematically with mood state although power analyses indicated a larger sample size was needed to detect an effect. Interestingly, however, individuals diagnosed with bipolar disorder had significantly lower social rhythmicity scores compared to matched controls. Compared to control participants, individuals diagnosed with bipolar disorder less regularly had breakfast, started work, had lunch, exercised, had an evening snack, watched an evening news program, watched another television program, or took a nap.

PAGE 27

27 Regularity and affect Anxiety In addition to the association of social rhythmicity with depression and manic symptoms, researchers have examined the relationship between the rhythmicity of daily activities and anxiety (Shear et al., 1994) Results indicated that individuals who were diagnosed with an anxiety disorder had si gnificantly lower regularity of social rhythms compared to control participants. Consequently, regularity was shown to be associated with levels of anxiety. Regularity of social rhythms and sleep Research has investigated the relationship between social rhythms and sleep. In general, regularity of social rhythms has been associated with better sleep. In a study of college students (age ranging from 18 39 years) poor sleepers (as defined according to the Pittsburgh Sleep Quality Index [PSQI]) had lowe r frequency and less regular social rhythms compared to good sleepers (Carney, Edinger, Meyer, Lindman, & Istre, 2006) Additionally, greater variability in wake time, consuming a morning beverage, returning home for the last time, and bedtime were associ ated with more disturbed sleep (Carney et al., 2006) Monk, Reynolds, Buysee, DeGrazia, & Kupfer (2003) examined social rhythms in individuals ranging in age from 19 49. Social rhythmicity was negatively correlated with daytime dysfunction and sleep onset latency and positively correlated with sleep quality and sleep efficiency as measured by the PSQI. Individuals who had more irregular social rhythms were found to have worse sleep. Additionally, categorical analyses indicated that there were signific antly more poor sleepers in the irregular social rhythm group compared to the regular social rhythm group (Monk, Reynolds, Buysse, DeGrazia, & Kupfer, 2003) In a study examining the relationship between spousal bereavement, depression, and social rhyt hmicity among older adults, individuals who exhibited less social regularity experienced lower sleep quality (as rated on the PSQI) and had decreased REM percent as measured by one-

PAGE 28

28 night polysomnography assessment (Brown et al., 1996) Importantly, additi onal associations between social rhythmicity and sleep were eliminated once analyses controlled for depression among the participants. These results suggest that depression may be a mediator of the relationship between social rhythm stability and sleep (B rown et al., 1996) The present study will extend the study of the relationship between social rhythms and sleep by addressing the role of affect (in addition to light and arousal) as a possible mediator of this relationship. Theoretical Models Relation ship between Social Rhythms, Light, Affect, Arousal, and Sleep This study i nvestigate d the relationship between social rhythmicity and sleep among different age groups (young and older adults). The theoretical model presented by Mistleberger and Skene (2004) illustrates modes through which photic and nonphotic stimuli influence circadian systems. Additionally, more specific theoretical models have been proposed which focus on social zeitgebers (the construct of interest in the present study). Hence, the Mistleberger & Skene (2004) model is helpful for placing social zeitgebers in the context of all possible influences on circadian systems. Two additional models, the social zeitgeber and internal trigger hypothesis theories are introduced as more focuse d explanations for the association between social cues and biological rhythms. Finally, a model illustrating the aims of the present study is presented Theoretical models have been p roposed to explain the relationship between social rhythms and biologica l rhythms. First, the social zeitgeber theory is examined as a proposed model delineating the relationship between social cues and biological rhythms (Ehlers, Frank, & Kupfer, 1988) Second, the internal trigger hypothesis will be examined as a complemen tary model which further explains the relationship between social cues and biological rhythms (Grandin, Alloy, & Abramson, 2006)

PAGE 29

29 The social zeitgeber theory proposes that changes in mood result from disrupted social rhythms caused by life events (see Fig ure 2 2). Life events refer to a stressful occurrence that takes place close in time to the onset of the affective disorder (Grandin et al., 2006) Life events, in turn, can affect social zeitgebers. For instance, the individual can decide to stay home and not leave the house. Changes in social zeitgebers, in turn, can result in a disruption of social rhythms. A disruption in social rhythms could impact biological rhythms (e.g., sleepwake cycle, hormonal levels) which in turn could be manifested in somatic symptoms (e.g., decreased appetite, poor sleep). Finally, affective episodes are thought to result from the disruption of biological rhythms, resulting in somatic symptoms. The internal trigger hypothesis (see Figure 2 2 ) expands on the social zeit geber theory by proposing that social and biological rhythm disruptions in individuals are influenced by an underlying abnormality in the circadian pacemaker brought about through a physiological process such as a genetic mutation (Grandin et al., 2006) The authors propose that irregularity in the rhythms of the individual (either social or biological rhythms) is a result of a genetic mechanism, perhaps resulting from a genetic mutation. For example, an individual who is experiencing manic symptoms may h ave an underlying genetic mutation affecting the regularity of his/her circadian pacemaker which is manifested in irregular mood. The internal trigger hypothesis is an improvement over the social zeitgeber theory in that it proposes a mechanism underlying disruptions in social rhythms and biological rhythms. Unfortunately, there are two limitations of this model. First, the model does not propose a mechanism by which social rhythms influence biological rhythms. The authors suggest that abnormality in th e circadian pacemaker concurrently influences both social and biological rhythms. While the authors

PAGE 30

30 suggest that social rhythms can influence biological rhythms, they do not present an explanation for how social rhythms exert their influence. This weakne ss was addressed by the present study. A second drawback to the internal trigger hypothesis is that the mechanism proposed to underlie changes in social and biological rhythms is genetically based. The reliance on a genetic mechanism does not account for the psychosocial factors which can influence the regularity of rhythms. For example, an innovative treatment for bipolar disorder, Interpersonal and Social Rhythm Therapy, involves a psychosocial intervention for regularizing patients social rhythms (F rank et al., 1997) While the proponents of interpersonal and social rhythm therapy recognize that there are biological factors underlying bipolar disorder, they have found that the psychosocial intervention of regularizing social rhythms can be effective for regulating circadian rhythms and, consequently, improving sleep. The present study introduced the variables of light, arousal, and affect as possibly underlying the relationship between social and biological rhythms. These variables are amenable to change and could account for improvements in circadian rhythm disruption seen in psychosocial interventions such as interpersonal and social rhythm therapy. In su mmary, the present study focused on a specific component of the social zeitgeber and internal trigger hypothesis models. Of interest in the present study wa s the mechanism underlying the pathway from social rhythm disruption to biological rhythm disruption (see Figure 2 2 ). The current study proposed that mechanisms (light, arousal, and affect ) can account in part for the biological rhythm disruption resulting from social rhythm disruptions. The Mediated Model of Social Rhythmicity proposes that social zeitgebers influence biological rhythms (e.g., sleep) by regulating exposure to light, arous al, and affect (see Figure 2 3). Background research regarding the association between the mediators of light, arousal, and affect and the outcome variable of sleep is presented below.

PAGE 31

31 Sleep The Association Between Sleep and Light Research suggests that light can be associated with sleep in a number of ways. The timing of the sleep period can be affected by exposure to light. The sleep period was found to be delayed when individuals were exposed to light in evening (Czeisler et al., 1989) Exposure to light in the late night/early morning is associated with phase advances (the timing of the sleep period begins earlier in the day; (Duffy & Wright, 2005) Additionally, it appears that the human circadian system is responsive to light exposure throughout the day (Duffy & Wright, 2005) Light has also been found to exert effects on sleep through exposure to varying intensities of light. Initially, it was believed that bright light exposure (similar to outdoor natural light) was necessary to entrain circad ian rhythms (Wever, 1989) Recently, however, research suggests that exposure to much less intense levels of light (i.e., indoor lighting) can result in a phase shift (Duffy & Wright, 2005) The overall amount of light an individual is exposed to has be en associated with overall time spent asleep during the night, as well as, the amount of time spent awake during the night (Hood, Bruck, & Kennedy, 2004; Wallace Guy et al., 2002) Additionally, individuals who are exposed to insufficient light have been found to have poorer sleep (Campbell & Daston, 1991; Campbell, Dawson, & Anderson, 1993; Kohsaka et al., 1999; Kohsaka et al., 2000; Lack & Wright, 1993; Youngstedt, Kripke, Elliott, Baehr, & Sepulveda, 1999) The Association between Sleep and Arousal Ar ousal is investigated in the present study as one of the three mechanisms underlying the relationship between social rhythmicity and sleep. Arousal is operationalized as referring to both cognitive and somatic arousal. Arousal may be implicated in the re lationship between social rhythms and sleep, because in order for an individual to begin to fall asleep, the suppression of

PAGE 32

32 certain drives (appetitive, thirst, the need to void, desire to move, desire to socially interact) and behaviors must take place (As choff, 1965) Consequently, it is hypothesized that higher levels of arousal are detrimental to sleep. First, studies relating cognitive arousal to sleep will be reviewed followed by studies linking physiological arousal to sleep. Early research demonst rated that cognitive arousal played a role in the maintenance of insomnia. Lichstein and Rosenthal (1980) asked participants whether cognitive or somatic arousal was the main determinant of their insomnia. Participants credited cognitive arousal as being 10 times more likely to be the cause of insomnia relative to somatic arousal (Lichstein & Rosenthal, 1980) Similarly, Epsie, Brooks, and Lindsay (1989) suggested that cognitive arousal was the most significant factor in the maintenance of sleep complain ts. Using the Sleep Disturbance Questionnaire (SDQ), the authors investigated which of 12 items were most highly rated by patients with insomnia. The results showed that items that referred to intrusive or worrisome thoughts were more often endorsed by patients with insomnia (Espie, Brooks, & Lindsay, 1989) Additionally, pre sleep cognitive activity was found to be positively associated with sleep onset latency (Nicassio, Mendlowitz, Fussell, & Petras, 1985; Vanegeren, Haynes, Franzen, & Hamilton, 1983) Early experimental studies also investigated the relationship between cognitive arousal and sleep. Gross and Borkovec (1982) demonstrated that inducing ruminative worry interfered with sleep onset latency. Researchers informed good sleepers prior to a nap that they would have to give a speech upon awakening. The group that was told that they would have to give a speech took significantly longer to fall asleep compared to the control group (Gross & Borkovec, 1982) An innovative study by Haynes, Adams, & Franzen (1981) presented one group of participants with moderately difficulty arithmetic problems prior to falling asleep. Those who were given the

PAGE 33

33 arithmetic problems had significantly shorter sleep onset latency compared to the control group (Haynes, Adams, & Franzen, 1981) The findings have since been explained by the concept that the arithmetic effectively distracted the individuals in the experimental group from their cognitive arousal (Harvey, 2002) More recent studies have also supported a rel ationship between cognitive arousal and sleep. An experimental manipulation of the pre sleep stress of female good sleepers resulted in a positive correlation between stress related intrusive thoughts (assessed using a post sleep questionnaire) and sleep onset latency (assessed by polysomnography; Hall, Buysse, Reynolds, Kupfer, & Baum, 1994) Additionally, pre sleep arousal was associated with decreased delta wave intensity of the first sleep wave cycle (Hall, Buysse, Reynolds, Kupher, & Baum, 1994) De lta waves are associated with slow wave sleep, the deepest, most restorative sleep. Consequently, a decrease in delta waves would indicate a lighter sleep period. A positive correlation was found between pre sleep cognitive activity and sleep onset lat ency as measured by actigraphy (Wicklow & Espie, 2000) Pre sleep cognitive activity was monitored over a three night period using a voice activated recorder and a subscale of the PreSleep Arousal Scale (Nicassio et al., 1985) In a study assessing cogn itive arousal and sleep in young adults, individuals with insomnia were found to be significantly more likely to attend to worries and concerns, to try to solve problems in their lives, and to listen to noises in the house or outside compared to good sleepers (Harvey, 2000) Additionally, individuals with insomnia reported greater difficulty falling asleep because of thoughts and worries compared to the good sleeping group. In addition to research supporting the link between cognitive arousal and sleep, ph ysiological or somatic arousal has also been linked with poor sleep. Physiological arousal can

PAGE 34

34 involve elevated heart rate, body temperature, cortisol levels, and whole body metabolic rate (Cortoos, Verstraeten, & Cluydts, 2006) Vgontzas and colleagues (2001) demonstrated that the hypothalamic pituitary adrenal axis (HPA axis) is activated in individuals with insomnia. By comparing adults diagnosed with insomnia to healthy controls, the authors noted increased levels of 24hour adrenocorticotropic hormone (ACTH; involved in the release of cortisol) and cortisol in the individuals diagnosed with insomnia. Also, among the individuals diagnosed with insomnia, those with a high degree of objective sleep disturbance secreted a higher amount of cortisol compa red to those with a lower degree of sleep disturbance (Vgontzas et al., 2001) Similarly, Rodenbeck, Huether, Rther, and Hajak (2002) found that increased evening and nocturnal plasma cortisol levels correlated with impaired sleep in patients with severe primary insomnia (Rodenbeck, Huether, Ruther, & Hajak, 2002) Additionally, higher levels of metabolic rate were shown in individuals diagnosed with insomnia compared to healthy controls (Bonnet & Arand, 1995) The authors measured metabolic rate through out the day and night suggesting that insomnia patients may not only be aroused during the sleep onset period but also during the daytime. The same authors investigated the effects of various manipulations of physical movements on physiological arousal (B onnet & Arand, 1999) Wakefulness was enhanced the most after standing and doing knee bends, less after standing, less after sitting up, and least after the participants were spoken to by the experimenter. These findings illustrate the impact of physical movements on measures of arousal. Finally, using stress responsivity and free coritsol levels over a 24 hour period, Prinz and colleagues (2000) examined whether longer lasting glucococorticoid increase was associated with sleep disturbance in a sample of older adults. When placed under a mild stress condition

PAGE 35

35 (insertion of an indwelling intravenous catheter placement), individuals with higher levels of free cortisol had impaired sleep as represented by lower sleep efficiency, more time spent in sleep s tage one, and greater EEG beta activity during NREM sleep (Prinz, Bailey, & Woods, 2000) In summary, while numerous studies have supported the relationship between cognitive arousal and poor sleep, it is difficult to separate the effects of both cognitiv e and somatic arousal. For example, while an individual is cognitively aroused, they may simultaneously be physiologically aroused and vice versa. Additionally, each component of arousal may work to increase the other. In order to encompass both of these possibly interacting constructs, the present study will assess both components of arousal (cognitive and somatic). Theoretical model Linking sleep and arousal Several models of insomnia have been proposed which implicate the role of arousal in maintai ning poor sleep. The model that is most relevant to the present study is the cognitive model of insomnia proposed by Harvey (2000). While labeled the cognitive model of insomnia, Harveys model addresses both physiological and cognitive aspects of arous al. Harvey proposes that pre sleep cognitive arousal (occurring both during the day and just before bed) will result in impaired sleep (Harvey, 2002) The specific steps of the model are illustrated in Figure 2 4. First, the individual experiences negat ively toned cognitive activity (e.g., worry). Second, the negatively toned cognitive activity triggers autonomic arousal resulting from activation of the sympathetic nervous system. Third, based on the increased arousal, the individual will experience se lective attention and monitoring of threat related cues. These cues consist of external and internal cues that indicate that the individual is not sleeping. For example, the individual may become aware of external noises in the environment or become awar e of muscle tension. The threat related cues then feed back into the negatively toned cognitive activity. The forth stage of the model consists of a distorted

PAGE 36

36 perception of the perceived deficit in sleep and daytime performance. The distorted perception s then feed into the negatively toned cognitive activity (e.g., Im going to be exhausted tomorrow). Harvey also describes two additional processes that contribute to the maintenance of insomnia. Erroneous beliefs about sleep (e.g., I cannot survive on less than eight hours of sleep) can serve to fuel negatively toned cognitive activity. Also, an individual can engage in safety behaviors (e.g., calling in sick to work) based on negative cognitions. Safety behaviors can exacerbate negative cognitions (e.g., Now that Ive missed a day at work I will be behind in everything) and beliefs (by avoiding work the individual does not get to test their erroneous belief that they cannot survive on less than eight hours of sleep). Harveys cognitive model of i nsomnia (Harvey, 2002) is particularly relevant to the present study in that it provides a mechanism to explain the role of arousal as mediating the relationship between social rhythmicity and sleep. Based on the above mentioned literature, it is apparent that there is an association between arousal and sleep. Harveys model delineates the specific pathways by which arousal impacts sl eep. The present study proposed that the pathways eluded to in Harveys model also represent the method by which arousal im pacts sleep. The in novation of the current study wa s the inclusion of additional mediators (light and affect) of the relationship between social rhythmicity and sleep. The Association between Sleep and Affect A large body of literature has examined the relationship between affect and sleep. Sleep difficulties are often comorbid with psychological problems. In fact, according to the Diagnostic and Statistical Manual of Mental Disorders (DSM IV; American Psychiatric Association, 2000) sleep disruption is symptomatic of all mood disorders including Major Depressive Disorder, Dysthymic Disorder, Bipolar Disorder, and Cyclothymic Disorder. Depending on the disorder, individuals could experience insomnia or hypersomnia.

PAGE 37

37 In terms of research examining the association between affect and sleep, a small number of studies have examined the association between daily levels of affect and sleep. Berry and Webb (1985) examined the association between mood (measured by the Lorr Mood Test; Lorr, Daston, & Smith, 1967) and sleep (EEG data) in women aged 56 66 years. Affect was found to be significantly associated with sleep, such that positive affect (cheerful and energetic moods) was associated with decreased sleep onset latency and negative affect (anxiety) was associated with decreased sleep efficiency and greater sleep onset latency (Berry & Webb, 1985a) Another study by the same authors found similar results with positive affect associated with increased sleep efficiency and total sleep time and negative aff ect associated with increased wake time after sleep onset (Berry & Webb, 1985b) Affect was also found to be associated with sleep in a study by Totterdell and colleagues (1994). Individuals that were higher in positive affect experienced an earlier onset of the sleep period (Totterdell, Reynolds, Parkinson, & Briner, 1994) Finally, a recent study by McCrae and colleagues (2008) demonstrated an association between affect and sleep over an extended data collection period (14 days). On days when individua ls reported higher levels of positive affect they also reported higher levels of sleep quality. In terms of the overall sample, individuals who reported lower positive affect reported greater overall total wake time For negative affect, on days when individuals reported higher levels of negative affect they were more likely to report below average sleep quality. In terms of the overall sample, between groups analyses showed that on average, higher negative affect was associated with greater total wake t ime and lower sleep quality (McCrae et al., 2008) In summary, both affective disorders and negative affect appear to be associated with impaired sleep while positive affect is associated with better sleep outcomes.

PAGE 38

38 Age The proposed Mediated Model of Soc i al Rhythmicity was examined in two age groups younger and older adults. It was necessary to investigate age differences in order to examine age as a moderating effect on the variables of interest. Prior research is presented next that examines the relationship between age and: 1) the predictor variable of social rhythmicity, 2) the outcome variable of sleep, and 3) the mediator variables of light, arousal, and affect. Age and Social Rhythms In general, it appears that the regularity of social rhythms i ncreases as individuals age (Monk, Petrie, Hayes, & Kupfer, 1994; Monk et al., 1997) Healthy older adults in their 8th and 9thMo nk and colleagues (2006) examined seniors who fit into one of three groups: 1) bereaved or caring for a spouse at home, 2) those diagnosed with insomnia, and 3) healthy, well functioning older adults. Similar to previous research, across all groups, socia l rhythm regularity appeared to increase with age. Interestingly, however, the results showed an acceleration of the age versus social rhythm regularity relationship in the late middle aged and senior years. The authors suggested that the relationship be tween lifestyle regularity and age may be better represented by a power function of time with an asymptote at about age 75 rather than a linear function between age 55 to age 90 years (Monk, Buysse, et al., 2006) decade have been shown to have social rhythm regularity scores that are, on average, one standard deviation higher than adults in their 30s (Monk et al., 1997; Monk, Reynolds, Machen, & Kupfer, 1992) Monk and colleagues (1992, 1997) proposed that increased regularity with age may be an adaptive response to agerelated changes in sleep and circadian rhythms among older adults.

PAGE 39

39 Age and Sleep Difficulty sleeping is a p roblem faced by the majority of Americans across different age groups. The National Sleep Foundations 2008 Sleep in America poll found that 65% of working Americans, age 18 and above, report experiencing a sleep problem at least a few nights a week (Natio nal Sleep Foundation, 2008) Similarly, an estimated 57% of adults age 60 and older complain of difficulty sleeping (Kryger, Monjan, Bliwise, & Ancoli Israel, 2004) There are significant negative outcomes associated with difficulty sleeping for older ad ults including excessive daytime sleepiness, attention and memory problems, depressed mood, falls, overuse of hypnotic drugs, and lower quality of life (Kryger et al., 2004) One type of sleeping disorder, insomnia, is characterized by difficulty falling asleep, staying asleep, waking early, and experiencing nonrestorative sleep (American Psychiatric Association, 2000) Sleep undergoes various physiological changes across the lifespan. The most prominent changes are reductions in slow wave sleep (SWS) and Rapid Eye Movement (REM) sleep. Slow wave sleep consists of stage three and four nonrapid eye movement (nonREM) sleep. SWS is described as restorative sleep in that the individual experiences a slowing of brain activity; enabling repairs and rest for brain areas (Horne, 1988) REM sleep involves low voltage brain waves and dreaming in addition to the movement of the eye muscles and involuntary muscle jerks. REM sleep is important for functioning in that it helps to develop synaptic connections during early life and is needed to maintain these connections in late life (Shaffery, Sinton, Bissette, Roffwarg, & Marks, 2002) Additionally, older adults can experience a phase advance in their sleep wake rhythm whereupon they become sleepier earlier in the day and have earlier wake times. Despite underlying physiological changes, however, recent research has suggested that agerelated differences in sleep are not as drastic as previously believed (Ohayon, Carskadon, Guilleminault, & Vitiello, 2004) Sleepi ng problems in older adults may not be due

PAGE 40

40 to age related changes in sleep per se but rather due to a combination of factors associated with aging (Vaz Fragoso & Gill, 2007; Vitiello, Moe, & Prinz, 2002) For example, a metaanalysis of studies examining objectively measured sleep (PSG or actigraphy) in healthy community dwelling individuals across the lifespan found that the majority of changes seen in sleep patterns across the life span occur between early adulthood and age 60, with minimal declines from age 60 to age 102 (Ohayon et al., 2004) It has been argued that sleep problems in nonhealthy older adults are the result of factors associated with aging such as loss of physical function, reduction in health status, and higher levels of primary sleep disorders (e.g., sleepdisordered breathing; Vaz Fragoso & Gill, 2007) Consequently, sleep disruption in older adulthood that is likely multifactorial in origin (Vaz Fragoso & Gill, 2007) In summary, it appears that difficulty sleeping remains a pr oblem for the majority of both younger and older adults, but the origin of sleeping difficulties may differ by age groups. Although there are arerelated changes in the physiology underlying sleep, it is likely that sleep disruption in older age is due to multiple factors associated with aging. Due to differences in perhaps the origin and expression of sleeping difficulties across age groups, the proposed model was examined separately for each age group. Age in Relation to Mediators of Light, Arousal, and Affect It wa s necessary to examine the possible mediating effects of light, arousal, and affect within two different age groups as age could interact with the variables of interest in numerous ways. For example, we know that older adults typically are m ore regular in their social rhythms but experience changes in their sleep (due to a combination of factors). Consequently, while increased regularity could serve as an adaptive function for older adults, we may not see better sleep outcomes in the older a dults compared to the younger adults due underlying age related

PAGE 41

41 factors (e.g. increased loss of health or physical function) impacting sleep. Additionally, we know that light can differentially affect both younger and older adults due to a reduction in pupil diameter as well as a yellowing of the lens with age which can attenuate the amount of light reaching the retina (Hughes & Neer, 1981; Teresi, Lawton, Ory, & Holmes, 1994) Consequently, if light wa s found to mediate the relationship between social rh ythmicity and sleep, it may do so to a lesser extent in the older adult sample due to factors affecting exposure to light. In terms of affect, we know that older adults typically report higher levels of positive affect and wellbeing compared to younger adults (Carstensen, Pasupathi, Mayr, & Nesselroade, 2000; Charles, Reynolds, & Gatz, 2001) Therefore there may be an interaction effect of age where older adults have higher levels of positive affect and, consequently, affect differentially mediates the r elationship between social rhythmicity and sleep for the two age groups. In fact, a recent study (McCrae et al., 2008) found a relationship between affect and sleep in older adults where poorer sleep was reported on days when individuals had higher negative affect and lower positive affect. Therefore, poorer affect in older adults may be associated with poorer sleep in the present study. Age differences in levels of arousal have not (to our knowledge) been previously reported in the literature. In summa ry, there are a number of different ways in which age could interact with the variables of interest, resulting in variations to the extent that light, arousal, and affect mediate the relationship between social rhythmicity and sleep. By examining the proposed model in two separate age groups, it is possible to differentiate the possible moderating effects of age on the variables of interest. Summary Research indicates that the circadian system plays a vital role in regulating many human functions, including the sleep/wake cycle. Additionally, as social stimuli can exert effects on the

PAGE 42

42 circadian system, and consequently impact sleep, it is warranted to examine the social stimuli associated with circadian rhythm changes. One aspect of social stimuli, regu larity of daily activities has been linked to sleep and age with greater rhythmicity in social activities associated with better sleep and older age. While the regularity of social stimuli has been linked to the sleep/wake cycle, the mechanisms underlyi ng this relationship have not been studied. Based on the negative association between arousal and sleep, it was hypothesized that arousal may be a factor mediating the relationship between social rhythmicity and sleep. To our knowledge, only one study (S tetler et al., 2004) has examined the relationship between social rhythmicity and arousal. Mandlers theory of mind and emotion (1975) states that interruption of ones plans and organized behavior can generate arousal which can be interpreted as various emotions (Mandler, 1975) Consequently, it is hypothesized that greater social rhythmicity will be associated with decreased arousal and better sleep. In terms of affect, we know that irregular social rhythms have been associated with affective disorders such a s depression (Brown et al., 1996; Prigerson et al., 1994) mania (Ashman et al., 1999; Malkoff Schwartz et al., 1998; Malkoff Schwartz et al., 2000) and anxiety (Shear et al., 1994) Additionally, literature on motivation indicates that more regular, habitual behavior is associated with less emotional intensity. Wood, Quinn, & Kashy (2002) found that individuals who engaged in habitual behaviors experienced less intense emotions compared to when they engaged in nonhabitual behaviors (Wood, Quinn, & Kashy, 2002) Frijdas laws of emotion state that continued pleasures wear off; continued hardships lose their poignancy (Frijda, 1988, p. 353). Consequently, it was hypothesized that higher levels of social rhythm regularity will be associated with de creased affect and subsequently better sleep. Finally, as light is the dominant zeitgeber affecting our circadian rhythms, it is

PAGE 43

43 essential to examine the role of light as mediating the relationship between social rhythmicity and sleep. Based on the possi ble differential effects of age on the relationships of interest, it wa s warranted to examine whether the mediation of the social rhythmicity and sleep relationship is similar for both younger and older adults.

