Version of Record: https://www.sciencedirect.

com/science/article/pii/S0028393222002330
Manuscript_8831d5b09f45ce2db094fc4b46058737

Sleep duration moderates the associations between immune markers and corticolimbic

function during stress in adolescents

Jessica P. Uy1, Macrina Dieffenbach1, Carrianne J. Leschak1, Naomi I. Eisenberger1,

Andrew J. Fuligni1,2, Adriana Galván1

1Department of Psychology, University of California, Los Angeles, Los Angeles, CA

2Department of Psychiatry and Biobehavioral Sciences, University of California, Los

Angeles, Los Angeles, CA

Corresponding Author:

Dr. Jessica P. Uy

1285 Franz Hall, Box 951563

University of California, Los Angeles

Los Angeles, CA 90095

Email: [email protected]

The authors have no conflicts of interest to declare.

© 2022 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
 Abstract

Adolescence is characterized by biological changes in hormonal and circadian systems

that, with concurrent psychosocial changes, result in increased sleep disturbances and

stress sensitivity. Sleep disturbance has been associated with heightened stress

sensitivity and elevated levels of inflammation in adults and adolescents, yet the neural

correlates are unknown in adolescents. The current study investigated whether and how

individual differences in peripheral immune markers (IL-6, TNF-α) related to neural

response to stress in adolescents and whether these immune-brain associations were

moderated by adolescents’ sleep duration. Thirty-seven adolescents (14-15 years) who

met quality control criteria for fMRI reported daily sleep duration for 7 days and

performed an fMRI stressor task. A subsample of 23 adolescents additionally provided

blood samples that were assayed for inflammatory markers using a multiplex assay.

Results revealed that average sleep duration moderated associations between TNF-α

and medial frontolimbic circuitry (amygdala, medial prefrontal cortex) during the stressor

task such that, among adolescents who reported shorter sleep duration, higher levels of

TNF-α were associated with greater deactivation in those regions during stress, which

was associated with greater self-reported anxiety. These findings suggest that

insufficient sleep duration coupled with greater levels of peripheral inflammation may

promote a neural profile characterized by alterations in frontolimbic circuitry during

stress, which can exacerbate sleep disturbances and/or peripheral inflammation.

Keywords: inflammation, stress, adolescents, sleep, amygdala, prefrontal cortex

 Sleep duration moderates the associations between immune markers and

corticolimbic function during stress in adolescents

During adolescence, there is a shift in chronotype such that adolescents prefer a

later bedtime and waketime (Carskadon et al., 1993). Sleep pressure, the homeostatic

mechanism by which the need for sleep increases, also accumulates at a slower rate

(Jenni et al., 2005). Coupled with early school start times (Carskadon et al., 1998),

adolescents represent one of the most sleep-deprived populations (CDC, 2011; Kann et

al., 2014). In addition to negatively affecting learning and memory (Walker & Stickgold,

2006), cognition (Anderson & Platten, 2011; Telzer et al., 2013), and decision making

(Killgore et al., 2011), individuals with insufficient sleep are also more likely to

experience physical and psychological health problems (Vgontzas et al., 2004). For

example, across healthy and clinical populations, various forms of poor sleep (e.g.,

experimental partial or total sleep deprivation, naturalistic sleep disturbance) have been

associated with elevated levels of inflammatory markers such as C-reactive protein

(CRP), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α) (Irwin, 2015; Irwin

et al., 2016). However, the neural correlates of sleep and inflammation are unknown in

adolescents.

Sleep influences the systems that respond to stress (e.g., sympathetic nervous

system [SNS] and hypothalamic-pituitary-adrenal [HPA] axis) (Irwin, 2015), which have

been shown to regulate immune responses (Irwin, 2019). Sleep disruption results in

increased SNS and HPA axis activity during sleep, which have implications for stress

responding while awake. Indeed, compared to well-rested adults, sleep-deprived adults

exhibited higher baseline cortisol levels and heightened cortisol response to

psychosocial stress (Minkel et al., 2014), suggesting that poor sleep might influence

immune function and health by sensitizing the systems that respond to stress.

