Delayed Sleep Phase
Conditions
Keywords
sleep, circadian, adolescence, substance use, reward, impulse control
Brief summary
This study will (1) comprehensively characterize the substance use disorder (SUD) risk profile associated with adolescent Delayed Sleep Phase (DSP), and (2) probe whether SUD risk is diminished by altering sleep/circadian timing.
Detailed description
Mounting evidence indicates that delayed sleep phase (DSP) may confer risk for adolescent substance use (SU) and SUDs. However, the exact nature of this link and the mechanisms underlying it remain unclear. Circadian misalignment, a mismatch between late sleep hours and early school start times, is a compelling potential contributor to elevated SU in adolescent DSP with plausible neurobehavioral mechanisms. The investigators hypothesize that DSP-associated circadian misalignment decreases impulse control and increases reward sensitivity, thereby increasing SUD risk. This study will, for the first time, (1) comprehensively characterize the SUD risk profile associated with adolescent DSP, and (2) probe whether SUD risk is diminished by altering sleep/circadian timing. The study will assess both established markers of SUD risk and putative neurobehavioral mechanisms (impulsivity and reward sensitivity). Specifically, the investigators will employ a comprehensive, multi-method approach to examining DSP's role in SUD risk, combining laboratory, experimental, and longitudinal studies. The investigators will recruit a sample of 150 eleventh and twelfth graders (16-19 y/o), divided between 100 DSP and 50 normal phase teens. The investigators will focus on cannabis and alcohol use given their prevalent use in adolescents and evident links to DSP. In the laboratory study, the investigators will compare a group of DSP adolescents to a group of normal phase adolescents on behavioral and neuroimaging (fMRI) tasks tapping impulsivity and reward sensitivity, as well as a circadian phase assessment. In the experimental study, the investigators will probe whether stabilizing circadian phase in the DSP group (n=100) by using sleep scheduling and chronotherapeutic approaches (i.e., dim light in the evening and bright light in the morning) improves sleep and neurobehavioral function relevant to SUD risk. NOTE: When this ClinicalTrials.gov protocol was initially submitted, there were some mistakes made. The initial submission focused only on the Experimental study, which thus only included the DSP group (aka Late Sleep Timing group), and thus out the Laboratory study along with the normal phase group (aka Early/Middle Sleep Timing group). At that time, we also only listed a limited range of the primary outcomes listed in the funded grant, inadvertently leaving out several primary outcomes (weekday sleep duration - actigraph, circadian timing - dim light melatonin onset, neural correlates of reward receipt, and baseline cannabis and alcohol use). Finally, we mistakenly listed cannabis use from the Longitudinal protocol as a secondary outcome when it was actually an exploratory outcome in the funded grant, and thus we removed it.
Interventions
Participants will wear Re-Timer bright glasses for 30 minutes each morning
Participants will wear tinted glasses that block blue wavelength light for 2 hours before bed
BEHAVIORALSleep scheduling
Participants will advance their weekday bedtime and maintain their weekday risetime on weekends
BEHAVIORALMonitor sleep, mood, and substance use
Participants will monitor sleep, mood, and substance use via smartphone-based platform and wrist actigraphy
Sponsors
National Institute on Drug Abuse (NIDA)
University of Pittsburgh
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
BASIC_SCIENCE
Masking
NONE
Intervention model description
This study combines Laboratory, Experimental, and Longitudinal protocols. The study includes 2 initial groups (Early/Middle Sleep Timing and Late Sleep Timing) that all complete the initial Laboratory (baseline) protocol. The Late Sleep participants are also randomized to complete one of two arms (Manipulation or Control) in the Experimental (intervention) protocol. The Early/Mid Sleep group does not complete the Experimental protocol. Finally, participants are also follow-up assessments through the life of the grant in the Longitudinal protocols.