PAGE 44

44 Social stimuli Behavioral state (open eyes) Circadian clock(s) Figure 2 1. Mistleberger & Skenes (2004) theoretical model of temporal influences on biological rhythms.

PAGE 45

45 Figure 2 2. Social Zeitgeber Theory and Internal Trigger Hypothesis. Blue indicates addition of Internal Trigger hypothesis to the Social Zeitgeber Theory. R ed denotes area of investigation in present study. Life Events Changes in Social Zeitgebers Affective Episodes Social Rhythm Disruption Biological Rhythm Disruption Other Somatic Symptoms Abnormality of Circadian Pacemaker Area of investigation in present study

PAGE 46

46 Figure 2 3. Mediated Model of Social Rhythmicity examined for both age groups in the present study. Social Rhythmicity Arousal Sleep Affect Li ght

PAGE 47

47 Figure 2 4. Cognitive Model of Insomni a (Harvey, 2000). Excessive negatively toned cognitive activity Arousal and distress Selective attention and monitoring Safety behaviors Beliefs Distorted perception of deficit In bed/during the day Real deficit

PAGE 48

48 CHAPTER 2 STATEMENT OF THE PRO BLEM Purpose of the Study Previous research has demonstrated a relationship between social rhythmicity and sleep, with decreased regularity associated with poorer sleep. Little research, however, has examin ed mechanisms that may underlie the relationship between social rhythmicity and sleep. Furthermore, while light is assumed to be the dominant zeitgeber affecting our biological rhythms, previous research has not attempted to examine the relationship betwe en light and social rhythmicity and sleep. Consequently, previous studies have presented an incomplete picture of the relationship between social regularity and sleep by not accounting for the dominant zeitgeber light. The present study addresses this c oncern by investigating the role of light, arousal, and affect as mediators of the relationship between social rhythmicity and sleep. Additionally, the present study investigates the role of age as moderating the relationships between social rhythmicity, light, arousal, and affect, and the outcome variable of sleep. Innovations of the present study include the examination of possible mechanisms underlying the relationship between social rhythmicity and sleep. Also, the present study will examine an often overlooked aspect of functioning the variations that occur within individuals. While between person differences in behavior are typically examined, the within person variations, or intraindividual variability, can provide useful information about the fluctuations of behaviors within and individual. Additionally, the study will use an innovative statistical approach (multilevel modeling) to examine daily fluctuations in the constructs of interest. In addition to providing new information on variabilit y in these constructs, the present study will examine the relationship between variability in these constructs and sleep. For example, we know that more light exposure is associated with better sleep. What has not been studied,

PAGE 49

49 however, is the impact of i rregularity in light exposure in relationship to their sleep. Consequently, the present study s ought to answer the following questions: 1. To what extent do individuals vary within themselves in terms of their social rhythmicity, light, arousal affect, and sleep and how does this compare to levels of between persons variation? 2. Is variability in social rhythmicity, light, arousal, and affect associated with better or worse sleep? 3. Are social rhythmicity, light, arousal, and affect coupled with sleep on a da y to day basis and how does this compare to meanlevel relationships? 4. To what extent do light, arousal, and affect mediate the relationship between social rhythmicity and sleep? Each of these goals was examined in both younger and older adults. Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep Fluctuate Daily both Within and Between Persons Analysis The purpose of aim one was to establish the extent to which individuals vary in their social rhythmicity, light, ar ousal, affect, and sleep. Intraindividual variability (IIV) refers to a fluctuation/change around an individuals mean (Nesselroade, 1991) IIV can be examined through repeated measurement of behavior on a moment to moment, daily, weekly, monthly, or yea rly basis. The present study examined daily fluctuations in the variables of interest over 14 consecutive days. Fourteen days of data collection were warranted in order to capture the variability inherent in the variables of interest. Additionally, as i t is possible that participants varied in their weekday and weekend behaviors 14 days of data collection allowed for analyses of data on both the weekdays and weekends. Meanbased analyses (e.g., ANOVA or regressions) consider variability within the indi vidual to be error or noise. However, it is the error or within person variability that is particularly of interest for highly variable behaviors. In fact, the examination of within person variability has been described as absolutely essential

PAGE 50

50 f or accurately capturing behaviors within the individual (Nesselroade, 2002) Furthermore, examining intraindividual variability is particularly pertinent when analyzing behaviors that are highly variable (e.g., sleep). Fluctuations or variability in the variables were analyzed using the intraclass correlation coefficient (ICC). The ICC consists of the between persons variance/total variance (between and within persons). The ICC is calculated by running a null multilevel model with the variables of interest. The amount of within person variance can be calculated (by subtracting the ICC from one) to determine if further multilevel analyses examining effects at the individual level are warranted. Hypothesis It was hypothesized that sufficient within person variability will be detected to warrant further multilevel analyses. A previous study examining variability among sleep variables demonstrated that within person variability accounts for at least fifty percent of the between person variability in napping and sleep variables in older adults (Dautovich, McCrae, Rowe, & Dzierzewski, 2008) Consequently, it is hypothesized that while within person variability may be less than between person variability, the amount of within person variability in the v ariables of interest is hypothesized to be at least fifty percent of the between person variability. Aim 2: Identify how Variability in Social R h ythmicity, Light, Arousal, and A ffect is R elated t o S leep Analysis The purpose of aim two wa s to determine the extent to which within person variability in social rhythmicity, light, arousal, and affect is related to sleep The intraindividual standard deviation (ISD) was calculated for each individual for the social rhythmicity, light, arousal, and affect vari ables. The ISD was calculated based on the distribution of values obtained across the

PAGE 51

51 14 days of repeated measurement for each individual. The ISD describes the extent to which an individuals scores tend to vary around their mean score. The data was de trended in order to control for the effects of time (e.g., time was residualized out of the variables of interest). B ivariate correlations were then conducted between the ISD of the predictor variables (social rhythmicity, light, arousal, and affect) and sleep. Given the exploratory nature of this aim, all predictor and sleep variables were included in the analyses. Significant correlations b etween the ISD s and sleep suggested that within person variability in social rhythmicity, light, arousal and aff ect is associated with sleep. Aim 2 also allow ed for examination of habitual social behaviors within the individual in relation to sleep. For example, a person may be regular or irregular in their social activities but that pattern may be habitual for the m (see Figure 5). For example, an individual could habitually be regular in social rhythms and days of irregular social rhythms would be atypical or variable for them. Conversely, someone could habitually be irregular in their social rhythms and days w here they are more regular in their activities would be considered atypical or variable for them. Cons equently, the level 1 or within person analysis examine d the extent to which variability within the individuals social rhythmicity was related to sleep. In addition to social rhythmicity, aim 2 also provided information on fluctuations in the remaining predictor variables (light, arousal, affect) in relation to sleep. Hypothesis In general, it wa s proposed that higher levels of variability in the predi ctor variables w ould be associated with poorer sleep. In terms of social rhythmicity, one study (Monk et al., 1991) found that individuals diagnosed with depression were more likely to have greater intraindividual variability on the SRM 17 compared to hea lthy controls. This study provides support for the association of variability in social rhythmicity with depressed mood but does not

PAGE 52

52 provide evidence for the association between variability and sleep. Examining variability in a broader research context s uggests that variability in physiological functioning is associated with poorer overall functioning (Britton, 1997; Pagani, 1999) Additionally, individuals who are diagnosed with insomnia typically have highly variable sleep patterns (Edinger, Marsh, McC all, Erwin, & Lininger, 1991) Therefore, although this aim wa s exploratory to a degree, it wa s hypothesized that in general, higher levels of variability in the predictor variables w ould be associated with poorer sleep. Aim 3 : Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light, Arousal, and Affect are Coupled with Daily Fluctuations in Sleep Analysis The goal of aim two was to examine the ability of social rhythmicity, light, arousal, and affect to predict sleep on a daily level and mean level A multilevel (MLM) approach was used to examine how daily data for the predictor variables predicts sleep. MLM, also known as hierarchical linear modeling or mixed effects modeling, is an alternative to univariate or multivariate analys is of repeated measures. MLM allows for the analyses of change in individual cases over an extended time period (Tabachnick & Fidell, 2007) MLM is an extension of the general linear model and does not require observations to be independent. Additionall y, MLM allows all available data to be included in the analyses as it assumes random missing data and consequently does not exclude a case due to one missing data point (Bryk & Raudenbush, 1992) MLM is particularly suited to the analysis of daily data due to the hierarchical structure of the data, with daily observations nested within the same participant (Singer, Davidson, Graham, & Davidson, 1998; Singer, Fuller, Keiley, & Wolf, 1998; Singer & Willett, 2003)

PAGE 53

53 Due to the hierarchical nature of the data (14 consecutive days nested within 50 participants), it was possible to examine to what extent the predictor variables were related to sleep at both level 1 and level 2. Level 1 analyses examined the day to day relationship between the variables for each individual. For example, On days when an individual is more or less regular in their social activities, do they have more or less SOL. (see Figure 3 1 for example). Level 1 analyses were computed using person centered variables which are the intraindi vidual fluctuations around the individuals own mean. Level 2 analyses examined the overall relationship between mean level variables. For example, Are higher or lower levels of overall regularity associated with more or less SOL overall. In order to compute level 2 analyses, mean levels were calculated for the variables of interest. Fixed and random effects were also estimated in addition to examining level 1 and level 2 relationships. Fixed effects are the effects that hold for all persons. Fixed effects can be measured at both level 1 and level 2. For example, fixed effects at level 1 refer to the relationship between the predictors and sleep, on a day to day level, that hold for all persons, on all occasions. Fixed effects at level 2 refer to the overall, mean level relationships, between the predictors and sleep for all persons on all occasions. Random effects, on the other hand, test whether there are significant individual differences in the observed fixed effects. A random effect of the f ixed effect of social rhythmicity predicting sleep at level 1 would indicate that although social rhythmicity predicts sleep on a day to day basis, there is significant within person variation in the magnitude of this relationship. Hypothesis The hypothe sis for the level 1 analysis was exploratory in that few studies have examined variability in the predictor variables within the individual and no studies (to our knowledge) have examined the daily relationships between the predictor variables and sleep. C onsequently, the

PAGE 54

54 tentative hypothesis for the level 1 analysis was that higher levels of variability in the predictor variables will be coupled with poorer sleep outcomes. The hypothesis for the level 2 analysis was that days with higher levels of social rhythmicity, light, and positive affect would be associated with better sleep and days with lower levels of arousal and negative affect would be associated with better sleep. Aim 4 : Identify to What Extent Light, Arousal, and Affect Mediate the Relationsh ip Between Social Rhythmicity and Sleep Analysis In order to determine the extent to which arousal mediates the relationship between social rhythmicity and sleep, Baron and Kennys (1986) four step method of testing mediation was examined using multiple regression analyses. The four steps of this process are as follows (Baron & Kenny, 1986) : Determine whether the predictor variables are correlated with sleep. This step will establish whether or not there is an effect between the predictor variables and sleep to be mediated. Determine whether social rhythmicity is correlated with the mediator variables of light, arousal, and affect. This step will investigate the mediators as outcome variables. Determine whether the mediators (light, arousal, and affec t) are related to sleep. Both social rhythmicity and the mediators are included in this analysis in order to control for social rhythmicity to establish the relationship between the mediator variables and sleep. In order to determine if light, arousal, and affect mediate the social rhythmicity and sleep relationship, the strength of the relationship between social rhythmicity and sleep should be significantly reduced when controlling for the mediator variables. A multiple mediator model was to be tested as we are interested in whether multiple variables (light, arousal, and affect) mediate the relationship between social rhythmicity and sleep.

PAGE 55

55 Hypothesis Little research has examined possible mediators of the relationship between social rhythmicity and s l eep. Consequently, aim four wa s primar ily exploratory in nature. It was hypothesized that arousal would partially mediate the relationship between social rhythmicity and sleep over and above the effe ct of light. Additionally, it wa s hypothesized that a ffect would also mediate the relationship between social rhythmicity and sleep over and above the effects of light and arousal. In other words, the relationship between social rhythmicity and sleep w ould vary depending on the levels of light, arousal, and affect.

PAGE 56

56 Figure 3 1. Example of hypothesized daily c oupling of s ocial r hythmicity and s leep. Y axis represents amount of social rhythmicity and sleep. Days in Study

PAGE 57

57 CHAPTER 3 METHODS The present study examined the relationships between s ocial rhythmicity, light, arousal, affect, and sleep in older and younger adults. S tudy materials were completed via the internet. Prior to beginning participation in the study, participants complete d the informed consent, demographics and health questionnaire, the BDI II, and the STAI Y1. After completing the pre study assessment, participants completed the social rhythmicity, light, arousal, affect, and sleep questionnaires on a daily basis for fourteen days. Materials Demographics and Health Survey T his questionnaire (see Appendix 1), collects information on demographics, sleep disorder symptoms, physical health, and mental health (Lichstein, Durrence, Taylor, Bush, & Riedel, 2003) The Demographics and Health Survey was administered prior to the st art of the 14 days of data collection. Health conditions that are assessed include: heart attack, cancer, AIDS, hypertension, neurological disorder (e.g., Parkinsons, seizures), breathing disorder (asthma, emphysema), urinary problems (kidney disease, pr ostate problems), diabetes, pain (arthritis, back pain, migraines), and gastrointestinal disorders (stomach, irritable bowels, ulcers, and gastric reflux). Self report sleep questions on the survey ask whether the participant had a sleep problem and if t hey or a bed partner noticed heavy snoring, difficulty breathing or gasping for breath, frequent leg jerks, restlessness before sleep onset, sleep attacks during the day, or paralysis at sleep onset. If they answer yes to any of these problems they are as ked to describe the problem and indicate how often and how long the symptoms have occurred. Additionally, the participants are asked for the names of all medications they are currently taking and about any mental health

PAGE 58

58 concerns. Finally, the participant s are asked to report whether they or someone they care about has experienced a stressful/disruptive life event (e.g., divorce, birth, marriage, move, death, illness) over the past year, month, or week. Beck Depression Inventory Second Edition Depression was measured using the Beck Depression Inventory Second Edition (BDI II; Beck, 1996) This is a 21 item measure with a scale ranging from 0 3 measuring the severity of depressive symptoms (3 being the most severe; see Appendix 1). Scores range from 0 63. Scores within the 0 13 range indicate none or minimal depression, 14 to 19 indicate mild depression, 20 to 28 indicate moderate depression, and 29 to 63 indicate severe depression. Participants are asked to respond to the questions based on the previous two weeks. The BDI II has demonstrated sufficient internal consistency reliability (.90) and concurrent validity (.69 .76) within psychiatric patients, college students, and community dwelling adults (Storch, Roberti, & Roth, 2004) The BDI II was completed at the start of the study. State Trait Anxiety Inventory, Form Y1 (STAI Y1 ) Anxiety will be measured using the State Trait Anxiety Inventory Form Y1 (STAI Y1; Spielberger, 1983) This is a 20item measure with a 4 point scale indicating how often the statement is true (see Appendix 1). Scores range from 20 80 with higher scores indicating greater maladjustment. The scale assesses how often the individual experienced apprehension, tension, nervousness, and worry during the previous two wee ks. The STAI Y1 was completed at the start of the study. The STAI Y1 demonstrates testretest reliability exceeding 0.70 for younger adults and reliably distinguishes patient and normal groups (Spielberger, 1983) The STAI Y1 was found to be a suitable in strument for identifying mental disorders in geriatric individuals (Kvaal, Ulstein, Nordhus, & Engedal, 2005) and demonstrated adequate test retest reliability for older adults

PAGE 59

59 (Himmelfarb & Murrell, 1983) Additionally, the scale is a sensitive indicator of transitory anxiety and was selected for its long standing reliability and validity of state anxiety in Caucasian, African American, and Hispanic populations (Novy, Nelson, Goddwin, & Rowzee, 1993 ) Social Rhythm Metric The Social Rhythm Metric (SRM 17; Appendix 2) was completed each morning by the participants. The SRM 17 is a scale designed to assess the extent to which an individual is regular or irregular on a daily basis in terms of the timing of daily events (Monk et al., 2003) The scale was o riginally developed by Monk and colleagues (1990). Events were selected for inclusion in the original measure if they were involved in the expression of a biological or emotional drive and consequently likely to be tied to circadian rhythms (Monk et al., 1990) As a result the scale contains information about the timing of meals, exercise, going outside, napping, and bed and wake times. There are two spaces (Activity A and Activity B) where the individual can write in an activity that is specific to the m. Additionally, as interpersonal interactions have been found to have strong entrainment potential (Klein & Wegmann, 1975) the scale also assesses whether other individuals were present for each activity. Once a day, in the morning, individuals complet e the scale retrospectively for the previous day. Individuals record the clock time that the activity was completed and then indicate if they were alone, and if they were not alone, who else was present. Based on the daily rhythm scores, a total score can be calculated. First, a habitual time is calculated for each activity using an outlier elimination algorithm1 1 Since an event could occur at 9:00, 9:10, 9:30 and 12:02, to calculate a mean habitual time would not be an accurate representation of the actual time of the event. Consequently, Monk et al., (1990), developed an algorithm for eliminating outliers that could unduly influence the calculatio n of a habitual time. First the clock time is decimalized (e.g. 9:45 becomes 9.75). Then 12 hours are added to all p.m. times and 24 hours are added to all times Next, the number of times in

PAGE 60

60 a week that an event occurs within 45 minutes of the habitual time for the event is summed. The final step is to add up the counts for all the items on the questionnaire and divide by the total number of items (e.g., 17). This will yield a score between 0 and 7 with higher scores indicating greater regularity. Monk et al. (1990) found that there was no signifi cant difference between diaries completed during the day (concurrent with the activity) and those completed retrospectively at night in terms of the means and standard deviations. This finding suggests that there should be no significant losses of data by having the participants complete the SRM 17 retrospectively. The SRM 17 showed good test retest reliability for older adults (rho = 0.44, N = 49, p<.001; Monk et al., 1990) and in a sample of adults (mean age = 33 years; r= 0.48, p < .002; Monk et al., 1994) Additionally, the SRM 17 demonstrated good construct validity in that during the validation study, adult participants who had a major life disruption (e.g., took a vacation) experienced considerably lower SRM scores than when they returned to their normal routine (Monk et al., 1990) The SRM 17 demonstrated adequate criterion validity for adults when the subjective reporting of bedtime and out of bed time was compared to objective actigraphy estimates (r = 0.41 and r = 0.96 respectively; Monk et al., 1994) From the SRM 17, both a weekly score as well as a daily rhythmicity score can be generated. Light Exposure To our knowledge, there are no previously developed instruments designed to measure exposure to light using a self report format. The present study used a measure designed for this study titled the Light Exposure Scale ( LES; see Appendix 2) which was completed each occurring between midnight and 5 a.m. The mean and standard deviation are then calculated f or each activity. Then if the daily time for each activity is more than 1.5 standard deviations from the mean, that daily time is eliminated from the analysis. The mean and standard deviation are then recalculated once outliers have been removed, resulti ng in a habitual time for each activity.

PAGE 61

61 morning by the participants. Previous research examining the relationship between light and sleep has assessed the intens ity, duration, and time of day of exposure to light. The LES measures these variables by asking participants to record the intensity of the light exposure (ranging from darkness/moonlight to daylight with sun shining scored as 1 5), the duration of expos ure (participants will estimate their exposure in one hour intervals), and the time of day of exposure (ratings using a 24hour clock). Consequently, the LES has face validity in terms of the typical components of light commonly assessed. Four variables are generated from the LES: AM light (the first hour of the day exposed to cloudy or bright light [> 1000 lux]), PM light (the last hour of the day exposed to cloudy or bright light [> 1000 lux]), light duration (the number of hours exposed to cloudy or br ight light over a 24hour period), and light intensity (the total amount of light [ranging from 15] multiplied by the number of hours). In order to determine the criterion validity of the LES, a subgroup of randomly selected participants w ere supplied wi th the Actiwatch L (Mini Mitter Co. Inc., 2001) in order to provide an objective assessment of light for comparison. Power analyses estimates using G*Power (Faul, Buchner, Erdfelder, & Albert Georg, 2006) suggest ed that a proposed sample size of 12 indivi duals will provide sufficient power (beta > .80) to detect a large effect ( d = .80) at statistical probability values of .05. Sample sizes of 27 and 156 participants will provide sufficient power to detect medium ( d = .50) and small ( d = 0.20) effects at probability values of .05. Consequently, in order to detect a medium effect, a subsample of 15 participants from each age group was selected to provide validation of the LES using the ActiwatchL. Correlations were run to compare the scores received fo r the LES variables of light duration and light intensity to the comparable Actiware variables of TALT and Total Exposure. There were no significant correlations between the objective and subjective measures of light. Possible

PAGE 62

62 explanations for the lack of association include the discrepancy between subjective and objective measurement not unlike what is observed when assessing sleep using sleep diaries versus Actiwatch L. It is possible that the perception of light exposure is a different construct than the objective assessment obtained through the ActiwatchL. An additional explanation is that the LES used a retrospective assessment format where individuals had to recall activities the day earlier. This method could have resulted in a different memor y of the experience of light compared to what was captured using the Actiwatch L device. Additionally, the geographical location of the majority of study participants, southeastern United States, could have affected the amount of light exposure participa nts received. Greater variability in light exposure could have resulted from a different study location. Finally, the instructions on the LES asked participants to record the brightest amount of sunlight they were exposed to in each hour. Consequently, participants may have only been exposed to five minutes of bright sunlight but the LES measure would reflect a full hour of bright sunlight exposure. Therefore, there could be overestimations of light exposure with the LES compared to the Actiwatch LArousal dev ice The Affect Grid, a scale measuring pleasure and arousal was completed each day by the participants (Russell, Weiss, & Mendelsohn, 1989 ; see Appendix 2). The Affect Grid is designed to capture two dimensions of affect and arousal: pleasure dis pleasure and arousal sleepiness. A variation of the grid was formatted for online use. For the two items, participants were asked to indicate how they generally felt the day before ranging from high arousal to sleepiness and unpleasant to pleasant feelin gs (9 point scale ranging from 1 9). The resulting variables are the arousal grid variable which indicates level of arousal and the pleasant grid which indicates the level of pleasant feelings.