Surprisingly, the effects of sleep and stress processes on brain development and

immune activity are rarely studied together in adolescents. The confluence of changes

in stress reactivity, sleep habits, and corticolimbic circuitry that occur during

adolescence confers a period of vulnerability to negative health outcomes. Burgeoning

research examining the links between sleep habits and immune markers during

adolescence has found that shorter sleep duration were associated with higher levels of

CRP in adolescents (Park et al., 2016, 2020). In relation to the upstream molecular

immune processes, shorter sleep duration was associated with greater gene expression

of pro-inflammatory proteins, increased signaling of pro-inflammatory transcription factor

NF-kB, downregulation of antiviral gene expression, and decreased signaling of

interferon (IFN) response factors in adolescents (Chiang et al., 2019). Moreover, shorter

sleep duration strengthened the associations between daily stress and NF-kB activity.

That is, greater daily stress was more strongly associated with greater inflammatory NF-

kB activity among adolescents with shorter sleep duration (Chiang et al., 2019). These

findings suggest that one way by which insufficient sleep might contribute to increased

inflammation may be through sensitizing the brain to stress, which would amplify the

body’s inflammatory state and compound the immune system’s effects on the brain.

However, the effects of sleep and inflammation on the developing brain’s response to

stress are unknown.

In addition to sleep influencing neural and immune responses, research has also

shown that heightened inflammation can enhance neural responses to stressful

experiences. Extant research on immune-to-brain signaling in humans demonstrated

that, compared to placebo, inflammatory challenge was associated with heightened

neural reactivity to negative social experiences in regions implicated in socioemotional

and pain processing, particular in the dorsal anterior cingulate cortex (dACC), anterior

insula, and amygdala (Eisenberger et al., 2009, 2017; Inagaki et al., 2012; Muscatell et

al., 2016). Moreover, those who showed greater increases in pro-inflammatory

cytokines (IL-6) in response to the inflammatory challenge showed greater activity in the

dACC and anterior insula in response to social exclusion (Eisenberger et al., 2009) and

increased depressed mood (Eisenberger et al., 2010; Moieni et al., 2015; Reichenberg

et al., 2001). In addition to heightened threat sensitivity, peripheral immune markers

have also been shown to influence prefrontal cortex (PFC) structure and associated

functioning (Harrison et al., 2009; Marsland et al., 2008). A meta-analysis on the

associations between peripheral immune markers and brain function in adults revealed

that inflammatory markers showed consistent effects in limbic and basal ganglia regions

(amygdala, hippocampus, striatum, thalamus), brainstem regions, cortical regions

(ACC, dorsomedial PFC, ventromedial PFC, orbitofrontal cortex, insula), and temporal

regions (Kraynak et al., 2018).

Given the protracted development of the PFC into adulthood, the implications of

inflammatory processes sensitizing an already sensitive limbic system in conjunction

with diminishing function of a developing regulatory system for mental and physical

health warrants further research in adolescents. However, very few studies have

investigated the associations between inflammatory processes and frontolimbic circuitry

function in adolescents. One study found that amygdala reactivity to threatening faces
was positively associated with inflammation (a standardized composite of CRP, IL-6, IL-

8, IL-10, and TNF-α) among adolescents who lived in poverty (Miller et al., 2020).

Studies examining the associations between peripheral inflammatory markers and

resting-state functional connectivity in adolescents found that higher levels of

inflammation were associated with lower functional connectivity in the emotional

regulation and central executive networks (Nusslock et al., 2019) and altered

connectivity within limbic and frontoparietal networks in adolescents (Swartz et al.,

2021). Together, these studies suggest that corticolimbic circuitry are targets of

inflammation during adolescence.