Eligibility
Sex/Gender
ALL
Age
16 Years to 19 Years
Healthy volunteers
Yes
Inclusion criteria
* Age 16-19 years * Currently in 11th or 12th grade and enrolled in a traditional high-school; or cyber school with synchronous classes (not home-schooled) * Physically and psychiatrically healthy, as determined by instruments described below * Provision of written informed consent and assent Additional inclusion criterion for Experimental protocol * Meets operational definition of delayed sleep phase (DSP; weekend bedtime ≥1 AM)
Exclusion criteria
* Significant or unstable acute or chronic medical conditions * Past or current bipolar disorder or psychotic disorder * Past or current substance use disorder other than alcohol use disorder or cannabis use disorder * Past month recreational drug use other than alcohol, cannabis, and nicotine * Current syndromal sleep disorders other than insomnia and delayed sleep phase disorder * Medications that interfere with sleep and/or reward function (antidepressants, and stimulants prescribed for ADHD are permitted) * Conditions that would interfere with the MRI procedures (e.g., non-removal ferromagnetic devices)
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Weekday Sleep Duration - Actigraphy | T1 (1 Week), T2 (2 Weeks) | Total Sleep Time as determined by wrist actigraphy data (averaged across weekdays during 1 week of T1 and during 2 weeks of T2) |
| Circadian Timing - Dim Light Melatonin Onset | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. | Circadian Timing as determined by dim light melatonin onset (DLMO) assessed during saliva sampling using the 4pg/ml threshold. |
| Circadian Alignment | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. Based on DLMO assessed on weeknight lab overnight visit, and including the two nights of actigraphy data prior to the lab visit. | Circadian alignment is operationalized as the interval between the dim light melatonin onset (DLMO) and sleep midpoint based on the prior two nights of actigraphy data. |
| Reward Motivation (Behavioral) | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. | Adjusted average pumps on Balloon Analogue Risk Task, a computerized measure of risk taking behavior in participants are presented with a series of balloons and offered the chance to earn money by pumping each balloon up by clicking a button. The adjusted average only includes non-burst trials. |
| Behavioral Inhibition | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. | Accuracy on Cued Go/No-Go Task, specifically correct response (withholding response) on No-Go trials following an incongruent Go cue |
| Neural Correlates of Reward Anticipation | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. | Activation within the reward network during the Monetary Incentive Delay task. Specifically, activation is defined as bold signal in regions of the reward network (from NeuroSynth) on reward anticipation trials (large reward) versus neutral (no money) trials. Higher values represent increased reactivity to reward, as compared to neutral trials. |
| Neural Correlates of Reward Receipt | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. | Monetary Incentive Delay Task: Win Outcome vs No Win contrast within the reward network (from Neurosynth). Higher values represent increased reactivity to reward wins, as compared to neutral trials. |
| Neural Correlates of Impulse Control | Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. | Activation within the Executive Control Network during the Stop Signal Task. Specifically, activation is defined as bold signal in regions of the Executive Control Network on unsuccessful Stop trials versus successful Go trials. Higher values represent increased activity to unsuccessful Stop versus successful Go trials. |
| Cannabis Use | Days of cannabis use in 3 months prior to baseline. | Days of cannabis use based on timeline followback interview administered during baseline interview during consent/diagnostic interview visit. |
| Alcohol Use | 3 months prior to baseline, based on timeline followback interview. | Days of alcohol use based on timeline followback interview administered during baseline interview during consent/diagnostic interview visit. |
Countries
United States
Participant flow
Recruitment details
Participants were recruited from the surrounding community using flyers, postings on Pitt+Me (a registry for clinical and translational science research) and Read Green, word-of-mouth referrals, social media (Facebook/Instagram, Snapchat, YouTube, Spotify), Peach Jar (an online source for distributing flyers to schools), bus or church advertisements, local teen agencies, and school recruitment (contacting guidance counselors, teachers, and administrative employees).
Pre-assignment details
The 142 enrolled participants include individuals who consented to the study, were randomized to a group, and completed a limited battery of baseline assessments during that consent/diagnostic interview visit, but did not necessarily continue on into the Laboratory (Baseline/T1) protocol.
Participants by arm
| Arm | Count |
|---|---|
| Early/Middle Sleep Timing Group definition: Participants who report a weekend bedtime \<1AM.
Also known as the Laboratory protocol.