PAGE 63

63 Validity was demonstrated in a sample of undergraduates who w ere asked to assess the meaning of 28 emotionrelated words using the Affect Grid. C orrelations indicated high convergent (0.89 0.95) and low discriminate validity (0.01 0.17). Undegraduates also used the Affect Grid to assess the feelings expressed in 20 photographs of faces (Ekman & Friesen, 1976; Russell & Bullock, 1985) Th e splithalf reliability was 0.99 for pleasure and 0.97 for arousal. Convergent and discriminate validity scores were similarly high and low as for the emotion related words ( 0.91 0.97 and 0.01 0.28). For measures of criterion validity, the Affect grid predicted scores on the Scales of Pleasure and Arousal (Mehrabian & Russell, 1974; 0.77 and 0.80) and scores on the PANAS (Watson, Clark, & Tellegen, 1988 ; 0.48 and 0.62). The PreSleep Arousal Scale (PSAS; Nicassio et al., 1985) was completed each day by the participants. The PSAS is used to measure both cognitive and somatic symptoms of arousal ex perienced at bedtime (Appendix 2 ). Two global measures are obtained from the scale: Somatic and Cognitive arousal. Individuals rate eight cognitive arousal items (e.g., racing mind) and eight somatic arousal items (e.g., muscle tension) on a scale ranging from 1 (not at all) to 5 (extremely). A total score ranging from 8 to 4 0 is computed for each subscale (cognitive and somatic). Higher scores are associated with higher levels of arousal. The items for the PSAS were derived from clinical observation interviews with individuals experiencing a disturbance of their sleep. The internal consistency and test retest reliability of the scale are adequate. Cronbachs alpha for the cognitive and somatic subscales respectively are 0.88 and 0.79 for college students, 0.67 and 0.84 for normal sleepers (age 19 65 years), and 0.76 and 0.81 for individuals diagnosed with insomnia (age 21 65 years; Nicassio et al., 1985) All items were found to significantly correlate with their respective subscales (cognitive and somatic). While subscale scores are significantly correlated with each o ther (r = 0.51, p < .001),

PAGE 64

64 the amount of unshared variance (74%) suggests that the subscales are statistically independent to some extent (Nicassio et al., 1985) In terms of the validity of the scale, criterion validity was exhibited by significant corr elations of the cognitive and somatic subscales of the PSAS with measures of depression (CES D, rs = 0.40, 0.41 respectively, ps < .001) and anxiety (Taylor MAS, rs = 0.50, 0.58 respectively, ps < .001; Nicassio et al., 1985) for individuals age 19 through 65 years of age. Face validity of the categorization of the scale into cognitive and somatic categories was verified by psychologists agreeing 100% on the categorization of the 16 items (Nicassio et al., 1985) The PSAS has been described as being most useful for differentiating between high versus low arousal, insomniacs versus good sleepers, and cognitive versus physiological arousal (Wicklow & Espie, 2000) In addition to the Affect Grid and the PSAS, participants completed the UWIST Mood Adjective Ch ecklist Revised (UMACL R; Winiger, under preparation) daily. The UMACL R consists of 24 items broken into six unipolar or three bipolar subscales for scoring (Winiger, under preparation) The three bipolar scales are hedonic tone, tense arousal, and ene rgetic arousal. Given the need to assess for arousal in the study, and to limit the participants fatigue, only the tense and energetic arousal scales were used in the study. Items for the tense arousal scale include feeling relaxed, nervous, and tense w hile the energetic arousal scale items include feeling alert, drowsy, and lively. Participants are asked to describe to what extent the adjectives describe their mood for the previous day (four point likert scale ranging from definitely not to definitely) Two variables are derived from this scale: the UMACL R energetic (higher scores indicate greater energy, alertness) and the UMACL R tenseness (higher scores indicate greater tenseness, nervousness).

PAGE 65

65 Reliability and validity of the UMACL R was examined i n a sample of undergraduate students. Cronbachs alpha estimates ranged from 0.82 to 0.87 for the energetic arousal scale and 0.78 to 0.83 for the tense arousal scale. Convergent validity estimates for the UMACL R and the Profile of Mood States for Adole scents (POMS A ) were 0.84 for the energetic arousal scale, 0.68 for the tension arousal scale (Terry, Lane, Lane, & Keohane, 1999) Convergent validity estimates for the UMACL R and Thayers ActivationDeactivation Adjective Checklist (AD ACL) ranged from 0.79 to 0.85 for the energetic arousal scale and 0.60 to 0.69 for the tense arousal scale (Thayer, 1986) Affect The Positive and Negative Affect Scale (PANAS; Appendix 2 ) was completed each day by the participants. The PANAS is a measure of mood s tates and consists of two 10item mood scales asking participants to rate specific feelings and emotions associated with positive affect (e.g., happy) and negative affect (e.g., distress ) Individuals respond to each item using a likert scale ranging from 1 (very slightly or not at all) to 5 (extremely) indicating how they feel at the moment (Watson et al., 1988) The PANAS has adequate reliability with internal consistency estimates for college students ranging from Cronbachs alpha .85 (negative affect sc ale) to .89 (positive affect scale; Watson et al., 1988). For community dwelling adults, the internal consistency of the negative affect scale was Cronbachs alpha of .87 and for the positive affect scale .86 (Watson et al., 1988). Test retest correlatio ns of the positive affect scale within college samples ranged from .47 to .68 while the test retest correlation of the negative affect scale ranged from .39 to .71 (Watson et al., 1988) The PANAS has good criterion validity for both college students and c ommunity dwelling adults with the negative affect scale correlating with other brief inventories of negative affect

PAGE 66

66 (Hopkins Symptom Checklist, r = .65; Beck Depression Inventory II, r = .56; Watson et al, 1988) In addition to the PANAS, the pleasant unpleasant feeling grid from the Affect Grid scale (see above) was used to assess for positive/negative affect. S leep Diary Sleeping behaviors were measured using a sleep diary. The sleep diary (Appendix 2) was completed each day by the participant s The sleep diary provides spaces for the individual to record the time spent napping, bedtime, the amount of time to fall asleep, the number of awakenings, the total time spent awake during the night, wake up time, out of bed time, and sleep quality rating (Li chstein, Riedel, & Means, 1999) Sleep diaries have become a primary form of sleep assessment due to their ease of use, ecological validity, and reliability for assessing sleep (Lichstein & Morin, 2000) Sleep diaries have been recommended over polysom nography (typically the gold standard in sleep assessment) and are typically more widely used than either polysomnography or actigraphy in research settings (McCrae et al ., 2006) The variables of interest for the present study that will be derived from the sleep diary are: sleep onset latency (SOL; initial time from lights out until sleep onset), wake time after sleep onset (WASO; time spent awake after initial sleep onset until last awakening), total wake time (TWT; total time spent awake during the ni ght), total sleep time (TST; computed by subtracting total wake time from time in bed); number of awakenings (NWAK; the number of times you awaken during the night), and sleep quality rating ( SQR ; a fivepoint rating of the quality of sleep ranging from po or to excellent ).

PAGE 67

67 Procedure Participants were recruited from the North Central Florida area and online. Recruitment materials consisted of flyers posted around the University of Florida and the community of Gainesville, FL, online advertisements in comm unity classifieds such as craigslist, undergraduates were recruited from undergraduate psychology courses at the University of Florida and from the undergraduate psychology research pool, and older adults were recruited using the University of Floridas Ag e Network Participant Registry. For the younger adults, 2% of the participants were recruited by word of mouth (e.g., from a friend participating in the study), 4% were recruited using an online description of the study, 12% were recruited using craigslis t, 38% were recruited from advanced psychology classes, and 44% were recruited from the undergraduate psychology research pool. For older adults, 6.1% were recruited using online descriptions of the study, 24.5% were recruited using the Age Network Partic ipant Registry, and 34.6% were recruited both by word of mouth and f rom craigslist. The study materials were completed online. After a brief introduction to the study individuals complete d the electronic informed consent approved by the University of Florida Institutional Review Board Once consent was given participants were screened for the exclusionary criteria consisting of 1) currently working shift work; 2) unable to complete study materials using a computer; 3) visual deficits that preclude parti cipation (e.g., have severe self reported difficulty reading the newspaper); 4) are currently on vacation or plan to take a vacation during the next two weeks; 5) have been diagnosed with a dementia disorder; 6) sleep disorders other than insomnia (e.g., periodic limb movement disorder, sleep apnea); 7) are currently pregnant. If participants met the criteria for participating in the study, they completed the baseline surveys online (i.e., the demographic and health questionnaire, the BDI II, and the STAI Y1). Participants were given the option to wear the Actiwatch device if they were within

PAGE 68

68 traveling distance to the University of Florida. 30 individuals (15 younger and 15 older adults) agreed to wear the watch (see discussion in the L ight Exposure Scale measure section of the methods). After completing the baseline measures, participants received an email the following morning with the first of fourteen daily surveys consisting of the Social Rhythm Metric 17, Light Exposure Scale, PSAS, Affect G rid, UMA CL R, PANAS, and the Sleep Diary The first 30 participants who completed the study were awarded $10 compensation. Undergraduate students who participated as part of the psychology research pool received research credit while undergraduate students who participated from higher level psychology classes received extra credit. Additionally, all participants received general information regarding their sleep and social rhythms as an incentive for participating. Participants The total number of individual s who participated in the study were 100 younger and older adults (50 younger adults and 50 older adults; see Tables 41 and 4 2 for descriptive statistics). The mean age of the younger adults was 19.88 years ( SD = 2.76), they were primarily female (72.00%), White (72.00%), college educated (80.00%), lived with another individual (86.00%), were single (96.00%), rated their heath as 3.84 out of 5 ( SD = 0.89), and reported on average 0.34 ( SD = 0.69) health conditions. The mean age of the older participants was 67.81 years ( SD = 6.73). The older adults were primarily female (62.00%), White (90.00%), college educated (92.00%), lived with another individual (71.40%), were married (58%), rated their health as 3.48 out of 5 ( SD = 0.84), and reported an average of 1.24 health conditions ( SD = 1.25).

PAGE 69

69 Table 4 1. Participant descriptive statistics for younger adults (N=50) Variable Value age, mean (SD), range 19.88 (2.85) 18 30 gender % Male 28.00 Female 72.00 race % White 72. 00 African American 10.00 Carribean American 2.00 East Asian 12.00 Middle Eastern 0.00 South Asian 4.00 education % high school 20.00 college/trade school 64.00 Associates's degree 4.00 Bachelor degree 6.00 Mast er's degree 4.00 professional degree (e.g., PhD, MD) 2.00 spouses' education % high school 20 .00 college/trade school 55 00 Associates's degree 5 0 0 Bachelor degree 0.00 Master's degree 10 0 0 professional degree (e.g., PhD, MD) 10 0 0 live alone % 14.00 marital status married 2.00 single 96 .00 common law 2 .00 widowed 0.00 divorced 0.00 seperated 0.00 health rating a 3.8 4 (0.89) number of health conditions b 0.34 (0. 69) sleep stat us % c good sleepers 86.00 poor sleepers 14.00

PAGE 70

70 aSelf rating of health ranging from poor (1) to excellent (5). bTotal number of medical conditions from the following list: heart problems, cancer, hypertension, neurological disorder, breathing disorder, urinary problems, diabetes mellitus, pain, gastrointestinal disorder, mental health disorder, and other. cParticipants were categorized as poor sleepers if they reported 31 minutes or more of unwanted awake time during the night (sleep onset lat ency or wake after sleep onset) 3 nights or more per week.

PAGE 71

71 Table 4 2. Participant descriptive statistics for older adults (N=50) Variable Value age, mean (SD), range 67. 81 (6.7 3 ) 60 95 gender % Male 38. 00 Female 6 2.00 race % White 90.00 African American 4. 00 Carribean American 0.00 East Asian 4.00 Middle Eastern 2.00 South Asian 0.00 education % high school 8. 0 0 college/trade school 30.00 Associates's degree 10. 0 0 Bachelor degre e 22. 0 0 Master's degree 20.00 professional degree (e.g., PhD, MD) 10.00 spouses' education % high school 19.40 college/trade school 19.40 Associates's degree 5.60 Bachelor degree 36.10 Master's degree 11.10 professional degree (e.g., PhD, MD) 8. 3 0 live alone % 28.60 marital status married 5 9 18 single 2.00 commonlaw 6.00 widowed 14. 29 divorced 18.37 seperated 0 .00 health rating a 3.48 (0.84) number of health conditions b 1. 24 (1 .2 5 ) sleep status % c good sleepers 40.00 poor sleepers 60.00

PAGE 72

72 aSelf rating of health ranging from poor (1) to excellent (5). bTotal number of medical conditions from the following list: heart problems, cancer, hypertension, neurological disorder, breathing disorder, urinary problems, diabetes mellitus, pain, gastrointestinal disorder, mental health disorder, and other. cParticipants were categorized as poor sleepers if they reported 31 minutes or more of unwanted awake time during the ni ght (sleep onset latency or wake after sleep onset) 3 nights or more per week.

PAGE 73

73 CHAPTER 4 RESULTS The overall aim of the proposed study was to investigate the role of light, arousal, and affect in mediating the relationship between social rhythmicity an d sleep. The overall aim of the study was examined by investigating a number of specific aims (aim 1 through aim 4). Each of the following three aims was analyzed separately for the younger and older adult samples. See Table 5 1 for a summary and descri ption of the variables used in the analyses. Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep Fluctuate Daily both Within and Between Persons The purpose of aim one was to establish the extent to which individuals vary both within and between persons in their social rhythmicity, light, arousal, affect, and sleep. Social R hythmicity The intraclass correlation coefficient (ICC) for the social rhythmicity variable for younger adults was 0.23 indicating that 77% of t he variability in social rhythmicity is a within person phenomenon (see Figure 51). For older adults, the ICC was 0.28 indicating that 72% of the variability in social rhythmicity occurs within persons. Younger adults were significantly more variable wi thin persons in SRM compared to older adults t (98) = 6.42, p <.001. The mean social rhythmicity value (average variability during the study) was 2.36 ( SD = 0.81) for younger adults and 4.10 ( SD = 0.82) for older adults. The difference i n mean variability was significant t (98) = 10.71, p <.001 (see table 5 2). Light V ariables The ICC for the light variables ranged from 0.25 to 0.47 for the younger adults and 0.43 to 0.52 for the older adults (see Figure 52). For younger and older adults respectively, the amount of within person variability in the AM light scale was 74% and 55%, for the PM light scale 74% and 57%, for the light duration scale 57% and 48%, and for the light intensity scale 53% and

PAGE 74

74 49%. Younger adults were significantly more variable wit hin persons compared to older adults for AM light t (96) = 3.85, p <.001, PM light t (97) = 3.67, p <.001, light duration t (97) = 3.25, p <.01, and light intensity t (98) = 4.12, p <.001. Older adults were exposed to significantly earlier AM light t (97) = 2.94, p <.01 and more intense light t (98) = 2.89, p <.01 (see table 5 2) than younger adults. Arousal V ariables The ICC for the arousal variables ranged from 0.29 to 0.50 for younger adults and 0.47 to 0.78 for older adults (see Figure 5 3). For younger and older adults respectively, the amount of within person variability in the arousal grid was 50% and 39%, for the PSAS somatic scale 64% and 53%, for the PSAS cognitive scale 71% and 31%, for the UMACL R energetic scale 64% and 22%, and for the UMACL R ten seness scale 54% and 22%. Younger adults were significantly more variable within persons compared to older adults for the arousal grid t (97) = 4.18, p <.001, the PSAS cognitive scale t (97) = 3.02, p <.01, the UMACL R tenseness scale t (97) = 4.45, p <.001, and the UMACL R energetic scale t (97) = 6.07, p <.001. In terms of mean level values, older adults reported significantly higher cognitive arousal according to the PSAS cognitive scale t (98) = 3.27, p <.01 compared to younger adults (see table 52). Aff ect V ariables The ICC for the affect variables ranged from 0.34 to 0.58 for the younger adults and 0.56 to 0.89 for the older adults (see Figure 54). The percentage of overall variability that could be attributed to fluctuations within persons, respectively for younger and older adults, was 66% and 44% for the pleasant grid, 42% and 20% for the PANAS positive subscale, and 59% and 11% for the PANAS negative subscale. Younger adults were significantly more variable within persons compared to older adults for the pleasant grid t (95) = 2.90, p <.01, the PANAS positive scale t (97) = 6.82, p <.001, and the PANAS negative scale t (94) = 4.08, p <.001. For meanlevel

PAGE 75

75 affect, older adults reported significantly more positive affect than younger adults t (98) = 3.87, p <.001 (see table 5 2). Sleep V ariables The ICC for the sleep variables ranged from 0. 10 to 0.33 for the younger adults and 0.35 to 0.66 for the older adults (see figure 5 5). The amount of within person variability in sleep for younger and older adults respectively was 67% and 59% for SOL, 90% and 65% for WASO, 78% and 58% for TWT, 84% and 60% for TST, 81% and 34% for NWAK, and 78% and 64% for SQR. Younger adults were significantly more variable within persons compared to older adults for SOL t (98) = 2.19, p <.05, WASO t (97) = 3.68, p <.001, TWT t (98) = 2.95, p <.01, TST t (98) = 8.95, p <.001, NWAK t (96) = 2.11, p <.05, and SQR t (97) = 2.40, p <.05. Age related differences in the mean levels of sleep variables included greater SOL t (98) = 3.54, p <.001, TWT t (98) = 2.10, p <.05, NWAK t (98) = 4.95, p <.001, and less TST t (98) = 2.34, p <.05 for older adults (see table 5 2). Aim 2: Identify H ow V ariability in S ocial R hythmicity, L ight, A rousal, and A ffect is R elated t o S leep There were sig nifica nt correlations of within person variability in the social rhythmicity variable and sleep for younger adults (see Table 5 3). More variability in the timing of social activities was associated with worse overall SOL ( r = 0.75, p < .001) WASO ( r = 0. 38, p < .05 ) and TWT ( r = 0. 33, p < .05 ). Younger adults who were more variable in the timing of their daily activities tended to take longer to fall asleep, spend more time awake during the night before awakening in the morning, and spend more time awake ove rall. There were no significant correlations between variability in social rhythmicity and the sleep variables for older adults (see Table 5 7).

PAGE 76

76 In addition to examining within person variability in the social rhythmicity summary variable (derived from th e SRM 17), correlations were run to explore the possible association between variability in the discrete events of the SRM 17 (e.g., having breakfast, exercise, going outside, etc.) and sleep. For younger adults, greater within person variability in the t ime of getting out of bed was associated with worse SOL ( r = 0.29, p < .05), TST ( r = 0.28, p < .05), and NWAK ( r = 0.40, p < .01). More variability in the timing of breakfast ( r = 0.40, p < .01) and lunch ( r = 0.29, p < .04) was associated with more NWA K. For older adults, more within person variability in the timing of starting work ( r = 0.47, p < .001) and eating lunch ( r = 0.32, p < .05) was associated with less SOL. More variability in the timing of naps ( r = 0.51, p < .05) and in the time arri ving at home ( r = 0.43, p < .01) was associated with more NWAK. Greater variability in the timing of the evening snack was associated with more TWT ( r = 0.34, p < .05). More variability in the timing of the first exposure to light in the morning was assoc iated with a higher number of awakenings during the night for younger adults ( r = 0. 32, p < .05) and was not significantly associated with sleep for older adults (see Tables 5 4 and 58). Higher levels of within person variability in the arousal grid, PS AS somatic, PSAS cognitive, UMACL R energetic, and UMACL R tenseness scales were associated with worse sleep in younger adults (see Table 5 5). More variability in the arousal grid rating was associated with worse WASO ( r = 0. 37, p < .01) TWT ( r = 0. 42, p < .01) and SQR ( r = 0.33, p < .05) Higher levels of variability in the PSAS somatic scale were associated with worse WASO( r = 0. 32, p < .05) NWAK ( r = 0. 31, p < .05 ) and SQR ( r = 0.32, p < .05) More variability in the PSAS cognitive scale was as sociated with worse SOL ( r = 0.36, p < .05) and WASO ( r = 0. 31, p < .05 ) More variability in the UMACL R energetic scale was associated

PAGE 77

77 with worse TWT ( r = 0. 30 p < .05) and SQR ( r = 0.38, p < .01) Higher levels of variability in the UMACL R tenseness scale was associated with worse SQR ( r = 0.33, p < .05) Higher levels of variability in the PSAS somatic and PSAS cognitive scales were associated with worse SOL ( r = 0. 36, p < .05) and TST ( r = 0.30, p < .05 ) respectively, for older adults (see Tabl e 5 9). In terms of variability in affect, higher levels of within person variability were associated with worse sleep overall for younger adults. More variability in the pleasant grid ratings was associated with worse SQR ( r = 0.31, p < .05 ) More var iability in the PANAS positive scale was associated with more NWAK ( r = 0. 29, p < .05 ) and more variability in the PANAS negative scale was associated with worse SQR ( r = 0.38, p < .01; see Table 5 6). There were no significant associations between variability in affect and mean level sleep in older adults. Aim 3 : Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light, Arousal, and Affect are Coupled with Daily Fluctuations in Sleep Prior to calculating the multilevel model ( MLM ) m ulticollinearity between the predictor variables of the light, arousal, and affect groups was examined through the estimation of a multivariate mixed effects null model (see Tables 5 11 through 5 16). Multicollinearit y is an assumption that is not routine ly examined in the MLM literature. In other parametric techniques, collinearity between predictors becomes especially problemsome at levels exceeding 0.90 (Hair, Anderson, Tatham, & Black, 1998) In MLM analyses, multicollinearity is of concern when cros s level interactions are formed (Tabachnick & Fidell, 2007) Given the lack of consensus on how best to handle Multicollinearity in MLM, and given the lack of cross level interactions in the present study, variables that were correlated at or above a level of 0.80 were residualized to control for the covariance. T here were t wo correlations that met or exceeded a correlation value of .80 at the between person level (the arousal grid and UMACL R energetic for younger adults r

PAGE 78

78 = 0.82, p < 0.001 and UMACL R energetic and UMACL R tenseness for older adults r = 0.90, p < 0.001). The residualized variables of the arousal grid and UMACL R energetic (for younger adults) and UMACL R tenseness and UMACL R energetic (for older adults) were used in the following M LM analyses. In the first step of the MLM, a null model was estimated for each of the six sleep outcomes (SOL, WASO, TWT, TST, NWAK, and SQR) that included only fixed and random intercepts for these variables. This model served as a comparison to later m odels that included the predictor variables and was used to derive the intraclass correlation coefficient used in aim 1 In step two of the MLM, the linear effects of time were added to the model to control for any systematic change in the data resulting from the effects of 14 days of measurement. In step three of the MLM, the estimates of the fixed and random intercepts and fixed and random linear slopes were added to the model. In the models with SRM predicting sleep, there were significant predictor s at the within and between person levels f or younger adults (see Table 5 17). At the within persons level time 0.26, t ( 90.64 ) = 2.46, p <.05) indicating that the amount of time it took younger adults to fall asleep decreased over the 14 study days. At the between person level, more irregularity overall was associated with overal 3.40, t (48.70) = 2.51, p < .05) 5.44, t (49.97) = 2.03, p < .05) 0.16, t (48.58) = 2.03, p <.05). These results suggest that the more ir regular younger adults were in their daily activities, overall, the longer it took to fall asleep, the more time they spent awake in total during the night, and the low er they rated their sleep quality. For older adults, there were significant predictors at the within person level ( level 1 ) with time significantly pre dicting 0.03, t (50.77 ) = 2.30, p <.05) suggesting that the number of awakenings decreased

PAGE 79

79 over the study period for older adults (see table 5 18). Although none of the SRM variables significantly predicted sleep for older adults, higher overa ll levels of irregularity approached 0.45, t (46.68) = 1.90, p =.06 ) For the prediction of sleep by light, there were significant predictors at levels 1 and 2 for the younger adults (see table 5 19). At the within person level (level 1) 0.30, t (472.46) = 2.35, p <.05) t (476.38) = 2.04, p <.05) significantly predicted SOL. Younger adults SOL decreased over the 14 days of the study and on days when younger adults were exposed to li ght later in the day, they took longer to fall asleep. AM light t (119.74) = 2.74, p <.01) indicating that days with later exposure to light during the morning were associated with more total time spent awake tha t t (140.03) = 2.06, p <.05) and less light intensity 0.05, t (413.80) = 2.67, p <.01) were associated with less TST for that night. Days with longer exposure to light but with less light intensity wer e associated with more total time spent sleeping t (53.73) = 2.99, p <.01) light t (54.41) = 3.48, p <.01) 0.09, t (53.88) = 3.34, p <.01) were associated wit h TST. Younger adults who overall received light exposure later in the day, higher levels of duration overall, and lower levels of light intensity reported spending more total sleeping overall. For older adults, there were significant predictors at the w ithin and between person levels (see table 5 20). AM light ( 1.70, t ( 547.99) = 2.20, p <.05) and light duration 0.04, t ( 580.29) = 2.21, p <.05) significantly predicted WASO and time 0. 03, t (5 3.18) = 2.51, p <.0 5) significantly predicted NWAK. These results suggest that days with exposure to light later in the morning and days with exposure to light for a longer duration was associated with less