The current study investigated whether and how individual differences in

peripheral immune markers related to neural response to stress in adolescents and

whether these immune-brain associations were moderated by adolescents’ sleep

duration. We focused on circulating levels of IL-6 and TNF-α because of their previous

associations with sleep and stress processes. We hypothesized that higher levels of

peripheral pro-inflammatory markers would be associated with heightened neural

response to stress in regions previously shown to respond to stress (e.g., anterior

insula, anterior cingulate cortex, amygdala) and/or diminished response in prefrontal

regions. We also hypothesized that the associations between peripheral immune

markers and neural response to stress would be stronger among adolescents who

reported shorter sleep duration.

2. Materials and Methods

2.1. Participants
 Sleep diaries and neuroimaging data were collected from 40 adolescents (14.03-

15.99 years, M = 15.076, SD = 0.646, 17 females). Participants were recruited using

flyers posted in local child and adolescent-friendly locations, on community websites,

flyer distributions at local high schools, and patient databases from the University of

California, Los Angeles (UCLA) Clinical and Translational Science Institute. Inclusion

criteria required all participants be right-handed, free from metal objects in the body,

speak fluent English, be in the appropriate age range, and have no previously

diagnosed psychiatric, neurological, or developmental disorders. Parents of adolescent

participants provided written consent and adolescents provided assent in accordance

with the UCLA Institutional Review Board. Participants were also provided the

opportunity to consent to an optional blood draw. All participants were compensated for

their participation.

Of the 40 participants, one participant was excluded from neuroimaging analyses

due to a neuroanatomical abnormality and two participants were excluded for excessive

motion across both runs of the task. Of the remaining 37 (18 females) participants with

usable neuroimaging data, 23 (62%; 9 females) participated in the blood draw. Analyses

were conducted with the maximum number of subjects for each analysis.

2.2. Procedure

Data were collected between January 2018 and October 2019. Participants

completed two visits at UCLA. During the first visit, after providing consent, participants

completed questionnaires about demographic information and were trained on how to

complete the daily diary measures. Research suggests good concordance between

daily sleep indices measured via self-report and actigraphy in adolescents (Lucas-
Thompson et al., 2021). For 7 days after the first visit, participants received a text

message each evening with a URL to an online survey asking them to complete

information about their day. After 7 days (but within two weeks), participants returned to

UCLA to complete their second visit. Participants who consented to the blood draw had

their blood drawn by a certified phlebotomist at the clinical lab in the Peter Morton

Medical Building at UCLA. After the blood draw, participants completed a brain scan

while performing the fMRI stressor task at the Center for Cognitive Neuroscience (CCN)

at UCLA. Participants who did not consent to the blood draw only completed the brain

scan portion of the study. Participants’ height and weight were measured to calculate

body mass index (BMI). BMI ranged from 14.337 to 45.154 (M = 23.298, SD = 6.299).

After the brain scan, participants completed additional questionnaires about their

experiences regarding the stressor task, were debriefed about the goals of the study,

and received compensation.

2.3. fMRI Stressor Task

The current study used a modified version of the well-validated Montreal Imaging

Stress Task (MIST) (Dedovic et al., 2005), which has been published previously

(Inagaki et al., 2016). During the stressor task, participants were asked to perform a

series of mental arithmetic challenges that have social evaluative components

integrated into the task (Figure 1). To assess the effects of stress, the stressor task

consisted of 2 experimental conditions (practice and test) that were presented in an

alternating block design. In the practice condition, participants completed a series of

easy mental arithmetic problems on the computer screen. Each series or block

contained 6 trials. On practice trials, easy arithmetic problems with no answer choices
were shown. Participants were given 5 seconds to solve each problem and were told to

press 1 once they mentally solved each problem. In the test condition, participants

completed a series of challenging mental arithmetic problems on the computer screen.