Participants are asked to complete the following during a 7-day baseline period:
\- Monitor sleep, mood, and substance use via smartphone-based platform and wrist actigraph At the conclusion of the 7 days, participants are asked to complete an overnight visit in the sleep lab. | 42 |
| Late Sleep Timing - Control Group definition: Participants who reported a weekend bedtime ≥ 1 AM.
Participants were randomized into either the Manipulation or Control Group.
Also known as the Laboratory Protocol.
Participants are asked to complete the following during a 7-day baseline period:
\- Monitor sleep, mood, and substance use via smartphone-based platform and wrist actigraph≥ At the conclusion of the 7 days, participants are asked to complete an overnight visit in the sleep lab. | 40 |
| Late Sleep Timing - Manipulation Group definition: Participants who reported a weekend bedtime ≥ 1 AM.
Participants were randomized into either the Manipulation or Control Group.
Also known as the Laboratory Protocol.
Participants are asked to complete the following during a 7-day baseline period:
\- Monitor sleep, mood, and substance use via smartphone-based platform and wrist actigraph≥ At the conclusion of the 7 days, participants are asked to complete an overnight visit in the sleep lab. | 40 |
| Total | 122 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 | FG002 |
|---|---|---|---|---|
| T1: Laboratory: 1 Week | Investigator withdrawn due to ineligibility learned post-enrollment. | 1 | 0 | 0 |
| T1: Laboratory: 1 Week | Lost to Follow-up | 7 | 1 | 1 |
| T1: Laboratory: 1 Week | Withdrawal by Subject | 6 | 1 | 3 |
Baseline characteristics
| Characteristic | Early/Middle Sleep Timing | Late Sleep Timing - Control | Late Sleep Timing - Manipulation | Total |
|---|---|---|---|---|
| Age, Continuous | 17.4 Age (in years) STANDARD_DEVIATION 0.64 | 17.6 Age (in years) STANDARD_DEVIATION 0.55 | 17.4 Age (in years) STANDARD_DEVIATION 0.74 | 17.5 Age (in years) STANDARD_DEVIATION 0.65 |
| Ethnicity (NIH/OMB) Hispanic or Latino | 2 Participants | 4 Participants | 3 Participants | 9 Participants |
| Ethnicity (NIH/OMB) Not Hispanic or Latino | 40 Participants | 36 Participants | 37 Participants | 113 Participants |
| Ethnicity (NIH/OMB) Unknown or Not Reported | 0 Participants | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) American Indian or Alaska Native | 0 Participants | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) Asian | 0 Participants | 2 Participants | 0 Participants | 2 Participants |
| Race (NIH/OMB) Black or African American | 9 Participants | 6 Participants | 4 Participants | 19 Participants |
| Race (NIH/OMB) More than one race | 1 Participants | 3 Participants | 1 Participants | 5 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 0 Participants | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 0 Participants | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) White | 32 Participants | 29 Participants | 35 Participants | 96 Participants |
| Sex: Female, Male Female | 28 Participants | 22 Participants | 26 Participants | 76 Participants |
| Sex: Female, Male Male | 14 Participants | 18 Participants | 14 Participants | 46 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk |
|---|---|---|---|
| deaths Total, all-cause mortality | 0 / 42 | 0 / 40 | 0 / 40 |
| other Total, other adverse events | 0 / 42 | 0 / 40 | 0 / 40 |
| serious Total, serious adverse events | 0 / 42 | 0 / 40 | 0 / 40 |
Outcome results
Primary
Alcohol Use
Days of alcohol use based on timeline followback interview administered during baseline interview during consent/diagnostic interview visit.
Time frame: 3 months prior to baseline, based on timeline followback interview.