PAGE 80

80 wake time during the night after falling asleep Also, the number of awakenings older adults reported decreased over the time of the study. At the betwe en person level, PM light was associated with SQR ( 0.14, t ( 54.31) = 2. 37, p <.05) light duration was associated with TST 0.22, t ( 49.40) = 2.41 p <.05) and light intensity was associated with SOL 0.03, t ( 47.67) = 2.29, p <.05) These results suggest that overall later exposure to light in evening was associated with overall better sleep quality ratings, and higher levels of overall duration of light exposure and light intensity were associated with more total time spent sleeping and longer time to fall asleep, respectively. For the relation ship between the arousal variables and sleep in younger adults, there were significant predictors at both level 1 and level 2 (see table 5 21). At level 1, the PSAS cognitive subscale significantly predicted SOL ( 1.11, t ( 45.44) = 4.93, p <.0 010.87, t ( 23.09) = 2. 47, p 2.09, t ( 35.37) = 4.11, p <.0 04.47, t ( 549.78) = 3.28, p 6, t ( 35.35) = 3.05, p <.0 1 0.05, t ( 33.80) = 4.69, p <.001). These results suggest that on days when younger adults have higher ratings of cognitive arousal before sleep, they take longer to fall asleep, are awake for longer during the night, have less total sleep time, more awakenings, and lower sleep quality ratin gs. Random effects were significant for the PSAS cognitive scale ( p < .05 to p < .01) predicting SOL, WASO, and TWT suggesting the level 1 effects mentioned above for SOL, WASO, and TWT vary in magnitude by the individual. T ime ( 0.01, t ( 145.51) = 1.97, p <.05) the 0.09, t ( 570.53) = 3.26, p <.0 1 ) and UMACL t (62.42) = 2.89, p <.01) also significantly predicted SQR suggesting that on days when younger adults experienced less physical tension before bed and felt more alert during the day they reported better sleep quality and that sleep quality ratings decreased over the study time

PAGE 81

81 For older adults, arousal predicted sleep at both level 1 and level 2 ( see table 5 22). For th e within person level (level 1), the PSAS cognitive subscale significantly predict ed SOL ( = 0.94, t (33.19) = 2.61, p <.05) WASO = 2. 11, t (50.85) = 3.04, p <.01), TWT ( = 3.55, t (40.00) = 3.56, p <.05), TST ( = 3.80, t (41.83) = 2.48, p <.05), NWAK ( = 0.08, t (33.81) = 4.81, p <.001 = 0.09, t (35.01) = 5.25, p <.001). Th ese results suggest that on days when older adults experience more cognitive arousal before bed, they take longer to fall asleep, have more awake time during the night, have less total sleep time, more awakenings and lower sleep quality ratings. Random ef fects were significant for the PSAS cognitive scale predicting WASO and the SQR suggesting that the effects mentioned above vary in magnitude between individuals. There was also a significant within person effect for time ( 0.02, t ( 62.28) = 2.25, p <.05) indicating that older adults sleep quality ratings decreased over the course of the study. This effect is qualified by a significant random effect for time indicating that the magnitude of this effect varies significa ntly within individuals. For level 2 (between person effects), the PSAS cognitive subscale significantly predicts SOL 1.78, t (46.49) = 2.38, p <.05) WASO = 2.58, t (48.02) = 2.68, p <.05) TWT 07, t (47.15) = 2.69, p <.05) and SQR 0.0 7, t (45.85) = 3.41, p <.001) Overall higher levels of cognitive arousal were associated with overall longer times to fall asleep, more wake time during the night, and poorer sleep quality ratings in older adults. Additionally, at level 2, the arousal g rid significantly predicted SQR ( 0.14 t ( 46.45) = 2.96, p <.0 5 ) Older adults who had overall higher scores on the arousal grid (indicating sleepiness) reported poorer sleep quality. For the affect variables, there were significant predictions of sleep at level 1 and 2 for young er adults (see table 5 = 0.25, t (93.38) = 2.26, p <.05), the pleasa nt grid predicted SQR 0.13, t ( 53.26) = 4.84, p

PAGE 82

82 <.001) the PANAS positive scale predicted SQR 0.02, t ( 54.13) = 2. 62, p <.05) and the PANAS negative scale predicted TWT 0.87, t ( 22.59) = 2. 07, p <.05) These results suggest that the time spent falling asleep decreased across the study days, on days with higher ratings of pleasant feelings and more positive affect, sleep quality was better. Also, days with negative affect were associated with more time spent awake during the night. For the level 2 effects, the pleasant grid significantly predicted TST 28.14, t ( 49.58) = 3.61 p <.0 1) and SQR 0.14, t ( 49.75) = 2.15, p <.05) while the PANAS negative significantly predicted SQR 0.05, t ( 50.79) = 2.22, p <.05) significantly predicted SQR. Therefore, younger adults who overall had higher ratings of pleasant feelings slept for longer periods of tim e and reported higher sleep quality while those with lower negative affect reported higher sleep quality. For the older adults, affect was associated with sleep at both level 1 and level 2 (see table 524). At the within person level, time 0.03, t ( 51.96) = 2.28, p <.05) predicted NWAK. The pleasant grid significantly predicted WASO 2.71, t ( 19.56) = 2.29, p <.05) TWT 3.81, t ( 16.36) = 2.41, p <.05) TST 5.50, t ( 7.49) = 2.55, p <.05) and SQR 0.13, t ( 25.04) = 4.08, p <.001) For older adults, the number of awakenings decreased over time, and days with higher ratings of pleasant feelings were associated with less wake time during the night, more total sleep time, and better sleep quality. At the between person level, the pleas ant grid was associated with WASO ( 7.07, t ( 50.40) = 2.25, p <.05) TST 20.14, t ( 47.91) = 2.44, p <.05) and SQR 0.17 t ( 49.83) = 2. 44, p <.05) the PANAS positive was associated with TST 2.94, t ( 47.05) = 2. 17 p <.05) and the PANAS negative 0.07, t ( 47.71) = 2.06, p <.05) was associated with NWAK. Overall, older adults who reported higher pleasant feelings reported less overall time spent awake, sleep for less total sleep time and reported higher sleep

PAGE 83

83 quality. Also, overall more positive affect was associated with more total sleep time while negative affect was associated with fewer awakenings. Aim 4 : Identify to What Extent Light, Arousal, and Affect Mediate the Relationship Between Social Rhythmicity and Sleep The goal of aim was to investigate the possible mediating effects of light, arousal, and affect on the relationship between social rhythmicity and sleep. This analysis was to be investigated using a four step mediation model proposed by Baron and Kenny (1986). In order to demonstrate mediation, both the predictor (SRM) and the mediators (light, arousal, and affect) needed to be significantly associated with the outcome variables (sleep). While significantly associations existed between variables for both younger and older adults, there were no sign ificant ( p < .05) relationships that were the same for both SRM and sleep. Consequently, the mediating effects of light, arousal, and affect could not be estimated as the first two steps of Baron and Kennys model could not be achieved.

PAGE 84

84 Table 5 1. Summar y of the variables used in the analyses. Construct Variable Explanation Social Rhythmicity SRM a weekly or daily score ranging from 0 7 with higher scores indicating greater regularity Light AMlight the first hour of the day exposed to cloudy or bri ght sunlight (equivalent to > 1000 lux) PMlight the last hour of the day exposed to cloudy or bright sunlight (equivalent to > 1000 lux) light duration the number of hours exposed to cloudy or bright light over a 24 hour period light intensity t he total amount of light [ranging from 1 5] multiplied by the number of hours in a 24 hour period Arousal Arousal grid 9 point scale measuring high arousal to sleepiness ; higher scores indicate greater sleepiness PSASsomatic higher scores indicate g reater physical tension before bed (e.g, heart racing) PSAScognitive higher scores indicate greater cognitive tension before bed (e.g., worry about problems) UMACL R energetic higher scores indicate feeling more alert, energetic, lively during the day UMACL R tenseness higher scores indicate feeling more nervous, anxious, tense during the day Affect Pleasant grid 9 point scale measuring unpleasant to pleasant feelings, higher scores indicate greater pleasant feelings PANAS positive higher scor es indicate more positive affect during the day (e.g., interested, excited, strong, enthusiastic) PANAS negative higher scores indicate more negative affect during the day (e.g., distressed, upset, scared, irritable, ashamed) Sleep SOL the amount of time it takes to fall asleep WASO the amount of time spent awake during the night after falling asleep and before awakening in the morning TWT the total amount of time spent awake during the night including the time to fall asleep, the time spent awak e during the night, and the time spent awake in the morning before getting out of bed TST the total amount of time spent sleeping during the night NWAK the number of awakenings during the night SQR the rating of the quality of sleep ranging form poor to excellent on a 5 point scale

PAGE 85

85 F igure 5 1. Percentage of total variability in social rhythmicity that can be attributed to within person variability. Note: Differences are significant at the .05 level (2 tailed).

PAGE 86

86 Figure 5 2. Percentage of total variability in light measures that can be attributed to within person variability. Note: *** D ifferences are significant at the 0.001 level (2tailed), ** differences are significant at the 0.01 level (2 tailed) ** ** *

PAGE 87

87 Figure 5 3. Percentage of total variability in arousal measures that can be attributed to within person variability Note: *** D ifferences are significant at the 0.001 level (2tailed), ** differences are significant at the 0.01 level (2 tailed ). ** ** **

PAGE 88

88 Figure 5 4. Percentage of t otal variability in affect measures that can be attributed to within person variability. Note: *** D ifferences are significant at the 0.001 level (2tailed), ** differences are significant at the 0.01 level (2 tailed) ** **

PAGE 89

89 Figure 5 5. Percentage of total variability in sleep variables that can be attributed to within person variability. Note: *** D ifferences are significant at the 0.001 level (2tailed), ** differences are significant at the 0.01 level (2 tailed), and differences are significant at the .05 level (2tailed). * * **

PAGE 90

90 Table 5 2. Age related differences in mean level values for social rhythmicity, light, arousal, affect, and sleep variables Construct Variable M (SD) M (SD) t df p Younger Older Social Rhythmicity SRM 2.36 0 .81 4.10 0.82 10.71 98 < 0 .00 1 Light AMlight 10.99 1.55 9.95 1.93 2.94 97 0.004 PMlight 15.03 1.54 15.34 1.77 0.95 97 0.35 light duration 179.40 120.74 243.15 162.67 1.72 98 0.09 light intensity 3535.23 356.75 3740.71 355.51 2.89 98 0.0 1 Arousal Arousal grid 4.07 1.50 3.80 1.59 0.87 98 0.39 PSASsomatic 8.68 0.87 9.02 1.43 1.44 96 0.15 PSAScognitive 10.82 2.31 13.05 4.25 3.27 98 0.001 UMACL R energetic 23.24 3.47 23.03 6.99 0.18 98 0.86 UMACL R tenseness 14.51 3.40 15 .20 6.71 0.65 98 0.52 Affect Pleasant grid 6.35 1.11 6.83 1.68 1.70 98 0.09 PANAS positive 26.44 7.29 32.84 9.15 3.87 98 < 0.00 1 PANAS negative 12.92 2.60 13.41 6.60 0.48 98 0.63 Sleep SOL 12.59 8.19 24.48 17.74 4.30 98 <0.001 WASO 5.32 5.53 25.40 22.77 6.06 98 <0.001 TWT 30.93 15.99 72.09 44.50 6.16 98 <0.001 TST 473.76 52.19 421.60 59.83 4.65 98 <0.001 NWAK 0.70 0.57 1.64 1.34 4.57 98 <0.001 SQR 3.82 0.45 3.66 0.59 1.48 98 0.14 Note: bolded p values are signific ant at least at <.05.

PAGE 91

91 Table 5 3. Correlations of within person variability in s ocial r hythmicity and mean level sleep in younger adults. (N=50) 1 2 3 4 5 6 7 1. SRM 1.00 2. SOL 0.75*** 1.00 3. WASO 0.38* 0.16 1.00 4. TWT 0.33* 0.24 0 .43*** 1.00 5. TST 0.03 0.13 0.35** 0.21 1.00 6. NWAK 0.04 0.10 0.20 0.11 0.19 1.00 7. SQR 0.09 0.18 0.11 0.04 0.11 0.07 1.00 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2 tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. Table 5 4. Correl ations of within person variability in light and meanlevel sleep in younger adults. (N=50) 1 2 3 4 5 6 7 8 9 10 1. AMlight 1 2. PMlight 0.20 1 3. light duration 0.07 0.06 1 4. light intensity 0.25 0.01 0.53*** 1 5. SOL 0.14 0.02 0.01 0.09 1 6. WASO 0.07 0.06 0.04 0.12 0.32* 1 7. TWT 0.27 0.01 0.05 0.24 0.75*** 0.57*** 1 8. TST 0.01 0.12 0.03 0.05 0.26 0.06 0.34* 1 9. NWAK 0.32* 0.24 0.11 0.19 0.27 0.44** 0.30* 0.13 1 10. SQR 0.06 0.15 0.13 0.13 0.13 0.20 0.23 0.43** 0.17 1 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2 tailed). Note: SOL= sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating.

PAGE 92

92 Table 5 5. Correlations of within person variability in arousal and mean level sleep in younger a dults. (N=50) 1 2 3 4 5 6 7 8 9 10 11 1. arousal grid 1 2. PSASsomatic 0.38** 1 3. PSAScognitive 5.11E+07 0.36** 1 4. UMACL R energetic 4.87E+07 0.21 0.39** 1 5. UMACL R tenseness 6.17E+07 0.37** 0.21 0.38** 1 6. SO L 0.25 0.15 0.36* 0.12 0.15 1 7. WASO 0.37** 0.32* 0.31* 0.24 0.05 0.32* 1 8. TWT 0.42** 0.24 0.27 0.30* 0.23 0.75*** 0.57*** 1 9. TST 0.003 0.07 0.01 0.19 0.14 0.26 0.06 0.34* 1 10. NWAK 0.23 0.31* 0.24 0.22 0.20 0.27 0.44** 0.30 0.13 1 11. SQR .334* 0.32* 0.21 0.38** 0.33* 0.13 0.20 0.23 0.43** 0.17 1 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating.

PAGE 93

93 Table 5 6. Correlations of within person variability in affect and me an level sleep in younger adults. (N=50) 1 2 3 4 5 6 7 8 9 1. pleasant grid 1 2. PANASpositive 0.52*** 1 3. PANASnegative 0.44** 0.36** 1 4. SOL 0.06 0.06 0.17 1 5. WASO 0.17 0.25 0.04 0.32* 1 6. TWT 0.18 0.13 0.12 0.7 5*** 0.57*** 1 7. TST 0.14 0.07 0.17 0.26 0.06 0.34* 1 8. NWAK 0.15 0.29* 0.03 0.27 0.44** 0.30* 0.13 1 9. SQR 0.31* 0.23 0.38** 0.13 0.20 0.23 0.43** 0.17 1 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c or relation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2 tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR= sleep quality rating. Table 5 7. Correlations of within person variability in s ocial r hythmicity and mean level sleep in older adults. (N=50) 1 2 3 4 5 6 7 1. SRM 1 2. SOL 0.11 1 3. WASO 0.22 0.17 1 4. TWT 0.26 0.71*** 0.69*** 1 5 TST 0.11 0.28* 0.07 0.31* 1 6. NWAK 0.02 0.04 0.53*** 0.33* 0.04 1 7. SQR 0.17 0.29* 0.57** 0.51*** 0.11 0.50** 1 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2 tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating.

PAGE 94

94 Table 5 8. Correlations of within person variability in light and meanlevel sleep in older adults. (N=50) 1 2 3 4 5 6 7 8 9 10 1. AMlight 1.00 2. PMlight 0.19 1.00 3. light duration 0.27 0.04 1.00 4. light intensity 0.12 0.18 0.58*** 1.00 5. S OL 0.09 0.12 0.15 0.10 1.00 6. WASO 0.03 0.17 0.02 0.15 0.17 1.00 7. TWT 0.09 0.02 0.04 0.16 0.71*** 0.69*** 1.00 8. TST 0.17 0.05 0.15 0.04 0.28* 0.07 0.31* 1.00 9. NWAK 0.26 0.03 0.22 0.17 0.04 0.53*** 0.33* 0.04 1.00 10. SQR 0.06 0.05 0.13 0.03 0.29* 0.57*** 0.51*** 0.11 0.50*** 1.00 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level ( 2tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating.

PAGE 95

95 Table 5 9. Correlations of within person variability in arousal and mean level sleep in older adults. (N=50) 1 2 3 4 5 6 7 8 9 10 11 1. arousal grid 1.00 2. PSASsomatic 0.10 1.00 3. PSAScognitive 0.25 0.50** 1.00 4. UMACL R energetic 0.18 0.11 0.12 1.00 5. UMACL R tenseness 0.37* 0.24 0.40** 0.59 *** 1.00 6. SOL 0.15 0.36* 0.21 0.08 0.12 1.00 7. WASO 0.11 0.04 0.16 0.11 0.01 0.17 1.00 8. TWT 0.08 0.19 0.23 0.08 0.14 0.71*** 0.69*** 1.00 9. TST 0.00 0.07 0.30* 0.01 0.05 0.28* 0.07 0.31* 1.00 10. NWAK 0.01 0.08 0. 05 0.07 0.03 0.04 0.53*** 0.33* 0.04 1.00 11. SQR 0.11 0.27 0.26 0.06 0.18 0.29* 0.57*** 0.51*** 0.11 0.50*** 1.00 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2 tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating.

PAGE 96

96 Table 5 10. Correlations of wi thin person variability in affect and mean level sleep in older adults. (N=50) 1 2 3 4 5 6 7 8 9 1. pleasant grid 1.00 2. PANASpositive 0.36* 1.00 3. PANASnegative 0.07 0.45** 1.00 4. SOL 0.13 0.03 0.01 1.00 5. WASO 0.06 0 .12 0.08 0.17 1.00 6. TWT 0.15 0.00 0.07 0.71*** 0.69*** 1.00 7. TST 0.05 0.03 0.04 0.28* 0.07 0.31* 1.00 8. NWAK 0.08 0.03 0.04 0.04 0.53*** 0.33* 0.04 1.00 9. SQR 0.14 0.05 0.16 0.29* 0.57** 0.51** 0.11 0.50** 1.00 Note: *** Correlation is significant at the 0.001 level (2tailed), ** c orrelation is significant at the 0.01 level (2 tailed), and Correlation is significant at the .05 level (2 tailed). Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=to tal wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. Table 5 11. Correlations among light variables between persons (below diagonal) and within person (above diagonal) for younger adults Amlight Pmlight light duratio n light intensity Amlight 0.5 0 *** 0.4 0 *** 0. 20*** Pmlight 0. 57 ** 0.69*** 0.50*** light duration 0.5 4 ** 0.6 7 *** 0.3 2 ** light intensity 0. 06 0.5 7 ** 0. 31 Note: *Correlation is significant at the p < 0.05 (2 tailed) ; **c orrelation is signific ant at the p < 0.01 (2tailed) ; ***c orrelation is significant at the p < 0.001 (2tailed)

PAGE 97

97 Table 5 12. Correlations among light variables between persons (below diagonal) and within person (above diagonal) for older adults Amlight Pmlight light duratio n light intensity Amlight 0.05 0.5 5 *** 0.4 5 *** Pmlight 0.1 2 0.5 3 *** 0.31*** light duration 0.74*** 0. 49 ** 0.67*** light intensity 0.6 4 *** 0.5 6 ** 0. 79 *** Note: *Correlation is significant at the p < 0.05 (2 tailed) ; **c orrelation is signif icant at the p < 0.01 (2tailed) ; ***c orrelation is significant at the p < 0.001 (2tailed) Table 5 13. Correlations among arousal variables between persons (below diagonal) and within person (above diagonal) for younger adults arousal grid PSAS somati c PSAS cognitive UMACL R energetic UMACL R tense arousal grid 0. 02 0.1 6*** 0.5 7 *** 0.1 7* ** PSAS somatic 0. 20 0.29*** 0.13** 0.22*** PSAS cognitive 0.2 7 0.52** 0.13** 0.25*** UMACL R energetic 0.8 2 *** 0.26 0.39* 0.36*** UMACL R tense 0.60** 0.39* 0.52* 0.72*** Note: *Correlation is significant at the p < 0.05 (2 tailed) ; **c orrelation is significant at the p < 0.01 (2tailed) ; ***c orrelation is significant at the p < 0.001 (2tailed)

PAGE 98

98 Table 5 14. Correlations among arousal variables between persons (below diagonal) and within person (above diagonal) for older adults arousal grid PSAS somatic PSAS cognitive UMACL R energetic UMACL R tense arousal grid 0.0 6 0.0 2 0. 25 0. 0 5 PSAS somatic 0.48** 0.37 0.11* 0.10* PSAS cogniti ve 0.34* 0.56** 0.02 0.09* UMACL R energetic 0.31* 0.07 0.11 0.58*** UMACL R tense 0.25 0.19 0.20 0.90*** Note: *Correlation is significant at the p < 0.05 (2 tailed) ; **c orrelation is significant at the p < 0.01 (2tailed) ; ***c orrelation is significant at the p < 0.001 (2tailed) Table 5 15. Correlations among affect variables between persons (below diagonal) and within person (above diagonal) for younger adults pleasant grid PANAS positive PANAS negative pleasant grid 0.3 2 *** 0. 31 *** PANAS positive 0.62 0.1 3 ** PANAS negative 0. 53 ** 0.1 6 Note: *Correlation is significant at the p < 0.05 (2 tailed) ; **c orrelation is significant at the p < 0.01 (2tailed) ; ***c orrelation is significant at the p < 0.001 (2tailed) Table 5 16. Correlations among affect variables between persons (below diagonal) and within person (above diagonal) for older adults pleasant grid PANAS positive PANAS negative pleasant grid 0. 24 *** 0.2 0 *** PANAS positive 0. 79 *** 0. 11* PANAS negative 0. 57 ** 0.3 6 Note: *Correlation is significant at the p < 0.05 (2 tailed) ; **c orrelation is significant at the p < 0.01 (2tailed) ; ***c orrelation is significant at the p < 0.001 (2tailed)

PAGE 99

99 Table 5 17. Multilevel model with social rhythmicity variab les predicting sleep in younger adults Fixed Effects SOL WASO TWT TST NWAK SQR Predictor Variable B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) Within person time 0.26 (0.11) 2.46 (90.64)* 0.12 (0. 13) 0.97 (630.64) 0.27 (0.26) 1.05 (625.80) 1.04 (1.00) 1.05 (625.73) 0.01 (0.01) 0.89 (108.14) 0.003 (0.01) 0.39 (105.33) SRM centered 0.05 (0.35) 0.14 (50.42) -0.47 (0.39) -1.21 (620.02) -1.53 (0.83) -1.84 (50.95) 0.35 (2.99) 0.12 (617.15) -0.01 (0.03) -0.49 (617.90) 0.02 (0.02) 0.84 (609.21) Between person SRM mean 3.40 (1.35) 2.51 (48.70)* 0.44 (0.97) 0.45 (50.97) 5.44 (2.68) 2.03 (49.97)* 8.17 (9.03) 0.91 (50.14) 0.14 (0.10) 1.40 (50.38) 0.16 (0.08) 2.02 (4 8.58)* Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.01 (0.08) 0.18 0.00 0.00 0.00 0.00 0.00 0.00 0.001 (0.001) 1.29 0.001 (0.001 ) 1.76 SRM centered 0.85 (1.08) 0.79 0.00 0.00 3.36 (6.09) 0.55 0.00 0.00 0.00 0.00 0.00 0.00 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total w ake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Correlation is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, although all convergence criteria were satisfied.