Each series or block contained 6 trials. On test trials, challenging arithmetic problems

with 4 possible answer choices were presented to participants and they had 5 seconds

to choose the correct answer before time ran out. The difficulty of the problems in the

test condition were chosen to be just slightly beyond individuals’ mental capacity to

solve within the time limit, though it is possible to solve the problems within the time

limit. After each arithmetic problem in the test condition, participants were shown

feedback on their performance (i.e., “correct”, “incorrect”, or “out of time” if participants

did not choose an answer in time). At the end of each test block, participants were also

shown a rating scale of their performance relative to that of their peers to increase the

social evaluative threat of the task. This performance evaluation rating was

manipulated, such that the participants’ performance evaluation rating declined over

time and at a faster rate than that of their peers. Participants were told that the

performance rating takes into account their accuracy and speed on the test trials to

circumvent suspicion of deception in those who may have better accuracy. After each

experimental block, participants were asked to rate their stress levels (1 = not at all

stressful, 4 = very stressful). Participants performed 4 practice blocks and 4 test blocks

that alternated in sequence.

Behavioral indices assessed from the task include average stress ratings for

each condition and average accuracy on the test (total number of correct responses

divided by the total number of test trials administered).

Figure 1. Diagram of the fMRI Stressor Task.

2.4. Measures

2.4.1. Sleep Duration. Each night, participants were asked to report how much

sleep they received the night before. Daily sleep duration was averaged across the 7

days to assess participants’ average nightly sleep duration. Average weekly sleep

duration for full sample ranged from 259.285 minutes to 585.857 minutes (M = 467.452

minutes, SD = 60.479).

2.4.2. State Trait Anxiety Inventory (STAI). After the fMRI Stressor Task,

participants were asked to indicate to what extent (1 = not at all, 4 = very much so) they

experienced 15 items relating to positive and negative affect and psychosomatic

symptoms during the test trials of the task (e.g., “I felt calm”, “My heart was beating

fast”, “I felt nervous”). After reverse-coding positive items, items on this scale were
summed to create an index of test-related anxiety. Higher scores indicate greater

anxiety symptoms. STAI scores ranged from 16 to 41 (M = 26.54, SD = 5.615).

2.5. Immunological Measures

Blood samples were collected in EDTA tubes. After collection, samples were

centrifuged at 4°C, plasma samples were harvested into multiple aliquots, and stored in

a -80°C freezer until all blood samples for the study have been collected. All plasma

samples from a single subject were assayed together on the same 96-well plate to

minimize effects of inter-assay variation. All samples were assayed in duplicate and an

internal quality control sample was included on every plate. Interleukin-6 (IL-6), IL-8, IL-

10, tumor necrosis factor-alpha (TNF-α), and interferon-gamma (IFNγ) were measured

in a multiplex assay utilizing a V-PLEX Custom Human Cytokine Proinflammatory Panel

on the Meso Scale Discovery (MSD) electrochemiluminesence platform (MSD,

Rockville, MD). Samples were assayed at a 2-fold dilution according to the

manufacturer’s protocol, with an eight-point standard curve with tripling dilutions.

Analyte-specific lower limits were calculated for each assay plate (IL-6: 0.21 pg/mL, IL-

8: 0.17 pg/mL, IL-10: 0.11 pg/mL, TNF-α: 0.11 pg/mL, IFNγ: .42 pg/mL). For all plasma

biomarkers, inter-assay coefficients of variation were less than or equal to 10% and

mean intra-assay coefficients of variation were less than 6.5%.

After excluding one subject with a self-reported acute viral infection and extreme

value on IFNγ (40.49 pg/mL), values for immune markers were natural log-transformed

to correct for non-normality. We focused on circulating levels of IL-6 and TNF-α

because of their previous associations with sleep and stress processes. Descriptive
statistics and bivariate correlations between immune markers, BMI, and sleep measures

are displayed in Table 1. Levels of immune markers did not differ by sex (p’s > .26).