Population: The numbers in this section include all participants who consented, were determined eligible and interested, and started the Laboratory (T1) protocol.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| Early/Middle Sleep Timing | Alcohol Use | 0.381 Days of use | Standard Deviation 0.697 |
| Late Sleep Timing - Control | Alcohol Use | 1.275 Days of use | Standard Deviation 2.961 |
| Late Sleep Timing - Manipulation | Alcohol Use | 1.325 Days of use | Standard Deviation 2.515 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a Poisson generalized linear model (GLM) to test the effect of group (Early/Mid vs Late) on a binary alcohol use outcome (yes/no in the past 3 months), accounting for age and sex.p-value: 0.46195% CI: [0.704, 2.171]Poisson GLM with robust standard error
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a negative binomial GLM to test the effect of group (Early/Mid vs Late) on days of alcohol use, accounting for age and sex.p-value: 0.00795% CI: [1.36, 7.9]negative binomial glm
Primary
Behavioral Inhibition
Accuracy on Cued Go/No-Go Task, specifically correct response (withholding response) on No-Go trials following an incongruent Go cue
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday.
Population: The numbers displayed in this section include the usable behavioral computer task data collected at T1 overnight visits. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Behavioral Inhibition | T1 | 0.910 Proportion accurate responses out of 1 | Standard Deviation 0.108 |
| Late Sleep Timing - Control | Behavioral Inhibition | T1 | 0.891 Proportion accurate responses out of 1 | Standard Deviation 0.096 |
| Late Sleep Timing - Control | Behavioral Inhibition | T2 | 0.861 Proportion accurate responses out of 1 | Standard Deviation 0.131 |
| Late Sleep Timing - Manipulation | Behavioral Inhibition | T1 | 0.846 Proportion accurate responses out of 1 | Standard Deviation 0.124 |
| Late Sleep Timing - Manipulation | Behavioral Inhibition | T2 | 0.829 Proportion accurate responses out of 1 | Standard Deviation 0.155 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.02495% CI: [0.07, 0.94]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoint, and sex.p-value: 0.6695% CI: [-0.55, 0.35]Mixed Models Analysis
Primary
Cannabis Use
Days of cannabis use based on timeline followback interview administered during baseline interview during consent/diagnostic interview visit.
Time frame: Days of cannabis use in 3 months prior to baseline.
Population: The numbers in this section include all participants who consented, were determined eligible and interested, and started the Laboratory (T1) protocol.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| Early/Middle Sleep Timing | Cannabis Use | 2.167 Days of use | Standard Deviation 5.708 |
| Late Sleep Timing - Control | Cannabis Use | 3.200 Days of use | Standard Deviation 10.115 |
| Late Sleep Timing - Manipulation | Cannabis Use | 2.148 Days of use | Standard Deviation 6.928 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a Poisson generalized linear model (GLM) to test the effect of group (Early/Mid vs Late) on a binary cannabis use outcome (yes/no in the past 3 months), accounting for age and sex.p-value: 0.4295% CI: [0.374, 1.513]Poisson GLM with robust standard error
Primary
Circadian Alignment
Circadian alignment is operationalized as the interval between the dim light melatonin onset (DLMO) and sleep midpoint based on the prior two nights of actigraphy data.
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday. Based on DLMO assessed on weeknight lab overnight visit, and including the two nights of actigraphy data prior to the lab visit.
Population: The numbers displayed in this section include the participants with usable data collected during (1) the saliva sampling protocol at T1 overnight visits and (2) the actigraphy protocol in the two nights prior to the T1 visit. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Circadian Alignment | T1 | 6.071 Hours | Standard Deviation 1.111 |
| Late Sleep Timing - Control | Circadian Alignment | T1 | 5.798 Hours | Standard Deviation 1.142 |
| Late Sleep Timing - Control | Circadian Alignment | T2 | 5.910 Hours | Standard Deviation 1.463 |
| Late Sleep Timing - Manipulation | Circadian Alignment | T1 | 5.937 Hours | Standard Deviation 1.071 |
| Late Sleep Timing - Manipulation | Circadian Alignment | T2 | 5.586 Hours | Standard Deviation 1.243 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.36595% CI: [-0.64, 0.24]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoints, and sex.p-value: 0.06795% CI: [-0.88, 0.03]Mixed Models Analysis
Primary
Circadian Timing - Dim Light Melatonin Onset
Circadian Timing as determined by dim light melatonin onset (DLMO) assessed during saliva sampling using the 4pg/ml threshold.
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday.