PAGE 100

100 Table 5 18. Multilevel model with social rhythmicity variables predicting sleep in older adults Fixed Effects SOL WASO TWT TST NWAK SQR Predictor Variable B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) Within person time 0.13 (0.24) 0.55 (70.86) 0.46 (0.31) 1.50 (82.64) 0.43 (0.61) 0.70 (65.82) 0.53 (0.74 ) 0.72 (69.16) -0.03 (0.01) -2.30 (50.77)* 0.01 (0.01) 1.22 (70.32) SRM centered 0.49 (0.71) 0.70 (577.48) 0.76 (1.22) 0.62 (57.65) 1.00 (2.02) 0.50 (0.62) 0.41 (2.98) 0.14 (42.44) 0.06 (0.04) 1.47 (40.53) 0.02 (0.03) 0.83 (577.49) Betwe en person SRM mean 2.19 (3.04) 1.71 (48.68) 2.01 (4.00) 0.50 (50.45) 0.87 (7.83) 0.11 (49.16) 11.92 (10.29) 1.16 (47.95) 0.45 (0.24) 1.90 (46.68) 0.13 (0.10) 1.35 (50.02) Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.53 (0.44) 1.21 0.20 (0.68) 0.30 5.03 (3.09) 1.63 1.64 (4.33) 0.38 0.003 (0.001) 1.90 0.002 (0.001) 2.29* SRM centered 0.00 0.00 19.73 (12.18) 0.11 45.65 (33.87) 1.35 135.90 (86.34) 1.57 0.01 (0.01) 1.09 0.00 0.00 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rat ing. *Correlation is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not posit ive definite, although all convergence criteria were satisfied.

PAGE 101

101 Table 5 19. Multilevel model with light variables predicting sleep in younger adults Fixed Effects SOL WASO TWT TST NWAK SQR Predictor Variable B(SE) t (df) B (SE) t (df) B(SE) t (df) B( SE) t (df) B (SE) t (df) B (SE) t (df) Within person time 0.31 (0.13) 2.35 (472.46)* 0.20 (0.15) 1.37 (484.62) 0.39 (0.31) 1.29 (476.11) 0.86 (1.17) 0.74 (462.94) 0.01 (0.01) 0.48 (478.06) 0.005 (0.009) 0.53 (110.74) AMlight centered 0.63 (0.31) 2.04 (476.38)* 0.19 (0.36) 0.52 (51.40) 2.06 (0.75) 2.74 (119.74)** 4.10 (2.71) 1.52 (491.25) 0.02 (0.03) 0.56 (45.91) 0.02 (0.02) 0.90 (65.56) PMlight centered 0.26 (0.34) 0.77 (476.93) 0.27 (0.40) 0.66 (86.67) 1.42 (0.79) 1.7 9 (477.74) 0.19 (2.98) 0.06 (492.28) 0.01 (0.03) 0.24 (459.47) 0.01 (0.02) 0.47 (470.63) light duration centered 0.01 (0.01) 1.13 (492.04) 0.01 (0.01) 0.97 (468.51) 0.02 (0.02) 1.04 (480.32) 0.14 (0.07) 2.06 (140.03)* 0.001 (0.001) 0.98 (451.03) 0.000 4 (0.0005) 0.82 (469.70) light intensity centered 0.002 (0.002) 0.79 (474.88) 0.001 (0.002) 0.41 (115.11) 0.01 (0.005) 1.18 (474.89) -0.05 (0.02) -2.67 (413.80)** 0.0001 (0.0002) 0.70 (75.51) 8.59E 5(0.0001) 0.61 (99.11) Between per son AMlight mean 0.61 (1.05) 0.58 (50.87) 0.39 (0.76) 0.52 (60.20) 0.12 (2.15) 0.06 (53.41) 16.58 (5.55) 2.99 (53.73)** 0.02 (0.07) 0.22 (53.34) 0.08 (0.06) 1.49 (50.68) PMlight mean 0.67 (1.07) 0.63 (53.94) 0.21 (0.78) 0.26 (66.5 9) 2.71 (2.18) 1.24 (57.26) 4.76 (5.75) 0.83 (60.97) 0.01 (0.08) 0.07 (58.22) 0.005 (0.06) 0.08 (54.79) light duration mean 0.02 (0.02) 1.05 (52.13) 0.002 (0.01) 0.18 (61.29) 0.05 (0.03) 1.50 (54.76) 0.30 (0.09) 3.48 (54.41)** 0.002 (0.001) 1.32 (54.89) 0.001 (0.001) 1.14 (52.46) light intensity mean 0.003 (0.005) 0.65 (51.54) 0.0002 (0.004) 0.04 (61.12) 0.01 (0.01) 1.49 (54.09) 0.09 (0.03) 3.34 (53.88)** 0.0001 (0.0003) 0.40 (54.97) 0.0003 (0.0003) 1.28 (52.55)

PAGE 102

102 Table 5 19. Continued Random Effects SOL WASO TWT TST NWAK SQR B(SE) Z B (SE) Z B(SE) Z B(SE) Z B (SE) Z B (SE) Z Within person time 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0004 (0.001 ) 0.85 AMlight centered 0.00 0.00 0.29 (0.66) 0.43 1.16 (1.68) 0.69 0.00 0.00 0.001 (0.004) 0.20 0.002 (0.002) 0.80 PMlight centered 0.00 0.00 0.71 (0.82) 0.86 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 light duration centered 0.00 0.00 0.00 0.0 0 0.00 0.00 0.01 (0.02) 0.51 0.00 0.00 0.00 0.00 light intensitycentered 0.00 0.00 2.05E 5 (2.21E -5) 0.93 0.00 0.00 0.00 0.00 4.20E 7 (2.57E 7) 1.63 6.95E 8 (8.58E -8) 0.81 Note: SOL=sleep onset latency, WASO=wake time after sle ep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Correlation is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, although all convergence criteria were satisfied.

PAGE 103

103 T able 5 20. Multilevel model with light variables predicting sleep in older adults Fixed Effects SOL WASO TW T TST NWAK SQR Predictor Variable B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) Within person time 0.21 (0.25) 0.84 (71.15) 0.46 (0.31) 1.46 (539.72) 0.39 (0.63) 0.62 (69.65) 0.80 (0.76) 1.05 (695. 41) 0.03 (0.01) 2.51 (53.18)* 0.01 (0.01) 1.32 (66.71) AMlight centered 0.24 (0.67) 0.36 (45.38) 1.70 (0.78) 2.20 (547.99)* 1.55 (1.53) 1.01 (40.36) 5.13 (2.18) 2.36 (1098.27) 0.04 (0.03) 1.51 (533.98) 0.03 (0.02) 1.55 (33.46) PMlight centered 0 .49 (0.59) 0.83 (533.94) 0.78 (0.99) 0.79 (77.75) 0.46 (1.81) 0.26 (74.13) 4.22 (2.23) 1.89 (74.04) 0.0001 (0.04) 0.003 (67.14) 0.005 (0.03) 0.18 (68.07) light duration centered 0.01 (0.01) 0.61 (569.51) -0.04 (0.02) -2.21 (580.29)* 0.03 (0.03) 0.9 6 (570.53) 0.16 (0.04) 3.69 (969.90) 0.001 (0.001) 1.89 (548.88) 0.0002 (0.0004) 0.51 (569.51) light intensity centered 0.004 (0.004) 1.02 (39.77) 0.01 (0.005) 1.20 (550.42) 0.01 (0.01) 1.07 (52.44) 0.03 (0.01) 2.29 (638.82) 0.0001 (0.0002) 0.90 (535.17) 8.71E 5 (0.0001) 0.67 (496.89) Between person AMlight mean 0.33 (1.97) 0.17 (51.44) 2.94 (2.49) 1.18 (55.24) 6.62 (4.94) 1.34 (50.66) 7.43 (6.52) 1.14 (52.46) 0.18 (0.15) 1.22 (48.04) 0.06 (0.06) 0.92 (54.82) PMlight m ean 0.87 (1.80) 0.49 (51.06) 2.34 (2.27) 1.03 (54.63) 5.22 (4.51) 1.16 (50.33) 8.91 (5.95) 1.50 (52.04) 0.13 (0.14) 0.90 (47.95) 0.14 (0.06) 2.37 (54.31)* light duration mean 0.4 (0.03) 1.55 (48.76) 0.01 (0.03) 0.18 (52.21) 0.03 (0.07) 0.40 ( 48.18) 0.22 (0.09) 2.41 (49.40)* 0.002 (0.002) 0.73 (47.10) 3.61E 5 (0.001) 0.04 (50.68) light intensity mean 0.03 (0.01) 2.29 (47.67)* 0.008 (0.02) 0.55 (50.34) 0.04 (0.03) 1.38 (47.12) -0.07 (0.04) -1.82 (48.69) -0.0003 (0.001) -0.37 (46.74) -0.001 (0. 0004) -1.38 (49.94) Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.57 (0.48) 1.19 0.00 0.00 3.76 (2.95) 1.27 0.00 0.00 0.002 (0.001) 1.26 0.001 (0.001) 1.58 AMlight centered 3.78 (3.15) 1.20 0.00 0.00 12.61 (15.70) 0.80 0.00 0.00 0.00 0.00 0.002 (0.003) 0.54 PMlight centered 0.00 0.00 11.97 (5.96) 2.01* 45.56 (20.68) 2.20* 47.75 (30.11) 1.59 0.02 (0.01) 2.34* 0.007 (0.004) 1.68 light duration centered 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 light intensity centered 3.91E 5 0.39 0.00 0.00 0.001 (0.001) 1.11 0.00 0.00 0.00 0.00 0.00 0.00 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Correlation is significant at the p < 0.05 (2tailed), ** c orrela tion is significant at the p < 0.01 (2tailed),*** c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, although all convergence criteria were satisfied.

PAGE 104

104 Table 521. Multilevel model with arousal variables predicting sleep in young er adults Fixed Effects SOL WASO TWT TST NWAK SQR Predictor Variable B(SE) t (df) B(SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) Within person ti me 0.08 (0.10) 0.85 (604.20) 0.08 (0.12) 0.67 (582.53) 0.12 (0.26) 0.49 (613.51) 1.73 (1.05) 1.65 (130.20) 0.0004 (0.01) 0.04 (112.48) 0.01 (0.01) 1.97 (145.51)* arousal grid centered 0.56 (0.49) 0.72 (19.85) 0.73 (0.53) 1.38 (65.40) 1.10 (0.95) 1 .15 (50.05) 1.17 (3.74) 0.31 (50.85) 0.003 (0.04) 0.08 (64.61) 0.01 (0.03) 0.41 (55.64) PSAS somatic centered 0.35 (0.49) 0.72 (19.85) 0.67 (0.74) 0.91 (23.84) 0.91 (1.17) 0.78 (17.74) 0.49 (4.94) 0.10 (22.82) 0.01 (0.04) 0.27 (16.21) 0.09 (0.03) 3 .26 (570.53)** PSAS cognitive centered 1.11 (0.22) 4.93 (45.44)*** 0.87 (0.35) 2.47 (23.09)* 2.09 (0.51) 4.11 (35.37)*** -4.47 (1.36) -3.28 (549.78)** 0.06 (0.02) 3.05 (35.35)** -0.05 (0.01) -4.69 (33.80)*** UMACL R energetic centered 0.24 (0.15) 1.66 (7 1.22) 0.15 (0.14) 1.04 (559.82) 0.18 (0.38) 0.47 (68.08) 0.90 (1.23) 0.73 (54.44) 0.01 (0.01) 0.62 (52.41) 0.03 (0.01) 2.89 (62.42)** UMACL R tense centered 0.05 (0.15) 0.33 (56.36) 0.23 (0.22) 1.03 (39.15) 0.34 (0.37) 0.91 (44.56) 1.97 (1.32) 1.50 (487.05) 0.02 (0.02) 1.44 (49.37) 0.01 (0.01) 1.22 (58.80) Between person arousal grid mean 5.61 (2.31) 2.43 (49.92)* 0.15 (1.73) 0.09 (49.65) 7.69 (4.84) 1.59 (49.62) 29.65 (15.17) 1.95 (49.05) 0.24 (0.18) 1.33 (49.76) 0.21 (0.12) 1. 78 (47.92) PSAS somatic mean 0.02 (1.46) 0.02 (49.86) 1.40 (1.09) 1.29 (49.57) 3.42 (3.05) 1.12 (49.54) 17.50 (9.57) 1.83 (49.25) 0.04 (0.11) 0.33 (49.72) 0.07 (0.07) 0.92 (48.01) PSAS cognitive mean 1.74 (0.50) 3.47 (50.14)** 0.64 (0.38) 1.71 (50.28) 2 .37 (1.05) 2.26 (49.96)* 0.008 (3.30) 0.003 (49.98) 0.06 (0.04) 1.44 (50.28) 0.02 (0.03) 0.86 (48.38) UMACL R energetic mean 2.68 (1.07) 2.50 (49.95)* 0.24 (0.80) 0.30 (49.75) 2.81 (2.24) 1.26 (49.67) 9.74 (7.03) 1.39 (49.23) 0.09 (0.08) 1.11 (49.74) 0.11 (0.05) 2.04 (47.93)* UMACL R tense mean 1.21 (0.43) 2.83 (50.01)** 0.30 (0.32) 0.94 (49.76) 1.72 (0.90) 1.91 (49.72) 1.11 (2.82) 0.40 (49.19) 0.03 (0.03) 0.80 (50.003) 0.03 (0.02) 1.62 (48.39)

PAGE 105

105 Table 5 21. Continued Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.00 0.00 0.00 0.00 0.00 0.00 2.07 (6.05) 0.34 0.001 (0.001) 1.32 0.0003 (0.0003) 1.00 arousal grid centered 0.06 (0.64) 0.10 4.19 (1.93) 2.17* 4.12 (6.33) 0.65 45.63 (94.64) 0.48 0.004 (0.008) 0.49 0.01 (0.01) 1.57 PSAS somatic centered 2.03 (1.97 ) 1.03 8.79 (4.94) 1.78 8.51 (11.34) 0.75 201.16 (188.02) 1.07 0.001 (0.01) 0.11 0.00 0.00 PSAS cognitive centered 1.22 (0.41) 2.95** 3.79 (1.55) 2.45* 5.14 (2.29) 2.24* 0.00 0.00 0.01 (0.003) 1.91 0.001 (0.001) 0.76 UMACL R energetic centered 0.33 (0.15) 2.17* 0.00 0.00 2.24 (1.07) 2.09* 1.49 (10.23) 0.15 0.003 (0.002) 1.62 0.0002 (0.001) 0.36 UMACL R tense centered 0.30 (0.19) 1.62 1.02 (0.49) 2.08 1.14 (1.19) 0.96 0.00 0.00 0.005 (0.002) 1.90 0.002 (0.001) 1.90 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Correlation is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orre lation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, although all convergence criteria were satisfied.

PAGE 106

106 Table 5 22. Multilevel model with arousal variables predicting sleep in old er adults Fixed Effects SOL WASO TWT TST NWAK SQR Predictor Variable B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) Within person time 0.07 (0.23) 0.31 (83.17) 0.35 (0.31) 1.15 (71.62) 0.36 (0.61) 0.59 (65.04) 0.34 (0.74) 0.46 (62.98 ) 0.02 (0.02) 1.66 (49.67) 0.02 (0.001) 2.25 (62.28)* arousal grid centered 0.20 (0.85) 0.24 (1125.78) 1.00 (1.30) 0.77 (27.23) 0.77 (1.80) 0.43 (551.30) 1.30 (2.44) 0.53 (532.21) 0.06 (0.04) 1.46 (26.59) 0.04 (0.03) 1.15 (33.39) PSAS somatic ce ntered 0.69 (1.20) 0.57 (3185.31) 1.27 (1.50) 0.85 (30.17) 1.46 (2.33) 0.63 (21.47) 0.89 (3.52) 0.25 (21.34) 0.01 (0.05) 0.18 (26.56) 0.05 (0.04) 1.29 (18.58) PSAS cognitive centered 0.94 (0.36) 2.61 (33.19)* 2.11 (0.70) 3.04 (50.85)** 3.55 (1.00) 3. 56 (40.00)* -3.80 (1.53) -2.48 (41.83)* 0.08 (0.02) 4.81 (33.81)*** -0.09 (0.02) -5.25 (35.01)*** UMACL R energetic centered 0.09 (.29) 0.32 (590.36) 0.82 (0.48) 1.72 (46.12) 0.14 (0.82) 0.17 (66.57) 0.91 (1.19) 0.76 (49.75) 0.01 (0.01) 0.77 (317.7 7) 0.01 (0.01) 0.93 (38.88) UMACL R tense centered 0.12 (0.29) 0.41 (559.93) 0.11 (0.41) 0.27 (420.95) 0.87 (0.76) 1.15 (44.68) 0.35 (1.13) 0.31 (38.42) 0.03 (0.01) 1.70 (55.18) 0.001 (0.01) 0.14 (427.39) Between person arousal grid me an 0.53 (1.78) 0.30 (46.43) 1.64 (2.28) 0.72 (47.60) 4.06 (4.48) 0.91 (47.34) 5.22 (6.10) 0.86 (47.27) 0.17 (0.14) 1.19 (47.15) 0.14 (0.05) 2.96 (46.45)** PSAS somatic mean 1.18 (2.03) 0.58 (46.48) 1.42 (2.60) 0.55 (48.04) 1.69 (5.09) 0.33 (47.13) 1.51 (6.95) 0.22 (47.76) 0.08 (0.16) 0.49 (47.01) 0.01 (0.05) 0.18 (45.25) PSAS cognitive mean 1.78 (0.75) 2.38 (46.49)* 2.58 (0.96) 2.68 (48.02)* 5.07 (1.88) 2.69 (47.15)* 4.91 (2.58) 1.91 (47.73) 0.09 (0.06) 1.40 (47.09) 0.07 (0.02) 3.41 (45.85) *** UMACL R energetic mean 0.73 (1.68) 0.43 (46.44) 0.87 (2.15) 0.40 (47.85) 2.99 (4.21) 0.71 (47.12) 7.90 (5.75) 1.37 (47.54) 0.01 (0.14) 0.07 (47.03) 0.003 (0.05) 0.07 (45.07) UMACL R tense mean 0.10 (1.72) .06 (46.46) 1.27 (2.20) 0.58 (47.89 ) 1.18 (4.30) 0.28 (47.13) 3.84 (5.88) 0.65 (47.59) 0.12 (0.14) 0.85 (47.07) 0.01 (0.05) 0.20 (45.11)

PAGE 107

107 Table 5 22. Continued Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.53 (0.38) 1.42 0.42 (0.69) 0.61 5.03 (3.03) 1.66 2.20 (4.33) 0.51 0.002 (0.001) 1.69 0.001 (0.001) 2.07* arousal grid centered 0.00 0.00 23.61 (17.14) 1.38 0.00 0.00 0.00 0.00 0.02 (0.02) 1.31 0.02 (0.01) 1.65 PSAS somatic centered 0.00 0.00 32.94 (16.59) 1.99* 59.42 (41.71) 1.43 170.28 (106.58) 1.60 0.03 (0.02) 1.84 0.02 (0.01) 1.62 PSAS cognitive centered 0.68 (0.84) 0.81 8.72 (3.38) 2.58** 8.68 (6.39) 1.36 31.68 (16.60) 1.91 0.001 (0.002) 0.78 0.004 (0.002) 1.96* UMACL R energetic centered 0.00 0.00 1.69 (1.45) 1.67 3.76 (3.78) 1.00 11.36 (9.17) 1.24 0.00 0.00 0.001 (0.001) 1.19 UMACL R tense centered 0.00 0.00 0.00 0.00 0.35 (3.33) 0.10 6.00 (8.22) 0.73 0.001 (0.001) 0.86 0.00 0.00 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Corr elation is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definit e, although all convergence criteria were satisfied.

PAGE 108

108 Table 5 23. Multilevel model with affect variables predicting sleep in young er adults Fixed Effects SOL WASO TWT TST NWAK SQR Predictor Variable B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t ( df) B (SE) t (df) B (SE) t (df) Within person time 0.25 (0.11) 2.26 (93.38)* 0.02 (0.13) 0.12 (614.36) 0.22 (0.27) 0.82 (617.79 1.67 (1.02) 1.63 (610.63) 0.01 (0.01) 0.67 (112.85) 0.003 (0.008) 0.40 (134.46) pleasant grid centered 0.58 (0.35) 1.67 (44.16) 0.58 (0.50) 1.15 (45.60) 1.37 (0.81) 1.70 (604.22) 2.53 (3.93) 0.64 (49.71) 0.04 (0.04) 1.07 (33.60) 0.13 (0.03) 4.84 (53.26)*** PANAS positive centered 0.16 (0.09) 1.76 (54.74) 0.02 (0.11) 0.18 (46.33) 0.24 (0.22) 1.10 (60.13) 0.31 (0.76) 0.40 (42.45) 0.01 (0.01) 1.19 (44.29) 0.02 (0.01) 2.62 (54.13)* PANAS negative centered 0.15 (0.19) 0.80 (44.69) 0.66 (0.28) 2.32 (32.19) 0.87 (0.42) 2.07 (22.59)* 2.47 (1.47) 1.68 (610.66) 0.03 (0.02) 1.25 (45.26) 0.0 1 (0.01) 0.94 (41.53) Between person pleasant grid mean 1.42 (1.49) 0.95 (48.87) 0.47 (1.01) 0.46 (50.75) 4.82 (2.86) 1.68 (49.84) 28.14 (7.79) 3.61 (49.58)** 0.03 (0.10) 0.31 (50.52) 0.14 (0.07) 2.15 (49.75)* PANAS p ositive mean 0.13 (0.20) 0.65 (48.17) 0.18 (0.13) 1.31 (49.58) 0.65 (0.38) 1.70 (49.04) 1.29 (1.04) 1.25 (48.32) 0.004 (0.01) 0.28 (49.99) 0.003 (0.01) 0.34 (48.94) PANAS negative mean 0.24 (0.52) 0.47 (49.68) 0.07 (0.36) 0.19 (52.29) 0.08 (1.01) 0.0 8 (50.75) 1.92 (2.75) 0.70 (51.19) 0.04 (0.04) 0.99 (50.89) 0.05 (0.02) 2.22 (50.79)* Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.01 (0.08) 0.15 0.00 0.00 0.00 0.00 0.00 0.00 0.001 (0.001) 1.06 0.003 (0.003) 0.93 pleasant grid centered 0.52 (0.93) 0.56 3.95 (2.23) 1.78 0.00 0.00 224.17 (128.72) 1.74 0.01 (0.01) 1.04 0.01 (0.01) 1.73 PANAS positive mean 0.08 (0.06) 1.38 0.10 (0.09) 1.0 9 0.42 (0.33) 1.27 1.78 (4.54) 0.39 0.001 (0.001) 1.21 0.0003 (0.0003) 1.29 PANAS negative mean 0.35 (0.25) 1.42 1.69 (0.81) 2.09* 1.01 (1.65) 0.61 0.00 0.00 0.01 (0.004) 2.30 0.0002 (0.0007) 0.28 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Correlation is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, although all convergence criteria were satisfied.

PAGE 109

109 Table 5 24. Multilevel model with affect variables predicting sleep in old er adults Fixed Effects SOL WASO T WT TST NWAK SQR Predictor Variable B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) B (SE) t (df) Within person time 0.19 (0.22) 0.87 (56.97) 0.40 (0.33) 1.24 (79.80) 0.40 (0.58) 0.69 (61.09) 0.50 (0.75) 0.67 (70.3 8) 0.03 (0.01) 2.28 (51.96)* 0.01 (0.01) 1.39 (80.36) pleasant grid centered 0.88 (0.79) 1.11 (21.51) 2.71 (1.19) 2.29 (19.56)* 3.81 (1.58) 2.41 (16.36)* 5.50 (2.16) 2.55 (7.49)* 0.06 (0.03) 1.96 (23.35) 0.13 (0.03) 4.08 (25.04)*** PANAS posi tive centered 0.11 (0.24) 0.44 (17.70) 0.20 (0.37) 0.55 (13.86) 0.02 (0.58) 0.04 (26.79) 0.95 (0.87) 1.09 (28.50) 0.02 (0.01) 1.55 (27.54) 0.02 (0.008) 1.95 (25.43) PANAS negative centered 0.09 (0.38) 0.23 (561.03) 0.63 (0.55) 1.15 (557.18) 1.0 4 (0.94) 1.11 (541.61) 1.88 (1.31) 1.43 (22.59) 0.0004 (0.02) 0.02 (558.45) 0.004 (0.01) 0.30 (579.75) Between person pleasant grid mean 2.12 (2.52) 0.84 (49.17) 7.07 (3.14) 2.25 (50.40)* 9.79 (6.04) 1.62 (49.72) 20.14 ( 8.25) 2.44 (47.91)* 0.28 (0.20) 1.44 (47.69) 0.17 (0.07) 2.44 (49.83)* PANAS positive mean 0.01 (0.41) 0.02 (48.54) 0.24 (0.52) 0.47 (49.70) 0.53 (0.99) 0.54 (49.37) 2.94 (1.35) 2.17 (47.05)* 0.03 (0.03) 1.03 (47.63) 0.01 (0.01) 0.56 (49.54) PANAS negative mean 0.16 (0.45) 0.37 (50.23) 0.56 (0.55) 1.02 (51.52) 1.11 (1.06) 1.04 (50.15) 2.51 (1.46) 1.72 (49.42) 0.07 (0.03) 2.06 (47.71)* 0.01 (0.01) 0.75 (49.91) Random Effects SOL WASO TWT TST NWAK SQR B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z B (SE) Z Within person time 0.11 (0.44) 0.26 0.26 (0.73) 0.36 6.68 (2.93) 0.92 0.57 (4.60) 0.12 0.002 (0.001) 1.61 0.001 (0.001) 0.12 pleasant grid centered 6.46 (5.79) 1.12 16.99 (14.13) 1.20 3.13 (19.65) 0.16 6.18 ( 54.71) 0.11 0.01 (0.01) 0.74 0.02 (0.01) 1.65 PANAS positive mean 0.80 (0.77) 1.04 2.28 (2.11) 1.08 4.10 (3.51) 1.17 13.10 (8.24) 1.59 0.001 (0.001) 1.03 0.001 (0.001) 0.96 PANAS negative mean 0.00 0.00 0.00 0.00 0.00 0.00 1.47 (9.50) 0.16 0.00 0.00 0.00 0.00 Note: SOL=sleep onset latency, WASO=wake time after sleep onset, TWT=total wake time, TST=total sleep time, NWAK=number of awakenings, SQR=sleep quality rating. *Correlati on is significant at the p < 0.05 (2tailed), ** c orrelation is significant at the p < 0.01 (2tailed),***c orrelation is significant at the p < 0.001 (2tailed), Variance is too small to be estimated; the final Hessian matrix was not positive definite, al though all convergence criteria were satisfied.