M (SD) TNF-α BMI Sleep

Duration
IL-6 0.59 (.70) .140 .719** .160
TNF-α 2.03 (.33) -.123 -.176
BMI 23.30 (6.39) .072
Sleep 458.79
Duration (71.89)

Table 1. Bivariate correlations between peripheral immune markers, BMI, and sleep

measures. Descriptive statistics are presented in raw values. Correlations were

conducted using natural log-transformed values of immune markers. ** p < .01

2.6. fMRI

2.6.1. fMRI Data Acquisition. Functional imaging data were collected on a 3

Tesla Siemens Magnetom Prisma MRI scanner with a 20-channel head coil using a

gradient-echo, echo-planar image (EPI) sequence (TR = 2000 ms, TE = 30 ms, flip

angle = 90 degrees, FOV = 192 mm, 260 volumes, 34 slices, slice thickness = 4 mm). A

T2-weighted, matched bandwidth (MBW), high-resolution anatomical scan (TR =

5000ms, TE = 35ms, FOV = 192mm, flip angle = 90 degrees, 34 slices, slice thickness

= 4.0 mm) and magnetization-prepared rapid-acquisition gradient echo (MPRAGE) scan

were acquired for registration purposes (TR = 2000 ms, TE = 2.52 ms, FOV = 256 mm,

matrix =, sagittal plane, slice thickness = 1 mm, 192 slices).

2.6.2. fMRI Preprocessing. Preprocessing and statistical analyses were

performed using FMRIB’s Software Library (FSL) 5.0.9. Preprocessing included motion
correction, non-brain matter removal using FSL brain extraction tool (BET), spatial

smoothing (5mm FWHM Gaussian kernel) to increase the signal-to-noise ratio and

filtered in the temporal domain using a nonlinear high-pass filter (100s). Images with

greater than 10% of TRs indicating framewise displacement > .9 mm were excluded

from analyses. EPI images were registered to the MBW scan, then to the MPRAGE

scan, and finally into standard Montreal Neurological Institute (MNI) space (MNI152, T1

2mm) using linear registration with FSL FMRIB’s Linear Image Registration Tool

(FLIRT).

2.7. Analytic Plan

Consistent with previous studies (e.g., Chiang et al., 2019; Miller et al., 2020), all

reported analyses covaried for sex and all analyses consisting of peripheral immune

markers additional covaried for BMI. Because variability in task performance could alter

participants’ psychological experience of the Stressor Task, all fMRI analyses

additionally covaried for task accuracy.

2.7.1. Behavioral Analysis of Stressor Task. Repeated-measures ANCOVAs and

regression analyses were conducted to determine whether stress ratings, test accuracy,

and test-related anxiety (participants’ responses to the test-related STAI) differed by

average sleep duration and immune markers.

2.7.2. fMRI Data Analysis. Imaging data were modeled using a block design.

General linear models (GLM) with multiple explanatory variables (regressors) were used

for fMRI analyses. For each run, 3 explanatory variables were modeled: 1) practice

blocks; 2) test blocks; 3) instruction and stress rating screens. Each explanatory

variable was convolved with a canonical double-gamma hemodynamic response

function (HRF). Onset time for practice and test blocks were defined as the onset of the

first arithmetic problem in each block. Offset time for practice blocks was defined as the

offset of the last arithmetic problem in the practice block. Offset time for each test block

was defined as the offset of the performance rating screen (Inagaki et al., 2016). The

duration of each block was the duration between each blocks’ respective onset and

offset times. “Rest” screens were not explicitly modeled and therefore served as an

implicit baseline.

Analyses focused on the contrast between test blocks and practice blocks (Test

> Practice, Practice > Test). A fixed effects voxel-wise analysis combined each of the

two runs at the second level. Regression analyses were conducted at the group level

using the FMRIB local analysis of mixed effects (FLAME1) module in FSL with mean-

centered regressors of interest entered in each respective model in whole brain

analyses. Consistent with previous research examining sleep and brain function in

adolescents (Baker et al., 2020), Z (Gaussianized T) statistic images were thresholded

at Z > 3.1 by a corrected cluster significant threshold of p < .05 using Gaussian Random

Field theory and corrected for family-wise errors. Anatomical localization within each

cluster were obtained by searching within maximum likelihood regions from the FSL

Harvard-Oxford probabilistic atlas.