Population: The numbers displayed in this section include the usable data collected during the saliva sampling protocol at T1 overnight visits. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Circadian Timing - Dim Light Melatonin Onset | T1 | 20.707 Clock Time (24-hour clock) | Standard Deviation 1.112 |
| Late Sleep Timing - Control | Circadian Timing - Dim Light Melatonin Onset | T1 | 21.849 Clock Time (24-hour clock) | Standard Deviation 1.395 |
| Late Sleep Timing - Control | Circadian Timing - Dim Light Melatonin Onset | T2 | 22.038 Clock Time (24-hour clock) | Standard Deviation 1.501 |
| Late Sleep Timing - Manipulation | Circadian Timing - Dim Light Melatonin Onset | T1 | 21.789 Clock Time (24-hour clock) | Standard Deviation 1.228 |
| Late Sleep Timing - Manipulation | Circadian Timing - Dim Light Melatonin Onset | T2 | 21.194 Clock Time (24-hour clock) | Standard Deviation 1.355 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: <0.00195% CI: [0.61, 1.59]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoints, and sex.p-value: <0.00195% CI: [-1.09, -0.43]Mixed Models Analysis
Primary
Neural Correlates of Impulse Control
Activation within the Executive Control Network during the Stop Signal Task. Specifically, activation is defined as bold signal in regions of the Executive Control Network on unsuccessful Stop trials versus successful Go trials. Higher values represent increased activity to unsuccessful Stop versus successful Go trials.
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday.
Population: The numbers displayed in this section include the usable fMRI task data collected at the T1 visit. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Neural Correlates of Impulse Control | T1 | 0.047 percent signal change | Standard Deviation 0.08 |
| Late Sleep Timing - Control | Neural Correlates of Impulse Control | T1 | 0.066 percent signal change | Standard Deviation 0.063 |
| Late Sleep Timing - Control | Neural Correlates of Impulse Control | T2 | 0.049 percent signal change | Standard Deviation 0.071 |
| Late Sleep Timing - Manipulation | Neural Correlates of Impulse Control | T1 | 0.070 percent signal change | Standard Deviation 0.072 |
| Late Sleep Timing - Manipulation | Neural Correlates of Impulse Control | T2 | 0.054 percent signal change | Standard Deviation 0.078 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.12495% CI: [-0.01, 0.06]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoint, and sex.p-value: 0.92595% CI: [-0.05, 0.06]Mixed Models Analysis
Primary
Neural Correlates of Reward Anticipation
Activation within the reward network during the Monetary Incentive Delay task. Specifically, activation is defined as bold signal in regions of the reward network (from NeuroSynth) on reward anticipation trials (large reward) versus neutral (no money) trials. Higher values represent increased reactivity to reward, as compared to neutral trials.
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday.
Population: The numbers displayed in this section include the usable fMRI task data collected at the T1 visit. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Neural Correlates of Reward Anticipation | T1 | 0.040 percent signal change | Standard Deviation 0.086 |
| Late Sleep Timing - Control | Neural Correlates of Reward Anticipation | T1 | 0.027 percent signal change | Standard Deviation 0.123 |
| Late Sleep Timing - Control | Neural Correlates of Reward Anticipation | T2 | 0.007 percent signal change | Standard Deviation 0.125 |
| Late Sleep Timing - Manipulation | Neural Correlates of Reward Anticipation | T1 | 0.074 percent signal change | Standard Deviation 0.132 |
| Late Sleep Timing - Manipulation | Neural Correlates of Reward Anticipation | T2 | 0.030 percent signal change | Standard Deviation 0.087 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.59295% CI: [-0.03, 0.06]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoint, and sex.p-value: 0.5595% CI: [-0.11, 0.06]Mixed Models Analysis
Primary
Neural Correlates of Reward Receipt
Monetary Incentive Delay Task: Win Outcome vs No Win contrast within the reward network (from Neurosynth). Higher values represent increased reactivity to reward wins, as compared to neutral trials.
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday.