PAGE 110

110 CHAPTER 5 DISCUSSION Review of the Findings While prior research has shown light to be the dominant zeitgeber influencing circadian rhythms, recent research has examined the influence of social activities. The regularity or irregularity of social rhythms has been linked to sleep in both younger and older adults. Despite the interest in the association between social rhythms and sleep, few studies have investigated the role of specific mechanisms in mediating this relationship. This study expanded on prior research by assessing the role of light, arousal, and affect in mediating the relationship between social rhythmicity and sleep. Given the circadian, or 24hour rhythm of sleep patterns, this study inve stigated sleep on a daily basis. In order to justify examining the relationship between the predictor variables and sleep on a daily basis within individuals, the constructs had to possess sufficient within person variability. Once the amount of within person variability was established, the relationship between variability and sleep outcomes was analyzed. Next, the daily coupling of the predictor variables and sleep was evaluated. Finally, the role of light, arousal, and affect as potential mediators o f the relationship between social rhythmicity and sleep was investigated. Each of these goals was examined in both younger and older adults. The results of each specific aim will be reviewed and interpreted separately, followed by consideration of the st rengths and major limitations of the study, discussion of the implications of the study (both theoretical and applied), and discussion of future directions. Aim 1: Identify the Extent to Which Social Rhythmicity, Light, Arousal, Affect, and Sleep Fluctuate Daily both Within and Between Persons Within person variability or intraindividual variability has become increasingly of interest in the social sciences. In fact, a recent issue of Psychology and Aging devoted a special section to the study of Intraindi vidual Variability and Aging. Intraindividual variability warrants

PAGE 111

111 attention given that our primary object of study humans are dynamic with changing and sometimes unpredictable behaviors (Ram, Lindenberger, & Blanchard Fields, 2009) Consequently, the use of relatively static representations of behaviors (e.g., one time of measurement) may not capture the full phenomena of human activity (West, 1985) Aim 1 of the study extracted the amount of between and within person variability in the constructs of interest. The importance of aim 1 is two fold. First, a minimum amount of within person variability must be present to warrant conducting the analyses for aim 2 (examining fluctuations over time). Second, intraindividual variability in general, and age related differences in intraindividual variability in particular, have not been investigated for many of the variables studied. Documenting the presence of intraindividual variability is an important first step in validating the use of intraindividual variability analyses in the social sciences (Boker, Molenaar, & Nesselroade, 2009) Boker, Nesselroade, and Molenaar (2010) commented on the need to focus on the time dynamics of variability once the amount of variability has been established. Consequentl y, the analyses of aim 1 contribute to the foundational literature in intraindividual variability by documenting the amount of variability present in social rhythmicity, light, arousal, affect, and sleep in younger and older adults. For social rhythmicity, two estimates of variability were generated. The first traditional estimate is an overall estimation of the variability of the individual that is based on two weeks of data. Younger adults, on average, had a social rhythmicity score of 2.36 while older adults on average scored 4.10. The social rhythm scale ranges from 0 to 7 with higher scores indicating more regularity. Therefore, older adults were, overall, more regular than irregular in their daily activities while younger adults were more irregu lar than regular in the timing of daily activities

PAGE 112

112 This result is consistent with previous research that has shown that regularity increases with age (Monk et al., 1994; Monk et al., 1997) The second estimate of variability was the intraindividual variability (the day to day) variation in regularity scores. Both younger and older adults were highly variable in the regularity of their daily activities. Of the total variability to be measured, 77% and 72% was due to intraindividual variations in younger and older adults respectively (younger adults were significantly more variable within person compared to older adults). These results indicate that although older adults are more regular overall, there is still considerable variation day to day in how re gular they are in their daily activities. In fact, older adults vary more in their behaviors within person than they do between person. The differences in the overall and day to day measures of regularity are important as they provide different views of the level of routine in older adults lives. Over a twoweek period, older adults generally appear more regular in the timing of their daily activities compared to younger adults. On a day to day basis, however, an older adults daily routine is more lik ely to resemble the routine of another individual than to resemble their own routine. This variation on a daily basis is considerably less than younger adults but still suggests there is more variation in the timing of daily events for older adults than w ould be seen simply by summing the two weeks of data. Light in the present study was examined subjectively using the Light Exposure Scale. The Light Exposure Scale was designed specifically for this study due to the lack of subjective measures for ass essing light. While the Light Exposure Scale has apparent face validity in terms of the measures of timing, duration, and intensity of light exposure, there are no established reliability or validity estimates for the measure. Consequently, the results must be interpreted with caution and have limited generalizability. Intraindividual variability in light has not, to our

PAGE 113

113 knowledge, previously been examined. Four aspects of light were examined: the hour of first exposure to light in the day, the hour of l ast exposure to light in the day, the duration of exposure to light, and the intensity of the exposure. Overall, there was considerable within person variability in light exposure with all light measures showing almost as much or more variability within than between persons. The most variability was seen for the timing of morning and evening light exposure for younger adults (74% for both variables) compared to the amount of light intensity and duration of exposure (53% and 57%). These results indicate t hat younger adults are extremely variable in the time at the start and the end of the day when they are first exposed to light. T here was similar variability across all four light measures for older adults (48% to 57%). Older adults were less variable wi thin person than younger adults in terms of the light variables. These results provide new information about the amount of variability within individuals in terms of their light exposure. The meanlevels of the light variables suggest that older adults received light earlier in the day and more intense light exposure overall. The finding that older adults are exposed to light earlier in the day is consistent with previous research findings that found older adults were exposed to higher light levels in t he morning compared to younger adults (Kawinska, Dumont, Selmaoui, Paquet, & Carrier, 2005; Scheuermaier, Laffan, & Duffy, 2010) Earlier light exposure for older adults could be explained by previous research documenting a trend towards morningness or earlier rising/alertness among older adults (Duffy, Rimmer, & Czeisler, 2001; Taillard, Philip, & Bioulac, 1999) Older adults are more likely to have morning versus evening chronotypes and consequently have greater opportunity to be exposed to dayli ght earlier in the day. In terms of the intensity of the light exposure, the greater light exposure for older adults runs counter to some previous studies (Jean -

PAGE 114

114 Louis et al., 2000; Kawinska et al., 2005; Mishima, Okawa, Shimizu, & Hishikawa, 2001) which f ound older adults had less or equal levels of light exposure compared to younger adults. Conversely, a minority of studies have shown older adults to have higher levels of light exposure compared to younger adults (Campbell, Kripke, Gillin, & Hrubovcak, 1988; Scheuermaier et al., 2010) It is possible that the use of a healthy, community dwelling sample of older adults resulted in a sample that was more active than previous studies with less healthy, residential older adults. Although, the overall age related differences in the light variables are important for understanding the relationship between age and light exposure they do not tell the complete story. For example, the within person variability in morning light exposure and intensity for younger and older adults was 53% and 49%. Therefore, older adults vary almost as much within person as they do between persons. The addition of intraindividual variability data sheds new light on age related differences in light exposure and adds an important ca veat to assumptions made about age differences based on mean level data. The mean level differences tell us that younger and older adults differ in the timing of their light exposure and the intensity of light received, overall The intraindividual varia bility data tells us that these age related differences do not hold across all older/younger adults and that there is significant variability in the timing and intensity of light received within person for younger and older adults. Overall, younger adult s showed as much or more variability in arousal within than they did between persons (50% to 71%) while older adults showed more between person variability and less variability day to day for four out of the five variables (22% to 53%). Few studies have e xamined variations in arousal on a daily basis within individuals. One recent study examined agerelated differences in variability in the cortisol awakening response (CAR; Almeida, Piazza, & Stawski, 2009). While the implications of the CAR are not completely understood, lower

PAGE 115

115 levels have been associated with burnout symptomatology (Pruessner, Hellhammer, & Kirschbaum, 1999) and heightened levels have been associated with neurotic (Portella, Harmer, Flint, Cowen, & Goodwin, 2005) and depressive symptoms (Bhagwagar, Hafizi, & Cowen, 2003) The authors found that across ages there was more variability within than between persons in cortisol levels and that there was an increase in variability with age for men (Almeida, Piazza, & Stawski, 2009) While the results from the present study use differing measures of arousal the high levels of variability within person for younger adults is consistent with these findings. Less variability for older adults in the present study may reflect the differences in measurement type (e.g., psychological versus physiological measures). Additionally, as cortisol is a hormone tied to the circadian rhythms, it is possible that the age related increased variability in cortisol with age in the Almeida, Piazza, and Stawski (2009) study could reflect a weakening of the circadian rhythm with age resulting in a less robust rhythm that is vulnerable to fluctuations. Possible explanations for the age differences in variability in arousal in the present study are that younger adults may have more variability in their daily routines, for example, engaging in school or work activities that are potentially arousing or anxiety provoking. The variability results from the section on social rhythmicity indicate that at the very least, youn ger adults are more variable in the timing of their daily activities. The one variable that younger and older adults did not differ on for variability was the Pre sleep Arousal Somatic subscale. Older adults were much more variable on this scale than on the other arousal measures. This scale includes physical sensations of arousal including shortness of breath, cold feelings in the extremities, dry mouth, and upset stomach. Older adults could be experiencing transient healthrelated conditions with symptoms that overlap with some of the physical manifestations of anxiety

PAGE 116

116 (older adults rated their health status as poorer than younger adults 3.48 versus 3.84). For mean level differences, older adults reported significantly higher cognitive arousal than younger adults. This finding is interesting given that the cognitive arousal occurred prior to bed and involved endorsing items such as worry about falling asleep, review or ponder events of the day, and being distracted by noises in the environment Higher levels of cognitive arousal for older adults could reflect concern about sleep (60% of the older sample met criteria as poor sleepers) or, conversely, the poorer sleep of the older adult sample could be a reflection of the higher level of cogni tive arousal. Additionally, the endorsement of specific items such as being distracted by noises in the environment could reflect genuine age related changes in sleep such as spending more time in the lighter stages of sleep. The amount of intraindividual variability in affect was greater for younger (39 to 66%) versus older adults (14 to 44%). Less variability in affect was observed for older adults than found in previous studies (e.g., 27% for positive and 28% for negative affect from McCrae et al., 2008) The age related differences are consistent with previous research showing that older adults have less intraindividual fluctuations in affect compared to younger adults (Rocke, Li, & Smith, 2009) Possible explanations for the lower levels of varia bility in affect measures for older adults are that 1) older adults have more stable external environments/lifestyles and are therefore less vulnerable to negative experiences (as evidenced by findings on social rhythmicity in this study and from Bouisson & Swendsen, 2003) ; 2) over time older adults have developed skills to regulate emotions (Carstensen et al., 2000) ; 3) evidence for habituation processes with age (Frijda, 1988) ; and 4) increasing use of life management techniques with age to regulate emotion (Carstensen, Isaacowitz, & Charles, 1999) For example, according to the Socioemotional Selectivity Theory, when time is perceived to be limited (e.g., older age),

PAGE 117

117 individuals are more likely to focus on emotional goals such as experiencing emotional s atisfaction and deriving emotional meaning rather than informational goals such as novel experiences or acquiring new knowledge (Carstensen et al., 1999) An alternative to the idea that emotional regularity increases with age, is that less variability in affect in older adults could be due to symptoms of alexithymia. Alexithymia was first described in 1972 and refers to difficulty identifying feelings, difficulty distinguishing between feelings and physical sensations of emotions, difficulty describing feelings, constricted imaginations, and an externally oriented cognitive style (Nemiah & Sifneos, 1970) It is possible that some of the older adult sample could be experiencing symptoms of alexithymia in part due to generational effects of being raised i n a less psychologically minded time period that did not as actively encourage the identification and expression of emotions. Older adults demonstrated significantly more positive affect compared to younger adults which is consistent with prior researc h demonstrating greater positive affect in older adults (Gross et al., 1997; Kunzmann, Little, & Smith, 2000; Mroczek & Kolarz, 1998; Rocke et al., 2009) Consistent with the above explanations for less intraindividual variability in affect for older adul ts, it is possible that higher levels of positive affect for older adults could reflect the actual experience of more positive affect or reflect a generational bias in older adults towards positive emotional expression and a tendency to contain negative em otions. Importantly, 20% of the overall variability in positive affect could be attributed to within person variability suggesting that the age related differences in positive affect do not necessarily represent stability in positive affect for all older a dults. Therefore, while overall, older adults exhibited more positive affect than younger adults, there is within person variability in these findings suggesting that a portion of older adults will not exhibit greater positive affect than younger adults.

PAGE 118

118 Sleep was the final variable assessed for intraindividual variability. Both younger and older adults were highly variable within person for sleep with more variability seen within person than betweenpersons for all variables except for number of awaken ings for older adults. Few studies have examined the amount of within person variability in sleep. These results are consistent with one study that assessed within person variability in sleep quality in college students (Galambos, Dalton, & Maggs, 2009) Specifically, that study found 69% of the variation in sleep quality was due to within person factors (comparable to 78% in the present study). Another study examined sleep in older and younger adults and found that younger adults were significantly mor e variable in their sleep as measured by a retrospective subjective report (the Pittsburgh Sleep Quality Index) and that the intraindividual variability in total sleep time was 0. 40 for younger adults and 0.28 for older adults (much lower than the variabil ity observed in the present study for younger [ 84% ] and older [ 60% ] adults Ready, Marquez, & Akerstedt, 2009) Importantly, across the study constructs, the greatest variability for older adults w as seen for the sleep variables. Greater variability in s leep for older adults compared to the other constructs could be attributed to age related changes in sleep including an increasing amount of time spent in the lighter stages (1 and 2) of sleep (Ohayon et al., 2004) Greater variability may reflect a great er susceptibility among older adults to awakenings during the night due to the lighter depth of their sleep. While older adults were highly variable on five out of the six sleep variables, they were less variable in the number of awakenings. This discrepancy in variability could be due to age related health conditions such as nocturia which could result in a higher, consistent pattern of nightly awakenings.

PAGE 119

119 The age related mean differences in sleep outcomes suggest that older adults have poorer sleep than younger adults in terms of the amount of time it takes to fall asleep, the amount of time spent awakening after falling asleep and before awakening, the total amount of time spent awake during the night, the total amount of time spent sleeping, and the number of awakenings. In addition to older adults demonstrating statistically significantly poorer sleep, the results are also clinically significant with older adults taking twice as long to fall asleep, spending more than twice as long awake during the night, and waking up twice as often. These results are consistent with previous research showing more time spent awake during the night for older adults (Ohayon et al., 2004) As mentioned in the literature review, poor sleep in older adults is not a function of age per se but rather is multifactorial in origin and due in part to the multiple comorbidities older adults encounter (for example, older adults in the study reported worse health status and more health conditions compared to younger adults). While research indicates that changes in with sleep with age occur earlier than previously thought (e.g., middle aged years; Ohyaon et al., 2004), the majority of the younger adult sample was considerably younger than middle age and likely had not yet encountered significant age related changes in their sleep. Interestingly, despite poorer overall levels of sleep on five out of the six sleep variables, and a higher percentage of poor sleepers compared to younger adults (60% versus 14%; see Tables 4 1 and 42), older adults reported comparable sleep quality to younger adults. This finding is interesting and underlines both the subjective nature of sleep and possible lowered expectations for sleep as individuals age In summary, aim 1 indentified the amount of intraindividual variability in social rhythmicity, light, arousal, affect, and sleep variables. The results indicated that there is considerable within person variability for both younger and older adults. Therefore, the next step

PAGE 120

120 of the analyses, examining day to day associations, was warranted. The importance of aim 1 is the identification of microlevel changes in behavior. Historically, the study of aging has studied changes at the macrolevel (Ram et al., 2009) occurring over years and decades. For example, the study of fluid and crystallized intelligence over the lifetime. As a result, the fluctuations or inconsistencies of short term processes have been overlooked. As demonstrated by the comparison of mean level versus intraindividual level out comes above, often times an age related difference at the mean level is qualified by significant intraindividual differences. Examining behaviors at the microlevel captures the complexity of psychological variables. Interestingly, older adults exhibited considerably less intraindividual variability compared to younger adults on most measures other than sleep. Rocke and colleagues (2009) stressed the importance of examining intraindividual variability in addition to mean level relationships given the ass umption that aging is associated with increased variability. Previous research has demonstrated increases in variability with aging for physiological (Lipsitz, 2002) cognitive (Hultsch, Strauss, & Hunter, 2008) and sensorimotor functioning (Woollacott, 2000) In turn, variab ility has been viewed as indicative of poorer functioning or a fault in the integrity of the system (Nesselroade, 1991) For example, variability in cognitive functioning can be indicative of neurological dysfunction. The view of i ncreasing variability with age, and the view that the increased variability is associated with negative outcomes, is consistent with a loss oriented view of aging. The results from the present study, however, present a more gain oriented scenario where ol der age is associated with less variability. The direction of the present results are consistent with recent findings (Rocke et al., 2009) suggesting that some constructs become or remain more stable with age. The majority of social rhythms, light, arous al, and affect variables

PAGE 121

121 exhibited less variability for older adults compared to younger adults suggesting a stabilizing rather than destabilizing function of age. The higher levels of variability for sleep compared to the other constructs for older adul ts fits with previous findings indicating greater variability in cognitive, sensorimotor, and physiological functioning. It is possible that sleep represents a more physiological process (albeit one that is affected by psychological processes) than social rhythmicity, light exposure, affect, or arousal which are more behavioral or psychological in nature. Consequently, it fits that more variability would be seen for sleep given that sleep also has a strong physiological component. Aim 2: Identify H ow V ar iability in S ocial R hythmicity, L ight, A rousal, and A ffect is R elated t o S leep While aim 1 of the study established the presence of within person variability, aim 2 investigated the positive and negative associations of variability. For example, do indivi duals who are more variability in their social rhythmicity, light exposure, arousal levels, and affect experience better or worse sleep? In general, within person variability was typically associated with negative outcomes for younger adults and was less related to positive or negative outcomes for older adults. In particular, greater variability in social rhythms was associated with worse sleep for younger adults. This result replicates previous findings showing that greater irregularity in social rhyt hms was associated with poor sleep for college students (Carney et al., 2006) In particular, greater fluctuations in the timing of daily activities were associated with a longer time to fall asleep and more time spent awake during the night. It is possi ble that varying the timing of daily activities could 1) affect the timing of the sleep period by undermining the circadian rhythm (hence longer time to fall asleep) and/or 2) increase levels of arousal/negative affect which in

PAGE 122

122 turn results in more wake ti me during the night. Research on motivation has shown that more regular, habitual behavior is associated with less emotional intensity (Wood et al., 2002) Therefore, younger adults who are more irregular in their daily activities may experience greater e motional intensity (arousal or affect) which in turn disrupts sleep. The lack of association of variability in social rhythmicity with sleep for older adults was surprising. As discussed for aim 3 below, it may be the strength rather than the presence of the zeitgeber that is important for entraining daily rhythms. Monk (2010) commented on the increasing difficulty of entraining circadian rhythms with age. Therefore, some activities may be more salient for older adults in terms of their ability to entr ain daily rhythms. The calculation of social rhythms involves an average of all completed daily activities. It may be that for older adults, certain activities have stronger entrainment ability than others. The examination of variability in the timing o f specific activities (see below) provides support for this explanation given that irregularity in specific activities was associated with sleep for older adults. The association between variability in specific daily events and sleep was also examined. G reater fluctuation in the time younger adults got out of bed was associated with worse sleep. This result is consistent with models of insomnia that advocate for regular wake times in order to entrain the circadian rhythm to in turn promote better sleep ( Bootzin, 1972; Spielman, Saskin, & Thorpy, 1987) Additionally, one of the few studies that examined the timing of daily events in relation to sleep found that variability in the out of bed time was associated with worse sleep for younger adults (Carney e t al., 2006) Meal times were also predictive for younger adults with varying breakfast and lunch times associated with worse sleep. For older adults, variability in nap times, time returning home, and time for the evening snack was associated with worse sleep. Sleep hygiene treatment recommendations often

PAGE 123

123 recommend restricting nap behavior and, if it is necessary to nap, selecting a consistent nap schedule. Napping can meet some of an individuals sleep need and consequently, disrupt nighttime sleep. B y napping at inconsistent times, older adults may be preventing their daily rhythms from adjusting to the disruption in sleep drive caused by the nap. Varying the return to home time could have an arousing effect by disrupting the daily schedule and short ening the time to wind down before bed. Fluctuations in the timing of evening snacks could negatively affect sleep by causing the digestive system to metabolize an unaccustomed late snack or by causing an alerting effect by engaging in a social interactio n during the snack. Variability in the timing of lunch and work was associated with better sleep for older adults. Both these activities occur towards the beginning to middle of the day. Possibly varying the timing of these events has a stimulating or arousing effect that activates older adults (thereby increasing activity levels during the day and interfering with napping behavior). Given the timing of these events, older adults would have enough time to wind down during the day and the extra stimulation/activation during the day could increase feelings of tiredness at night (hence less time to fall asleep). Interestingly, other than the timing of lunch, differing daily activities were associated with sleep for younger and older adults. The age di screpancies in results suggest that there may be agerelated differences in the ability of different daily events to entrain rhythms in younger and older adults. It may be that older and younger adults differ in the variability of timing of different even ts. For example, meals may occur at more regular times for older adults (perhaps due to the timing of medication or medical conditions requiring regular meals) while work may occur at more regular times for younger adults. The differing levels of variabil ity for discrete events may affect the relationship between variability in these events and sleep. Additionally,

PAGE 124

124 specific events may have stronger entrainment potential for the different age groups. Arriving at home may signal a winding down of the day for older adults, whereas for younger adults, it could signal the start of the next phase of activity. Variability in the timing of exposure to light in the mornings was associated with worse sleep for younger adults. Given that photic influences are t he strongest at entraining the circadian rhythm, it is not surprising that varying the timing of exposure to light could disrupt the circadian rhythm. Variability in light exposure was not significantly associated with sleep outcomes for older adults. Consistent with the hypothesis that it is the strength of the zeitgeber that is influential rather than the presence/absence (Monk, 2010) it is possible that the variations in light were not significant enough in magnitude to influence the sleep of older adults. Additionally, due to a reduction in pupil diameter, as well as a yellowing of the lens with age, the amount of light that reaches the retina in older adults may be attenuated to begin with (Hughes & Neer, 1981; Teresi et al., 1994) As a result, the strength of variations in light exposure may have to be even greater for older adults to compensate for age related changes. Fluctuations in arousal for all arousal measures for younger adults and for two measures for older adults were associated with worse sleep. Previous research has shown that overall higher levels of arousal can have a disrupting effect on sleep (Aschoff, 1965; Espie et al., 1989; Nicassio et al., 1985) These results are interesting because they suggest that in addition to highe r levels of arousal being associated with worse sleep, fluctuations in the level of arousal are also associated with poorer sleep. Consequently, it may be helpful to not only reduce the amount of arousal, but also to aim for more stable levels of arousal. Variability in affect was associated with worse sleep for younger but not older adults. Given that older adults showed the least intraindividual variability for the affect variables, it is