Levels of peripheral immune markers were entered as mean-centered regressors

of interest in separate GLMs for whole-brain fMRI analyses to assess their associations

with neural response to stress. Moderation analyses between sleep duration and each

immune marker were conducted at the whole-brain level to assess whether average
sleep duration moderated associations between immune markers and neural response

to stress.

2.7.3. Region-of-interest (ROI) Analyses. In addition to whole-brain analyses,

ROI analyses were also conducted in regions previously implicated in stress reactivity

and regulation (e.g., dACC, left and right anterior insula, left and right amygdala, and left

and right hippocampus). dACC and bilateral anterior insula ROIs were structurally

defined using the Automated Anatomical Labeling (AAL) atlas. The dACC ROI

combined Brodmann Areas 32 and 24 and used a rostral boundary of y = 36 and a

caudal boundary of y = 0 (Dedovic et al., 2016). The anterior insula ROIs were

constructed by dividing the AAL insula ROI at y = 0, approximately separating

dysgranular and granular insula (Slavich et al., 2010). Amygdala and hippocampus

ROIs were anatomically defined using the FSL Harvard-Oxford probabilistic atlas and

thresholded at 50%. (Figure 2). ROI analyses were corrected for multiple comparisons

using Holm-Bonferroni correction adjusted for 28 comparisons (main effects of

inflammation on 7 ROIs [dACC, left and right anterior insula, left and right amygdala, left

and right hippocampus] for 2 immune markers [IL-6, TNF-α], and interactions between

sleep duration and inflammation on 7 ROIs for 2 immune markers), resulting in an

adjusted alpha of .0018.

Figure 2. ROIs for ROI analyses.

2.7.4. Sensitivity Analyses. For each effect found for the Test relative to Practice

contrast, sensitivity analyses were additionally conducted for each condition relative to

implicit baseline separately to determine whether stress effects (Test > Practice) were

driven by greater activation or deactivation during the Test condition compared to

Practice condition.

Additionally, we conducted supplementary analyses where subjects who had test

accuracy scores lower than 25% (chance level) were excluded. Results remained

unchanged and are reported in Supplemental Materials.

3. Results

3.1. Behavioral Results

3.1.1. Stress ratings. On average, participants rated the test block (M = 2.899,

SD = .644) as more stressful than the practice block (M = 1.578, SD = .618), F(1, 35) =

171.809, p < .001. Differences in stress ratings (i.e., psychological stress reactivity) did

not differ by sleep duration (p = .544), levels of IL-6 (p = .985), or TNF-α (p = .530).
 3.1.2. Test accuracy. Accuracy on test problems ranged from 0% to 87.5% (M =

44.6%, SD = 19.79%). Adolescents who reported higher stress ratings on the test had

lower test accuracy (B = -.123, SE = .048, t(34) = -2.536, p = .016). Test accuracy did

not significantly differ by sleep duration (p = .621), levels of IL-6 (p = .403), or TNF-α (p

= .553).

3.1.3. Test-related Anxiety. Test-related anxiety symptoms were not related to

test accuracy (p = .439). Adolescents who endorsed higher stress ratings on the test

reported greater test-related anxiety (B = 5.292, SE = 1.200, t(34) = 4.411, p = < .001.

Test-related anxiety did not differ by sleep duration (p = .892), levels of IL-6 (p = .780),

or TNF-α (p = .353).