Population: The numbers displayed in this section include the usable fMRI task data collected at the T1 visit. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Neural Correlates of Reward Receipt | T1 | -0.009 percent signal change | Standard Deviation 0.103 |
| Late Sleep Timing - Control | Neural Correlates of Reward Receipt | T1 | -0.028 percent signal change | Standard Deviation 0.135 |
| Late Sleep Timing - Control | Neural Correlates of Reward Receipt | T2 | 0.009 percent signal change | Standard Deviation 0.1 |
| Late Sleep Timing - Manipulation | Neural Correlates of Reward Receipt | T1 | -0.019 percent signal change | Standard Deviation 0.139 |
| Late Sleep Timing - Manipulation | Neural Correlates of Reward Receipt | T2 | -0.008 percent signal change | Standard Deviation 0.074 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.5595% CI: [-0.07, 0.04]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoint, and sex.p-value: 0.55695% CI: [-0.1, 0.06]Mixed Models Analysis
Primary
Reward Motivation (Behavioral)
Adjusted average pumps on Balloon Analogue Risk Task, a computerized measure of risk taking behavior in participants are presented with a series of balloons and offered the chance to earn money by pumping each balloon up by clicking a button. The adjusted average only includes non-burst trials.
Time frame: Overnight visits at end of T1 (1 Week) and T2 (2 Weeks). Always occurred on a Wednesday or Thursday.
Population: The numbers displayed in this section include the usable behavioral computer task data collected at T1 overnight visits. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Reward Motivation (Behavioral) | T1 | 32.609 Adjusted Average Number of pumps | Standard Deviation 13.274 |
| Late Sleep Timing - Control | Reward Motivation (Behavioral) | T1 | 37.696 Adjusted Average Number of pumps | Standard Deviation 15.364 |
| Late Sleep Timing - Control | Reward Motivation (Behavioral) | T2 | 43.961 Adjusted Average Number of pumps | Standard Deviation 13.508 |
| Late Sleep Timing - Manipulation | Reward Motivation (Behavioral) | T1 | 34.448 Adjusted Average Number of pumps | Standard Deviation 14.416 |
| Late Sleep Timing - Manipulation | Reward Motivation (Behavioral) | T2 | 34.117 Adjusted Average Number of pumps | Standard Deviation 11.244 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.23495% CI: [-2.18, 8.82]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoint, and sex.p-value: 0.01895% CI: [-11.25, -1.06]Mixed Models Analysis
Primary
Weekday Sleep Duration - Actigraphy
Total Sleep Time as determined by wrist actigraphy data (averaged across weekdays during 1 week of T1 and during 2 weeks of T2)
Time frame: T1 (1 Week), T2 (2 Weeks)
Population: The numbers displayed in this section include the usable actigraph data collected for participants during T1. Early/Mid Sleep group only assessed at T1.
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Early/Middle Sleep Timing | Weekday Sleep Duration - Actigraphy | T1 | 6.309 Number of hours | Standard Deviation 1.07 |
| Late Sleep Timing - Control | Weekday Sleep Duration - Actigraphy | T1 | 5.866 Number of hours | Standard Deviation 0.963 |
| Late Sleep Timing - Control | Weekday Sleep Duration - Actigraphy | T2 | 5.845 Number of hours | Standard Deviation 0.817 |
| Late Sleep Timing - Manipulation | Weekday Sleep Duration - Actigraphy | T1 | 5.880 Number of hours | Standard Deviation 0.765 |
| Late Sleep Timing - Manipulation | Weekday Sleep Duration - Actigraphy | T2 | 6.554 Number of hours | Standard Deviation 0.801 |
Comparison: For these Aim 1 analyses comparing the Early/Middle vs Late Sleep groups, the Late Sleep-Control and Late Sleep-Manipulation were combined into a single Late Sleep group. We used a linear regression model to test the effect of group (Early/Mid vs Late) on each a priori outcome, accounting for age and sex.p-value: 0.01895% CI: [-0.77, -0.07]Regression, Linear
Comparison: For Aim 2, this mixed effect model tested the effect of group (Control vs Manipulation) on changes in each a priori outcome, accounting for group, timepoints, and sex.p-value: 0.00195% CI: [0.28, 1.07]Mixed Models Analysis