PAGE 125

125 not surprising that variability was not associated with sleep f or older adults. The results for younger adults are interesting given that in addition to higher levels of negative affect being maladaptive, inconsistencies in affect levels can also be maladaptive in terms of sleep outcomes In summary, higher levels of variability were associated with negative sleep outcomes for younger and older adults except for the timing of work and lunch for older adults. Importantly, given the cross sectional study design, we cannot assume directionality for these relationships. Aim 3 : Identify to What Extent the Daily Fluctuations in Social Rhythmicity, Light, Arousal, and Affect are Coupled with Daily Fluctuations in Sleep The goal of the third aim of the study was to examine how the predictors of social rhythmicity, light, arousal, and affect were coupled with daily fluctuations in sleep. Given that the predictor variables occurred earlier in the day than sleep, the direction of the relationships are discussed such that social rhythmicity, light, arousal, and affect are p redict sleep. While causation cannot be discerned due to the cross sectional study design, we can assume a temporal relationship between the predictors and sleep variables given that the behaviors measured by the predictor variables occurred earlier in the day than sleep. The results indicated that mean level social rhythmicity predicted sleep for younger adults with greater overall regularity associated with less time to fall asleep, more total time spent awake, and poorer sleep quality overall. There w ere no significant associations between regularity and sleep for older adults. The findings related to younger adults support previous research suggesting that greater regularity in daily activities is associated with better sleep (Carney et al., 2006; Monk et al., 2003) The lack of a day to day or level 1 association between social rhythmicity and sleep could be explained by the fact that regularity is not a construct suited to daily measurement. The social rhythmicity scale produces a mean level variable derived from multiple days of observation. From this meanlevel variable, we were able to

PAGE 126

126 calculate a daily variable. It is possible that a daily value of social rhythmicity is less meaningful than a summary variable. Conversely, it is possible that the lack of a day to day relationship between social rhythmicity and sleep reflects the idea that the regularity or irregularity of younger adults activities is not as impactful on sleep on a day to day basis. For example, a critical threshold may exist whereby irregularity in activities over a one day period is not dysregulating for sleep, but habitual irregularity disrupts sleep. Entrainment of circadian rhythms by social cues may require prolonged/repeated exposure to the cue. Therefore, higher levels of variability over an extended period (e.g., two weeks) may sufficiently disrupt an individuals circadian rhythms whereas irregularity on a daily basis is not sufficient to disrupt already entrained rhythms. The lack of association between social rh ythmicity and sleep for older adults is puzzling given previous research showing greater regularity is associated with better sleep in older adults (Brown et al., 1996; Zisberg, Gur Yaish, & Shochat, 2010) Possible explanations for the lack of findings a re 1) the lower levels of regularity in the older sample compared to previous studies and 2) omission of aspects of regularity relevant for regulating sleep. First, the social rhythmicity score of this sample of older adults (4.10 out of 7) was lower than the previously reported mean scores for older adults of 4.75 (Zisberg et al., 2010) and 4.40 (Monk et al., 1992) Given age related declines in endogenous circadian entrainment mechanisms the lower levels of regularity in the older adult sample may not have been sufficient to entrain the circadian system and consequently see an association with sleep. Monk (2010) made an interesting recommendation regarding the need to study zeitgeber enhancement. He suggests that it may not be the regularity/irregular ity or presence/absence of social cues that serve to entrain circadian systems, rather the strength of the entraining abilities of the zeitgeber that is of importance

PAGE 127

127 (Monk, 2010) For example, brighter light levels are associated with better sleep in nur sing home residents (Ancoli Israel, Martin, Kripke, Marler, & Klauber, 2002) By adopting a highly regular lifestyle, older adults may be able to entrain their circadian system but simply increasing regularity may not be sufficient. Consequently, the lac k of association between social rhythmicity and sleep for older adults in this sample may be due to the lower levels of regularity in the sample which, in turn, decreased the strength of social activities as zeitgebers. Second, Zisberg and colleagues (2010) recently commented on the need to assess multiple facets of regularity which could affect the timing of circadian rhythms. The authors suggest that in addition to the timing of daily activities (as measured by the Social Rhythm Metric), the frequency, duration, and weekly patterns of activities are important aspects of daily activity patterns. By examining only the timing of daily activities in the present study, the regularizing effects of the frequency, duration, and presence of weekly activities ma y have been overlooked. Aim 3 for the light variables examined the day to day and meanlevel associations between light and sleep. Importantly, as mentioned above, due to the lack of established reliability or validity estimates for the light measure, the light results should be interpreted as exploratory. For younger adults, days with earlier exposure to light, exposure to light for longer periods of time, and days with exposure to less intense light were associated with better sleep that night. There was a significant within person fixed effect for time indicating that younger adults time to fall asleep decreased over the course of the study. This effect could be attributed to the effects of daily measurement, a Hawthorne effect, whereby behaviors change/improve as a result of being observed. At the betweenpersons level, individuals who were exposed to light later in the day, for longer duration, and to less intense light, overall, experienced better sleep at night. These results reflect the role of light as the primary, most dominant zeitgeber influencing our circadian

PAGE 128

128 rhythms (Mistlberger & Skene, 2004) The discrepant findings at the within and between person levels underlines the importance of investigating lightsleep relationships at a daily in addition to a mean level. For example, later exposure to light during the day was associated with worse sleep that night but associated with longer total sleep time overall during the study. Examination of mean level relationships allows for the stud y of cumulative effects of later day light exposure while the day to day analyses permit the study of the associations with sleep that night. Both results suggest that later day light exposure is significantly related to sleep, albeit by different mechani sms. The pattern of results for light exposure in older adults suggests significant associations at both the within and between person levels. On a daily basis, exposure to light later in the day and for shorter duration was associated with worse sleep that night. At the between persons level, more intense, longer duration, and later in the day exposure to light, over the two weeks, was associated with better sleep overall. Overall, the findings for light duration are consistent at both the within and be tween person levels. For the other light variables, however, the analyses at each level provided unique information. The finding that exposure to light later in the day is associated with better sleep quality overall for older adults is interesting. One age related change in sleep is an advancing of the circadian rhythm ( become sleepier earlier; Monk, 2005) Additionally, older adults are more likely to complain of early morning awakening insomnia as opposed to difficulty falling sleep. It is possible that exposure to light later in the evening results in a phase delay in the sleep of older adults, consequently, delaying the bedtime and dealying the sleep period so that there are less difficulties with early morning awakenings. The results for the r elationship between arousal and sleep for younger and older adults indicated that arousal and sleep are linked both on a day to day basis and at an overall, mean -

PAGE 129

129 level. Cognitive arousal before bedtime as measured by the Pre sleep Arousal Scale cognitive subscale was highly predictive of sleep for both younger and older adults. On days when younger and older adults had more cognitive arousal before bed, they experienced poorer sleep. This effect was qualified by random effects for both age groups suggest ing that there are significant individual differences in the magnitude of this relationship. The finding that cognitive arousal predicts sleep on a daily basis is consistent with prior research. The fact that cognitive arousal so widely impacts sleep (ac ross multiple measures of sleep) and effects both age groups can be explained by Harveys model of insomnia (2000). According to Harveys model, cognitive activity leads to autonomic arousal. Not only does the increased arousal disrupt sleep, but it also triggers a selective attention and monitoring of the self and the environment (e.g., listening for noise, monitoring the body for signs of sleepiness). Consequently, a feedback loop is created where increased arousal leads to increased monitoring which i n turn leads to increased cognitive activity and arousal which, in turn, further disrupts sleep. Cognitive arousal may be particularly relevant to sleep disruption due to the ability of cognitive arousal to perpetuate further wakefulness. This perpetuati on, it can be argued, is less likely to be found with somatic arousal. The day to day relationship between cognitive arousal and poor sleep was replicated in the between person analyses suggesting that, overall, higher levels of cognitive arousal are asso ciated with worse sleep. Aim 2 showed that affect significantly predicted sleep on a day to day and between persons level for both younger and older adults. Days with higher ratings of pleasant feelings, positive affect, and less negative affect were as sociated with better sleep in younger adults. This result supports recent findings from a study with Canadian University students which found that days with higher positive affect were associated with better sleep quality that night (Galambos et

PAGE 130

130 al., 2009) The present study extends these findings to show that in addition to positive affect, negative affect and pleasant feeling ratings also predicted sleep the following night. Additionally, overall ratings of pleasant feelings and less negative affect w ere associated with overall better sleep. Similar results were found for older adults with days with higher ratings of pleasant feelings associated with better sleep. Overall, more pleasant feelings, positive affect, and less negative affect was associat ed with overall better sleep for older adults. These findings support previous research indicating that positive and negative affect is associated with sleep (Berry & Webb, 1985a; McCrae et al., 2008; Totterdell et al., 1994) The lack of an association between negative affect and sleep for older adults mirrors findings from a recent study which found that negative affect changed more in relation to sleep for younger adults than older adults (Ready et al., 2009) An explanation for the significant relat ionship between affect and sleep is that more positive affect could indirectly impact sleep by increasingly the likelihood that an individuals would engage in sleep promoting activities (e.g., increased activity levels, exercise, and exposure to sunlight outdoors). Additionally, negative affect could serve to activate the first states of the cognitive model of insomnia (Harvey, 2000) by creating negative thoughts and ruminations which beget a cycle of sleep disruption. In summary, aim 2 examined relationshi ps between the predictor variables and sleep at day to day and mean levels. Social rhythmicity predicted sleep at the between persons level for younger adults. Light, arousal, and affect predicted sleep at the between and within person levels for both younger and older adults. Interestingly, relationships were seen at the day to day level that were not present when examined at the between persons and relationships were observed at the between persons level and not at the within person level. The discrepancy in findings

PAGE 131

131 supports the importance of including both intraindividual and interindividual differences in the study of dynamic behaviors. Aim 4 : Identify to What Extent Light, Arousal, and Affect Mediate the Relationship Between Social Rhythmicity a nd Sleep The goal of aim four was to investigate the possible mediating effects of light, arousal, and affect on the relationship between social rhythmicity and sleep. The mediating effects of light, arousal, and affect could not be estimated as the step s of Baron and Kennys model of mediation could not be achieved. The predictor variable of social rhythmicity was correlated with sleep for younger adults (step 1) and correlated with the potential mediator variables of light, arousal, and affect (step 2) The mediator variables that were associated with social rhythmicity, however, were not associated with the sleep variables (step 3). For older adults, the predictor variable of social rhythmicity was not correlated with sleep (step 1). Figures 6 1 and 62 summarize the extent that the mediated model of social rhythmicity was explored in this study. Strengths and Weaknesses of the Present Study The major strengths of the present study were the examination of both intraindividual and between persons va riability, inclusion of younger and older adults, and the use of a microlongitudinal design. Few studies have examined within person variability and the results suggest that there is significant variability in these constructs that is overlooked by static measurements. By using both mean and intraindividual measures, significant overall effects could be examined on an individual level. The results suggest that both between and within person analyses revealed unique relationships between the predictor var iables and sleep. The use of a younger and older adult sample allowed for age comparisons in mean level, intraindividual variability estimates, and day to day associations. The use of a longitudinal data collection period (over 14 days) enabled both the estimation of variability as well as the examination of

PAGE 132

132 day to day relationships. The use of a daily diary measurement minimized the effects of recall bias. Additionally, the examination of social rhythmicity, light, arousal, affect, and sleep within the younger and older adults natural environments preserved the ecological validity of the relationships which might be missing from a laboratory sleep study. Weaknesses of the present study include the sampling methods, the use of internet data collection and the inability to deduce causation. The study employed a convenience sampling method. The implication of this recruitment method is that the sample may not be representative of the general population and consequently, the results have limited genera lizability. Although the younger adult sample was relatively racially diverse, the older adult sample consisted primarily of white participants Additionally, by not randomly sampling the population, participants may have self selected to participate in the study based on motivations that could affect the results. For example, individuals who have particularly poor or good sleep may be interested in participating in a sleep study. The use of internet data collection can be problematic given the lack of experimental control over the study conditions and the restriction of the sample to those who own a computer or have internet access. Finally, although we could discern directionality between the variables of interest based on the timing of the measures, an experimental manipulation is necessary in order to uncover causal relationships between the predictor variables and sleep. Implications The findings of the study have both theoretical and clinical implications. In terms of theoretical implications, the results underscore the importance of examining psychological behaviors at the intraindividual level and day to day level. First, a significant portion of variability can be overlooked when only mean level variables are considered. Second, relationships emerged on a day to day basis that were not present in static, mean level

PAGE 133

133 comparisons. This study examined three different modes by which the predictor variables could be related to sleep: 1) variability, 2) higher/lower levels on a day to day basis, and 3) higher/lower levels overall. The clinical implications of investigating these relationships using multiple modes include gaining a better understanding of the ways in which behaviors are related to sleep. For example, for younger adults variations in the timing of first light exposure, the timing of light exposure on a day to day basis and the timing of light exposure overall over a two week period were differentially associated with sleep. Consequently, for some constructs, a clinician may focus on eliminating variability, increasing or lowering average levels on a biweekly basis, or increasing or lowering daily levels of a behavior. Additionally, by investigating multiple facets of each of the constructs of interest, we could discern the specific aspects of light, arousal, and affect which are related to sleep. For example, cognitive arousal was particularly salient in association with sleep outcomes in both older and younger adults. Consequently, therapeutic techniques which emphasize cognitive versus somatic tension may be particularly helpful in targeting sleep complaints. Future Directions This study employed primarily subjective measures of the variables of interest. Future studies could investigate these variables using objective measures ( such as actigraphy) or physiological markers for psychological constructs (e.g., cortisol levels). In order to observe the changes in circadian rhythm resulting from social zeitgebers, physiological underpinnings of circadian rhythms (e.g., core body temperature, salivary, or plasma melatonin levels) could be observed. Additionally, the burgeoning use of smartphones may enable the use of moment to moment measurement to capture psychological states and behaviors. The use of a recently developed instrume nt to assess routine in older adults (the Scale of Older Adults Routine [SOAR]; (Zisberg, Young, & Schepp, 2009) ), in addition to the Social Rhythm Metric, could

PAGE 134

134 enable the assessment of multiple facets of routinized behavior such as the frequency, durati on, and timing of events. In addition to examining the relationship between variability in the predictor variables and sleep in community dwelling participants, it would be helpful to expand the context of study to other environments such as residential or assisted living facilities. In terms of data analyses, while the MLM permitted the study of day to day relationships, it would be interesting to examine the relationships that do not necessarily occur on the same day (i.e., time lagged relationships). Also, the recent use of measurement burst studies (e.g., Ram & Gerstoff, 2009) could have implications for examining the relationship between social rhythmicity and sleep over time. These studies employ a microlevel focus (e.g., daily measures over two weeks) embedded within a longitudinal timeframe (e.g., ten years). Therefore, age related changes in within person variability can be seen over time. Finally, the present study demonstrated that older adults exhibit more stability in terms of social rhy thmicity, light, arousal, and affect compared to younger adults. These results are consistent with a gain oriented view of aging. Future studies can continue to investigate aspects of aging that reflect stabilizing, versus destabilizing processes. This emphasis on strengths associated with aging is consistent with the strength based focus in the field of Counseling Psychology. Conclusion In summary, the results indicated that younger adults were highly variable within person for all variables. There w ere significant age related differences in variability with younger adults significantly more variable on all measures except for the PSAS somatic scale. Despite showing less variability compared to younger adults, older adults still demonstrated consider able variability. Age related differences in mean levels were observed for social rhythmicity, light exposure, arousal, affect, and sleep. Variability in social rhythmicity, light, arousal, and affect was associated with worse sleep for younger adults an d better and worse sleep for older adults.

PAGE 135

135 Multilevel modeling analyses showed that social rhythmicity predicted sleep at the between person level and light, arousal, and affect predicted sleep at both within and between person levels for younger adults. For older adults, light, arousal, and affect predicted sleep at both between and within person levels. The potential mediating effects of light, arousal, and affect could not be tested as there were insufficient associations between the predictor, potential mediators, and outcome variables for both age groups. Results provide support for the relationship between social rhythmicity, light, arousal, affect, and sleep for younger adults and for light, arousal, affect and sleep for older adults. Further re search is needed to test the mediating effects of light, arousal, and affect.

PAGE 136

136 Figure 6 1. Relationships in the mediated model of social rhythmicity that were observed in younger adults The dashed lines indicat e assoc iations that were not observed. The solid lines indicate associations that were observed. The solid lines indicate the presence of mean level relationships. The fluctuating black solid lines indicate the presence of day to day relationships. Social Rhythmicity Arousal Sleep Affect Light

PAGE 137

137 Figure 6 2. Relationships in the mediated model of social rhythmicity that were observed in older adults The dashed lines indicat e associations that were not observed. The solid lines indicate associations tha t were observed. The solid lines indicate the presence of mean level relationships. The fluctuating black solid lines indicate the presence of day to day relationships. Social Rhythmicity Arousal Sleep Affect Light

PAGE 138

138 APPENDIX A BASELINE SURVEY

PAGE 139

139

PAGE 140

140

PAGE 141

141

PAGE 142

142

PAGE 143

143

PAGE 144

144

PAGE 145

145

PAGE 146

146

PAGE 147

147

PAGE 148

148

PAGE 149

149

PAGE 150

150

PAGE 151

151

PAGE 152

152

PAGE 153

153

PAGE 154

154

PAGE 155

155

PAGE 156

156

PAGE 157

157

PAGE 158

158

PAGE 159

159

PAGE 160

160 APPENDIX B DAILY SURVEY

PAGE 161

161

PAGE 162

162

PAGE 163

163

PAGE 164

164

PAGE 165

165

PAGE 166

166

PAGE 167

167

PAGE 168

168

PAGE 169

169

PAGE 170

170

PAGE 171

171

PAGE 172

172

PAGE 173

173

PAGE 174

174

PAGE 175

175

PAGE 176

176 LIST OF REFERENCES Almeida, D. M., Piazza, J. R., & Stawski, R. (2009). Interindividual differences and intraindividual variability in the cortisol awakening response: An examination of age and gender. Psyc hology and Aging, 24(4), 819 827. American Psychiatric Association (2000). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, D.C.: American Psychiatric Association. Amir, S., Beaule, C., Arvanitogiannis, A., & Stewart, J. (2002). Modes of plasticity within the mammalian circadian system Plasticity in the Adult Brain: from Genes to Neurotherapy (Vol. 138, pp. 191203). Amir, S., & Stewart, J. (1996). Resetting of the circadian clock by a conditioned stimulus. Nature, 379(6565) 542545. Ancoli Israel, S., Martin, J. L., Kripke, D. F., Marler, M., & Klauber, M. R. (2002). Effect of light treatment on sleep and circadian rhythms in demented nursing home patients. [Proceedings Paper]. Journal of the American Geriatrics Society, 50(2), 282289. Aschoff, J. (1965). Circadian rhythms in man a self sustained oscillator with an inherent frequency underlies human 24hour periodicity. Science, 148 (3676), 1427. Aschoff, J. (1986). Anticipation of a daily meal A process of learning d ue to entrainment. Monitore Zoologico ItalianoItalian Journal of Zoology, 20(2), 195219. Aschoff, J., Fatransk.M, Giedke, H., Doerr, P., Stamm, D., & Wisser, H. (1971). Human circadian rhythms in continuous darkness Entrainment by social cues. Science, 171(3967), 213. Ashman, S. B., Monk, T. H., Kupfer, D. J., Clark, C. H., Myers, F. S., Frank, E., et al. (1999). Relationship between social rhythms and mood in patients with rapid cycling bipolar disorder. Psychiatry Research, 86(1), 1 8. Barger, L. K ., Wright, K. P., Hughes, R. J., & Czeisler, C. A. (2004). Daily exercise facilitates phase delays of circadian melatonin rhythm in very dim light. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 286(6), R1077R1084. Baro n, 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(6), 11731182. Beck, A. T., Steer, R.A., & Garbin, M.G. (1996). Beck Depression InventorySecond Edition. San Antonio, TX: The Psychological Corporation.

PAGE 177

177 Bendetti, F., Barbini, B., Campori, E., Fulgosi, M. C., Pontiggia, A., & Colombo, C. (2001). Sleep phase advance and lithium to sustain the a ntidepressant effect if total sleep deprivation in bipolar depression: New findings supporting the internal coincidence model? Journal of Psychiatric Research, 35, 323329. Berry, D. T. R., & Webb, W. B. (1985a). Mood and sleep in aging women. Journal of Personality and Social Psychology, 49(6), 17241727. Berry, D. T. R., & Webb, W. B. (1985b). Sleep and cognitive functions in normal older adults. Journals of Gerontology, 40(3), 331335. Bhagwagar, Z., Hafizi, S., & Cowen, P. J. (2003). Increase in conc entration of waking salivary cortisol in recovered patients with depression. American Journal of Psychiatry, 160(10), 18901891. Bogdan, A., Bouchareb, B., & Touitou, Y. (2001). Ramadan fasting alters endocrine and neuroendocrine circadian patterns. Meal time as a synchronizer in humans? Life Sciences, 68(14), 16071615. Boker, S. M., Molenaar, P. C. M., & Nesselroade, J. R. (2009). Issues in intraindividual variability: Individual differences in equilibria and dynamics over multiple time scales. Psycholo gy and Aging, 24(4), 858862. Bonnet, M. H., & Arand, D. L. (1995). 24hour metabolic rate in insomniacs and matched normal sleepers. Sleep, 18(7), 581588. Bonnet, M. H., & Arand, D. L. (1999). Level of arousal and the ability to maintain wakefulness. J ournal of Sleep Research, 8(4), 247254. Bootzin, R. R. (1972). A stimulus control treatment for insomnia. Proceedings of the 80th annual convention of the american psychological association, 395396. Bouisson, J., & Swendsen, J. (2003). Routinization and emotional wellbeing: An experience sampling investigation in an elderly French sample. Journals of Gerontology Series B Psychological Sciences and Social Sciences, 58 (5), P280P282. Britton, J. (1997). Measurement of peak flow variability in community populations: Methodology. European Journal of Respiration, 10, 4244. Brown, L. F., Reynolds, C. F., Monk, T. H., Prigerson, H. G., Dew, M. A., Houck, P. R., et al. (1996). Social rhythm stability following late life spousal bereavement: associations with depression and sleep impairment. Psychiatry Research, 62(2), 161169.

PAGE 178

178 Bryk, D. T. R., & Raudenbush, S. W. (1992). Hierarchial Linear Models for Social and Behavioral Research: Applications and Data Analysis Methods Newbury Park, CA: Sage. Campbell, S. S ., & Daston, P. (1991). Bright light treatment of sleep disturbance in older subjects. Sleep Research, 20 448. Campbell, S. S., Dawson, D., & Anderson, M. W. (1993). Alleviation of sleep maintenance insomnia with timed exposure to bright light. Journal of the American Geriatrics Society, 41(8), 829836. Campbell, S. S., Kripke, D. F., Gillin, J. C., & Hrubovcak, J. C. (1988). Exposure to Light in Healthy Elderly Subjects and Alzheimers Patients. Physiology & Behavior, 42(2), 141144. Card, J. P. (Ed.). (1999). Anatomy of the mammalian circadian time keeping system. Boca Raton: CRC Press. Carney, C., Edinger, J., Meyer, B., Lindman, L., & Istre, T. (2006). Daily activities and sleep quality in college students. Chronobiology International: The Journal of Biological & Medical Rhythm Research, 23(3), 623637. Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory of socioemotional selectivity. American Psychologist, 54(3), 165181. Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000). Emotional experience in everyday life across the adult life span. Journal of Personality and Social Psychology, 79(4), 644655. Charles, S. T., Reynolds, C. A., & Gatz, M. (2001). Age related differences and change i n positive and negative affect over 23 years. Journal of Personality and Social Psychology, 80(1), 136151. Chuo, T. C., Scammell, T. E., Gooley, J. J., Gaus, S. E., Saper, C. B., & LU, J. (2003). Critical role of dorsomedial hypothalamic nucleus in a wid e range of behavioral circadian rhythms. The Journal of Neuroscience, 23, 10691 10702. Cortoos, A., Verstraeten, E., & Cluydts, R. (2006). Neurophysiological aspects of primary insomnia: Implications for its treatment. Sleep Medicine Reviews, 10 (4), 255266. Czeisler, C. A., Kronauer, R. E., Allan, J. S., Duffy, J. F., Jewett, M. E., Brown, E. N., et al. (1989). Bright light induction of strong (Type 0) resetting of the human circadian pacemaker. Science, 244 (4910), 13281333.