3.2. Main effects of fMRI Stressor Task

ROI analyses. Mixed effects tests comparing activation between Test and

Practice conditions (relative to implicit baseline) in ROIs, controlling for sex and task

accuracy, revealed that participants engaged dACC and bilateral anterior insula more

during test relative to practice blocks (dACC: t(33) = 4.807, p < .001; left anterior insula:

t(33) = 3.375, p = .001; right anterior oinsula: t(33) = 3.102, p = .003). In contrast,

participants showed greater deactivation in bilateral amygdala and bilateral

hippocampus during test relative to practice blocks (left amygdala: t(33) = -4.521, p <

.001; right amygdala: t(33) = -3.372, p = .001; left hippocampus: t(33) = -3.536, p =

0.001; right hippocampus: t(33) = -2.908, p = .005) (Supplemental Figure 1).

Regression analyses were conducted to determine whether activation during

Test > Practice in ROIs (dACC, bilateral anterior insula, bilateral amygdala, bilateral

hippocampus) were associated with stress reactivity and test-related anxiety, over and
above sex and task accuracy. Analyses revealed that greater deactivation during Test >

Practice in bilateral amygdala and hippocampus were associated with greater test-

related anxiety (left amygdala: b = -0.1333, SE = 0.0429, t(32) = -3.106, p = .0040; right

amygdala: b = -0.1179, SE = 0.0369, t(32) = -3.196, p = .0031; left hippocampus: b = -

0.1597, SE = 0.06036, t(32) = -2.646, p = .0125; right hippocampus: b = -0.16614, SE =

0.05667, t(32) = -2.932, p = .0062. (Figure 3). In contrast, activation in dACC and

bilateral anterior insula were not associated with test-related anxiety. Additionally,

activation in ROIs during Test > Practice was not associated with differences in stress

ratings.

Figure 3. Greater deactivation in bilateral amygdala and hippocampus during Test >

Practice were associated with greater test-related anxiety, controlling for sex and test

accuracy.
 Whole-brain analyses. Whole-brain analyses revealed one cluster (cluster size =

53787 voxels) with peak activation in bilateral occipital poles that extends into lateral

prefrontal regions (dorsolateral prefrontal cortex, middle frontal gyrus, inferior frontal

gyrus), dACC, anterior insula, orbitofrontal cortex, thalamus, and visual cortex more

during test blocks compared to practice blocks (Test > Practice) (Supplemental Figure

2, Supplemental Table 1). In contrast, participants engaged medial prefrontal regions

(frontal pole, ventromedial prefrontal cortex), hippocampus, posterior insula, posterior

cingulate gyrus, temporal, and occipital regions more during practice than test blocks

(Practice > Test) (Supplemental Figure 3, Supplemental Table 1).

Examining each condition relative to implicit baseline separately, results revealed

one cluster with peak activation in left occipital fusiform gyrus (-26, -82, -14, Z = 8.51,

cluster size = 60533 voxels) that extends into bilateral thalamus, insula, orbitofrontal

cortex, frontal pole, inferior frontal gyri, and anterior cingulate cortex during Test relative

to implicit baseline (Supplemental Table 1). For Practice relative to implicit baseline,

participants engaged similar regions as those in the Test condition: occipital, temporal,

thalamic, insular, cingulate, and prefrontal regions (including frontal pole, inferior frontal

gyrus) regions (Supplemental Table 1).

3.3. Immune markers and neural response to stress

ROI analyses. There were no significant associations between activation during

Test > Practice in ROIs and IL-6 (ps > .141) or TNF-α (ps > .212).

Whole-brain analyses. Whole-brain analyses revealed that levels of IL-6 were

negatively associated with activation in left occipital cortex (left intracalcarine cortex [-8,
-86, 4, Z = 4.43, cluster size = 1179 voxels]) for Test > Practice contrast. There were no

significant associations between activation during Test > Practice and TNF-α.