PAGE 179

179 Dautovich, N. D., McCrae, C. S., Rowe, M., & Dzierzewski, J. M. (2008). How variable is napping behavior in older adults? Withinperson variability in napping in older adults in relation to sleep. Paper presented at the Annual Meeting of the Associated Professional Sleep Societies. Duffy, J. F., Rimmer, D. W., & Czeisler, C. A. (2001). Association of intrinsic circadian period with morningness eveningness, usual wake time, and circadian phase. Behavioral Neuroscience, 115 (4), 895899. Duffy, J. F., & Wright, K. P. (2005). Entrainme nt of the human circadian system by light. Journal of Biological Rhythms, 20(4), 326338. Edinger, J. D., Marsh, G. R., McCall, W. V., Erwin, C. W., & Lininger, A. W. (1991). Sleep variability across consecutive nights of home monitoring in older mixed DI MS patients. Sleep, 14(1), 13 17. Ehlers, C. L., Frank, E., & Kupfer, D. J. (1988). Social zeitgebers and biological rhythms A unified approach to understanding the etiology of depression. Archives of General Psychiatry, 45(10), 948 952. Ekman, P., & F riesen, W. (1976). Pictures of facial affect. Palo Alto, CA: Consulting Psychologists Press. Espie, C. A., Brooks, D. N., & Lindsay, W. R. (1989). An evaluation of tailored psychological treatment of insomnia. Journal of Behavior Therapy and Experimental Psychiatry, 20(2), 143153. Faul, F., Buchner, A., Erdfelder, E., & Albert Georg, L. (2006). G*Power (Version 3.0.3). Germany. Frank, E., Hlastala, S., Ritenour, A., Houck, P., Xin Ming, T., Monk, T. H., et al. (1997). Inducing lifestyle regularity in re covering bipolar disorder patients: Results from the maintenance therapies in bipolar disorder protocol. Biological Psychiatry, 41(12), 11651173. Frijda, N. H. (1988). The laws of emotion. American Psychologist, 43, 9. Fuller, P. M., Gooley, J. J., & Sa per, C. B. (2006. ). Neurobiology of the sleepwake cycle: Sleep architecture, circadian regulation, and regulatory feedback. Journal of Biological Rhythms, 21(6), 482493.

PAGE 180

180 Galambos, N. L., Dalton, A. L., & Maggs, J. L. (2009). Losing Sleep Over It: Daily Variation in Sleep Quantity and Quality in Canadian Students' First Semester of University. [Article]. Journal of Research on Adolescence, 19(4), 741 761. Grandin, L. D., Alloy, L. B., & Abramson, L. Y. (2006). The social zeitgeber theory, circadian rhyt hms, and mood disorders: Review and evaluation. Clinical Psychology Review, 26(6), 679694. Gross, J. J., Carstensen, L. L., Pasupathi, M., Tsai, J., Skorpen, C. G., & Hsu, A. Y. C. (1997). Emotion and aging: Experience, expression, and control. Psychology and Aging, 12(4), 590599. Gross, R. T., & Borkovec, T. D. (1982). Effects of a cognitive intrusion manipulation on the sleep onset latency of good sleepers. Behavior Therapy, 13(1), 112 116. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate Data Analysis Englewood Cliffs, NJ: Prentice Hall. Hall, M., Buysse, D. J., Reynolds, C. F., Kupher, D. J., & Baum, A. (1994). Stress related intrusive thoughts disrupt sleep onset and continuity. Sleep Research, 25, 163. Halle, S., & Stenseth, N. C. (Eds.). (2000). Activity Patterns in Small Mammals: An Ecological Approach (Ecological Studies, 141) Verlag: Springer. Harvey, A. G. (2000). Pre sleep cognitive activity: A comparison of sleep onset insomniacs and good sleepers. Britis h Journal of Clinical Psychology, 39, 275286. Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40(8), 869893. Haynes, S. N., Adams, A., & Franzen, M. (1981). The effects of pre sleep stress on sleep onset insomnia. J ournal of Abnormal Psychology, 90(6), 601 606. Himmelfarb, S., & Murrell, S. A. (1983). Reliability and validity of 5 mental health scales in older persons. Journals of Gerontology, 38(3), 333339. Hood, B., Bruck, D., & Kennedy, G. (2004). Determinants of sleep quality in the healthy aged: the role of physical, psychological, circadian and naturalistic light variables. Age and Ageing, 33(2), 159165. Horne, J. A. (1988). Why We Sleep New York: Oxford University Press.

PAGE 181

181 Howland, R. H., & Thase, M. E. (E ds.). (1999). Affective disorders: Biological aspects. London: Oxford University Press. Hughes, P. C., & Neer, R. M. (1981). Lighting for the elderly A psychobiological approach to lighting. Human Factors, 23(1), 6585. Hultsch, D. F., Strauss, E., & H unter, M. A. (2008). Intraindividual variability, cognition, and aging. In F. I. M. Craik & T. A. Salthouse (Eds.), The handbook of aging and cognition (pp. 491556). New York: Psychology Press. Jean Louis, G., Kripke, D. F., Ancoli Israel, S., Klauber, M R., Sepulveda, R. S., Mowen, M. A., et al. (2000). Circadian sleep, illumination, and activity patterns in women: influences of aging and time reference. Physiology & Behavior, 68(3), 347352. Jilge, B. (1995). Ontogeny of the rabbit's circadian rhythms without an external zeitgeber. Physiology & Behavior, 58(1), 131140. Kawinska, A., Dumont, M., Selmaoui, B., Paquet, J., & Carrier, J. (2005). Are modifications of melatonin circadian rhythm in the middle years of life related to habitual patterns of light exposure? Journal of Biological Rhythms, 20(5), 451460. Klein, D. C., Moore, R. Y., & Reppert, S. M. (1991). The Suprachiasmatic Nucleus: The Mind's Clock. New York: Oxford University Press. Klein, K. E., & Wegmann, H. M. (1975). The resynchronizati on of human circadian rhythms after transmeridian flights as a result of flight direction and mode of activity. In L. E. Scheving, F. Halberg & J. E. Pauly (Eds.), Chronobiology Tokyo: Igaku Shoin. Kohsaka, M., Fukuda, N., Honma, H., Kobayashi, R., Sakakibara, S., Koyama, E., et al. (1999). Effects of moderately bright light on subjective evaluations in healthy elderly women. Psychiatry and Clinical Neurosciences, 53 (2), 239241. Kohsaka, M., Fukuda, N., Kobayashi, R., Honma, H., Sakakibara, S., Koyama, E., et al. (2000). Effect of short duration morning bright light in elderly men: Sleep structure. Psychiatry and Clinical Neurosciences, 54(3), 367368. Kronfeld Schor, N., Dayan, T., Elvert, R., Haim, A., Zisapel, N., & Heldmaier, G. (2001). On the use o f the time axis for ecological separation: Diel rhythms as an evolutionary constraint. The American Naturalist, 158(4), 451457. Kryger, M., Monjan, A., Bliwise, D., & Ancoli Israel, S. (2004). Sleep, health, and aging Bridging the gap between science and clinical practice. Geriatrics, 59(1), 24 30.

PAGE 182

182 Kunzmann, U., Little, T. D., & Smith, J. (2000). Is age related stability of subjective wellbeing a paradox? Cross sectional and longitudinal evidence from the Berlin Aging Study. Psychology and Aging, 15(3), 511526. Kupfer, D. J., & Foster, F. G. (1972). Interval between onset of sleep and rapideye movement sleep as an indicator of depression. Lancet, 2(7779), 684&. Kvaal, K., Ulstein, I., Nordhus, I. H., & Engedal, K. (2005). The Spielberger StateTrai t Anxiety Inventory (STAI): The state scale in detecting mental disorders in geriatric patients. International Journal of Geriatric Psychiatry, 20(7), 629634. Lack, L., & Wright, H. (1993). The effect of evening bright light in delaying the circadianrhy thms and lengthening the sleep of early morning awakening insomniacs. Sleep, 16(5), 436443. Lichstein, K. L., Durrence, H. H., Taylor, D. J., Bush, A. J., & Riedel, B. W. (2003). Quantitative criteria for insomnia. Behaviour Research and Therapy, 41(4), 427445. Lichstein, K. L., & Morin, C. M. (2000). Treatment of Late life Insomnia Thousand Oaks, CA: Sage Publications Inc. Lichstein, K. L., Riedel, B. W., & Means, M. K. (1999). Psychological treatment of late life insomnia. In R. Schulz, G. Maddox & P. Lawton (Eds.), Anuual Review of Gerontology and Geriatrics: Vol. 18. Focus on Interventions Research with Older Adults. (pp. 74110). New York: Springer. Lichstein, K. L., & Rosenthal, T. L. (1980). Insomniacs perceptions of cognitive versus somatic de terminants of sleep disturbance. Journal of Abnormal Psychology, 89(1), 105 107. Lipsitz, L. A. (2002). Dynamics of stability: The physiologic basis of functional health and frailty. Journals of Gerontology Series aBiological Sciences and Medical Science s, 57(3), B115 B125. Lorr, M., Daston, P., & Smith, I. R. (1967). An analysis of mood states. Educational and Psychological Measurement, 27(1), 89. Malkoff Schwartz, S., Frank, E., Anderson, B., Sherrill, J. T., Siegel, L., Patterson, D., et al. (1998). Stressful life events and social rhythm disruption in the onset of manic and depressive bipolar episodes A preliminary investigation. Archives of General Psychiatry, 55(8), 702 707.

PAGE 183

183 Malkoff Schwartz, S., Frank, E., Anderson, B. P., Hlastala, S. A., Luther, J. F., Sherrill, J. T., et al. (2000). Social rhythm disruption and stressful life events in the onset of bipolar and unipolar episodes. Psychological Medicine, 30(5), 10051016. Mandler, G. (1975). Mind and emotion. New York: Wiley. McCarley, R. W., & Massaquoi, S. G. (1992). Neurobiological structure of the revised limit cycle reciprocal interaction model of REM cycle control. The Journal of Sleep Research, 1, 132137. McClintock, M. K. (1971). Menstrual synchrony and suppression. Nature, 229(5282) 244. McCrae, C. S., McNamara, J. P. H., Rowe, M. A., Dzierzewski, J. M., Dirk, J., Marsiske, M., et al. (2008). Sleep and affect in older adults: Using multilevel modeling to examine daily associations. Journal of Sleep Research, 17(1), 42 53. McCrae, C. S., Rowe, M.A., Dautovich, N.D., Lichstein, K.L., Durrence, H.H., Riedel, & B.W., T., D.J., & Bush, A.J. (2006). Sleep hygiene practices in two community dwelling samples of older adults. Sleep, 29(12), 9. Mehrabian, A., & Russell, J. A. (1974). An a pproach to environmental psychology Cambridge, MA: MIT Press. Mini Mitter Co. Inc. (2001). Actiwatch 16/Actiwatch 64/Actiwatch L/ Actiwatch Score Instruction Manual : Mini Mitter Co., Inc. Mish, F. C. E. (2000). In F. C. Mish (Ed.), Merriam Webster's Col legiate Dictionary (10 ed.). Springfield, Massachusetts: Merriam Webster, Incorporated. Mishima, K., Okawa, M., Shimizu, T., & Hishikawa, Y. (2001). Diminished melatonin secretion in the elderly caused by insufficient environmental illumination. Journal of Clinical Endocrinology & Metabolism, 86(1), 129134. Mistlberger, R. E., Antle, M. C., Glass, J. D., & Miller, J. D. (2000). Behavioral and serotongergic regulation of circadian rhythms. Biological Rhythm Research, 31, 43. Mistlberger, R. E., & Skene, D. J. (2004). Social influences on mammalian circadian rhythms: animal and human studies. Biological Reviews, 79(3), 533556. Monk, T. H. (2005). Aging Human Circadian Rhythms: Conventional Wisdom May Not Always Be Right. Journal of Biological Rhythms, 20(4), 366374. Monk, T. H. (2010). Enhancing Circadian Zeitgebers. [Editorial Material]. Sleep, 33(4), 421 422.

PAGE 184

184 Monk, T. H., Buysse, D. J., Hall, M., Nofzinger, E. A., Thompson, W. K., Mazumdar, S. A., et al. (2006). Age related differences in the lifestyle regularity of seniors experiencing bereavement, care giving, insomnia, and advancement into oldold age. Chronobiology International, 23(4), 831841. Monk, T. H., Flaherty, J. F., Frank, E., Hoskinson, K., & Kupfer, D. J. (1990). The Social Rhy thm Metric An instrument to quantify the daily rhythms of life. Journal of Nervous and Mental Disease, 178 (2), 120126. Monk, T. H., Houck, P. R., & Shear, M. K. (2006). The daily life of complicated grief patients -What gets missed, what gets added? De ath Studies, 30(1), 7785. Monk, T. H., Kupfer, D. J., Frank, E., & Ritenour, A. (1991). The Social Rhythm Metric (SRM): Measuring daily social rhythms over 12 weeks. Psychiatry Research, 36 (2), 195207. Monk, T. H., Petrie, S. R., Hayes, A. J., & Kupfer D. J. (1994). Regularity of daily life in relation to personality, age, gender, sleep quality and circadian rhythms. Journal of Sleep Research, 3 (4), 196205. Monk, T. H., Reynolds, C. F., III, Buysse, D. J., DeGrazia, J. M., & Kupfer, D. J. (2003). The relationship between lifestyle regularity and subjective sleep quality. Chronobiology International, 20(1), 97107. Monk, T. H., Reynolds, C. F., Kupfer, D. J., Hoch, C. C., Carrier, J., & Houck, P. R. (1997). Differences over the life span in daily life style regularity. Chronobiology International, 14(3), 295306. Monk, T. H., Reynolds, C. F., Machen, M. A., & Kupfer, D. J. (1992). Daily social rhythms in the elderly and their relation to objectively recorded sleep. Sleep, 15 (4), 322329. Mroczek, D. K., & Kolarz, C. M. (1998). The effect of age on positive and negative affect: A developmental perspective on happiness. Journal of Personality and Social Psychology, 75(5), 13331349. Mrosovsky, N. (1996). Locomotor activity and nonphotic influences on c ircadian clocks. Biological Reviews of the Cambridge Philosophical Society, 71(3), 23. National Sleep Foundation (2008). Sleep in America Poll Summary of Findings Washington, D.C. Nemiah, J. C., & Sifneos, P. E. (1970). Psychosomatic Illness Problem in Communication. Psychotherapy and Psychosomatics, 18(1 6), 154 160.

PAGE 185

185 Nesselroade, J. R. (1991). Interindividual differences in intraindividual change. In J. L. Horn & L. M. Collins (Eds.), Best Methods for the Analysis of Change: Recent Advances, Unans wered Questions, Future Directions. (pp. 92 107). Washington DC: American Psychological Association. Nesselroade, J. R. (2002). Elaborating the differential in differential psychology. Multivariate Behavioral Research, 37 (4), 543561. Nicassio, P. M., Mendlowitz, D. R., Fussell, J. J., & Petras, L. (1985). The phenomenology of the PreSleep State The development of hte Pre Sleep Arousal Scale. Behaviour Research and Therapy, 23(3), 263271. Novy, D. M., Nelson, D. V., Goddwin, J., & Rowzee, R. D. (1993 ). Psychometric comparability of the State Trait Anxiety Inventory for different ethnic subpopulations. Psychological Assessment, 5 343349. Ohayon, M. M., Carskadon, M. A., Guilleminault, C., & Vitiello, M. V. (2004). Meta analysis of quantitative s leep parameters from childhood to old age in healthy individuals: Developing normative sleep values across the human lifespan. [Article]. Sleep, 27(7), 12551273. Ohta, H., Honma, S., Abe, H., & Honma, K. (2002). Effects of nursing mothers on rPer1 and rP er2 circadian expressions in the neonatal rat suprachiasmatic nuclei vary with developmental stage. European Journal of Neuroscience, 15 (12), 19531960. Pagani, M. (1999). Cardiovascular oscillations: From a physiological curiosity to a probe of aging and health? Journal of Gerontology: Biological Sciences, 54A B225 B226. Portella, M. J., Harmer, C. J., Flint, J., Cowen, P., & Goodwin, G. M. (2005). Enhanced early morning salivary cortisol in neuroticism. American Journal of Psychiatry, 162(4), 807809. Prigerson, H. G., Reynolds, C. F., Frank, E., Kupfer, D. J., George, C. J., & Houck, P. R. (1994). Stressful life events, social rhytms, and depressive symptoms among the elderly: An examination of hypothesized causal linkages. Psychiatry Research, 51(1), 33 49. Prinz, P. N., Bailey, S. L., & Woods, D. L. (2000). Sleep impairments in healthy seniors: Roles of stress, cortisol, and interleukin1 beta. Chronobiology International, 17(3), 391404. Pruessner, J. C., Hellhammer, D. H., & Kirschbaum, C. (1999) Burnout, perceived stress, and cortisol responses to awakening. Psychosomatic Medicine, 61(2), 197204.

PAGE 186

186 Ram, N., Lindenberger, U., & BlanchardFields, F. (2009). Introduction to the special section on intraindividual variability and aging. Psychology and Aging, 24(4), 775777. Ready, R. E., Marquez, D. X., & Akerstedt, A. (2009). Emotion in Younger and Older Adults: Retrospective and Prospective Associations with Sleep and Physical Activity. [Article]. Experimental Aging Research, 35(3), 348368. Reppe rt, S. M. (1995). Interaction between the circadian clocks of mother and fetus Circadian Clocks and Their Adjustment (Vol. 183, pp. 198207). Rocke, C., Li, S., & Smith, J. (2009). Intraindividual variability in positive and negative affect over 45 days: Do older adults fluctuate less than young adults? Psychology and Aging, 24(4), 863878. Rodenbeck, A., Huether, G., Ruther, E., & Hajak, G. (2002). Interactions between evening and nocturnal cortisol secretion and sleep parameters in patients with severe chronic primary insomnia. Neuroscience Letters, 324 (2), 159163. Russell, J. A., & Bullock, M. (1985). Multidimensional Scaling of Emotional Facial Expressions Similarity from Preschoolers to Adults. Journal of Personality and Social Psychology, 48(5), 12901298. Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect Grid: A single item scale of pleasure and arousal. Journal of Personality and Social Psychology, 57(3), 493502. Scheuermaier, K., Laffan, A. M., & Duffy, J. F. (2010). Light Exposure Patterns in Healthy Older and Young Adults. Journal of Biological Rhythms, 25(2), 113122. Shaffery, J. P., Sinton, C. M., Bissette, G., Roffwarg, H. P., & Marks, G. A. (2002). Rapid eye movement sleep deprivation modifies expression of long term poten tiation in visual cortex of immature rats. Neuroscience, 110 (3), 431443. Shear, M. K., Randall, J., Monk, T. H., Ritenour, A., Tu, X., Frank, E., et al. (1994). Social rhythm in anxiety disorder patients. Anxiety, 1 (2), 90 95. Shimoda, K., Hanada, K., Y amada, N., Takahashi, K., & Takahashi, S. (1986). Periodic exposure to mother is potent zeitgeber of rat pups' rhythm. Physiology and Behavior, 36, 7. Singer, J. D., Davidson, S. M., Graham, S., & Davidson, H. S. (1998). Physician retention in community a nd migrant health centers Who stays and for how long? Medical Care, 36(8), 11981213.

PAGE 187

187 Singer, J. D., Fuller, B., Keiley, M. K., & Wolf, A. (1998). Early childcare selection: Variation by geographic location, maternal characteristics, and family structur e. Developmental Psychology, 34(5), 11291144. Singer, J. D., & Willett, J. B. (2003). Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. New York: Oxford University Press. Spielberger, C. D., Gorsuch, R.L., Lushene, R., Vagg, P.R., & Jacobs, G.A. (1983). Stait Trait Anxiety Inventory, Form Y Palo Alto, CA: Consulting Psychologists Press. Spielman, A. J., Saskin, P., & Thorpy, M. J. (1987). Treatment of chronic insomnia by restriction of time in bed. Sleep, 10(1), 4556. Stetler, C., Dickerson, S. S., & Miller, G. E. (2004). Uncoupling of social zeitgebers and diurnal cortisol secretion in clinical depression. Psychoneuroendocrinology, 29(10), 12501259. Storch, E. A., Roberti, J. W., & Roth, D. A. (2004). Factor Structure, Concurrent Validity, and Internal Consistency of the Beck Depression Inventory Second Edition in a Sample of College Students. Depression and Anxiety, 19(3), 187189. Tabachnick, B. G., & Fidell, L. S. (2007). Using Multivariate Statistics (5th ed.). New York: Pearson. Taillard, J., Philip, P., & Bioulac, B. (1999). Morningness/Eveningness and the need for sleep. Journal of Sleep Research, 8, 291295. Teresi, J., Lawton, M. P., Ory, M., & Holmes, D. (1994). Measurement issues in chronic care populations Dem entia special care. Alzheimer Disease & Associated Disorders, 8 S144S183. Terry, P. C., Lane, A. M., Lane, H. J., & Keohane, L. (1999). Development and validation of a mood measure for adolescents. Journal of Sports Sciences, 17 (11), 861872. Thayer, R E. (1986). ActivationDeactivation Adjective Check List Current Overview and Structural Analysis. Psychological Reports, 58(2), 607614. Totterdell, P., Reynolds, S., Parkinson, B., & Briner, R. B. (1994). Associations of sleep with everyday mood, minor symptoms and social interaction experience. Sleep, 17(5), 466475. Vanegeren, L., Haynes, S. N., Franzen, M., & Hamilton, J. (1983). Presleep cognitions and attributions in sleep onset insomnia. Journal of Behavioral Medicine, 6(2), 217232.

PAGE 188

188 Vaz Fragos o, C. A., & Gill, T. M. (2007). Sleep complaints in community living older persons: A multifactorial geriatric syndrome. Journal of the American Geriatrics Society, 55(11), 18531866. Vgontzas, A. N., Bixler, E. O., Lin, H. M., Prolo, P., Mastorakos, G., Vela Bueno, A., et al. (2001). Chronic insomnia is associated with nyctohemeral activation of the hypothalamic pituitary adrenal axis: Clinical implications. Journal of Clinical Endocrinology and Metabolism, 86(8), 37873794. Vitaterna, M. H., Takahashi, J. S., & Turek, F. W. (2001). Overview of circadian rhythms. Alcohol Research & Health, 25(2), 85. Vitiello, M. V., Moe, K. E., & Prinz, P. N. (2002). Sleep complaints cosegregate with illness in older adults Clinical research informed by and informing epidemiological studies of sleep. Journal of Psychosomatic Research, 53 (1), 555559. WallaceGuy, G. M., Kripke, D. F., JeanLouis, G., Langer, R. D., Elliott, J. A., & Tuunainen, A. (2002). Evening light exposure: Implications for sleep and depression. J ournal of the American Geriatrics Society, 50 (4), 738 739. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 54(6), 10631070. Watts, A. G., Swanson, L. W., & Sanchezwatts, G. (1987). Efferent projections of the Suprachiasmatic nucleus: Studies using retrograde transport of fluorescent dyes and simultaneous peptide immunohistochemistry in the rat. Journal of Comparat ive Neurology, 258(2), 204229. Weaver, D. R., & Reppert, S. M. (1989). Direct inutero perception of light by the mammalian fetus. Developmental Brain Research, 47(1), 151 155. West, B. (1985). An essay on the importance of being nonlinear. New York, NY: Springer Verlag. Wever, R. A. (1989). Light effects on human circadianrhythms A review of recent Andechs experiments. Journal of Biological Rhythms, 4(2), 161185. Wicklow, A., & Espie, C. A. (2000). Intrusive thoughts and their relationship to actigra phic measurement of sleep: towards a cognitive model of insomnia. Behaviour Research and Therapy, 38(7), 679693.

PAGE 189

189 Winiger, S. R. (under preparation). Examination and suggested revision of the UWIST Mood Adjective Checklist. Western Kentucky University. W ood, W., Quinn, J. M., & Kashy, D. (2002). Habits in everyday life: Thought, emotion, and action. Journal of Personality and Social Psychology, 83, 16. Woollacott, M. (2000). What are the underlying systems contributing to balance control? Can research on balance in children with cerebral palsy shed light on balance in all populations? Journal of Sport & Exercise Psychology, 22, S5S5. Youngstedt, S. D., Kripke, D. F., Elliott, J. A., Baehr, E. K., & Sepulveda, R. S. (1999). Light exposure, sleep quality, and depression in older adults. In M. F. Holick & E. G. Jung (Eds.), Biologic Effects of Light. Boston, MA: Kluwer Academic Publishers. Zisberg, A., Gur Yaish, N., & Shochat, T. (2010). Contribution of Routine to Sleep Quality in Community Elderly. Sleep 33(4), 509514. Zisberg, A., Young, H. M., & Schepp, K. (2009). Development and psychometric testing of the Scale of Older Adults' Routine. Journal of Advanced Nursing, 65(3), 672683.

PAGE 190

190 BIOGRAPHICAL SKETCH Natalie Deidre Dautovich obtained her Bachelo r of Arts degree in psychology from the University of Alberta, Edmonton, Alberta, Canada in 2000. She completed her Master of Science degree majoring in c ounseling psychology from the University of Florida in 2006.