3.4. Interactions between sleep duration and peripheral immune markers on

neural response to stress

ROI analyses. After correcting for multiple comparisons, there were no significant

interactions between sleep duration and TNF-α or IL-6 on neural response to stress in

ROIs, controlling for gender, BMI, and test accuracy. However, uncorrected, there was

an interaction between sleep duration and TNF-α on neural response to stress in the left

amygdala (b = 0.814, SE = 0.341, t(15) = 2.388, p = .0305). Follow-up simple effects

analyses revealed that among individuals who reported short sleep duration (1 SD

below mean = 384.845 minutes), greater levels of TNF-α were associated with greater

deactivation in left amygdala during Test > Practice (b = -57.2021, SE = 25.598, t(15) =

-2.235, p = .0411). Levels of TNF-α were not associated with left amygdala activation

during Test > Practice among those who reported average (457.248 minutes; b = -7.55,

SE = 18.797, t(15) = -0.402, p = .6936) or long sleep duration (529.651 minutes; b =

42.101, SE = 30.265, t(15) = 1.391, p = .1845) (Supplemental Figure 4).

Whole-brain analyses. Whole-brain analyses testing for interactions between

average sleep duration and immune markers on neural response to stress (Test >

Practice), controlling for sex, BMI, and test accuracy, revealed a significant interaction

between sleep duration and TNF-α in left MPFC (-2, 58, 24, Z = 3.75, cluster size = 125

voxels) during Test > Practice (Figure 4). Parameter estimates (5mm spheres around

peak activation) from left MPFC were extracted to probe the nature of the interaction.

Follow up simple slopes analyses revealed that among individuals who reported short
sleep duration (1 SD below mean = 384.845 minutes), greater levels of TNF-α were

associated with greater deactivation in left MPFC during Test relative to Practice (B = -

208.3376, SE = 81.776, t(14) = -2.548, p = .0232). Additionally, among those who

reported long sleep duration (1 SD above mean = 529.651), greater levels of TNF-α

were associated with less MPFC deactivation during Test relative to Practice (B =

254.627, SE = 95.604, t(14) = 2.663, p = .0185). Among those who reported average

sleep duration, levels of TNF-α were not associated with MPFC activation during Test >

Practice (B = 23.145, SE = 59.346, t(14) = 0.390, p = .702) (Figure 4).

When examining each condition relative to implicit baseline separately (i.e.,

Practice > implicit baseline, Test > implicit baseline), there were no significant

interactions between sleep duration and TNF-α in left MPFC for those contrasts.

Moreover, there were no significant interactions between sleep duration and TNF-α for

the Test > implicit baseline and Practice > implicit baseline contrasts at the whole-brain

level.

Controlling for sex, BMI, test accuracy, sleep duration, and levels of TNF-α,

activation in left MPFC was not associated with stress reactivity or test-related anxiety.
Figure 4. Sleep duration moderated the associations between TNF-α and activation in

left MPFC for Test > Practice, cluster-corrected at Z > 3.1, p < .05. Among individuals

who reported short sleep duration (1 SD below mean = 384.845 minutes), greater levels

of TNF-α were associated with greater deactivation in left MPFC during Test > Practice.

Additionally, among those who reported long sleep duration (1 SD above mean =

529.651), greater levels of TNF-α were associated with less MPFC deactivation during

Test > Practice. *p < .05

Whole-brain analyses revealed an interaction between sleep duration and IL-6 in

left occipital pole (-16, -104, -10, Z = 4.49, cluster size = 120 voxels) for Test > Practice

contrast. Examining each condition relative to baseline separately, whole-brain analyses

revealed a significant interaction between sleep duration and IL-6 during Practice

(relative to implicit baseline) in left frontal pole (-48, 44, -8, Z = 4.57, cluster size = 199
voxels), bilateral occipital cortex (left: -16, -80, -8, Z = 3.99, cluster size = 121 voxels;

right: 18, -88, 8, Z = 4.41, cluster size = 137 voxels), right temporal cortex (50, -52, -24,

Z = 3.68, cluster size = 102 voxels), and right ACC (6, -4, 34, Z = 3.87, cluster size =

102 voxels). Follow-up simple slopes analyses to delineate the nature of these

interactions are reported in Supplemental Materials. There were no significant

interactions between sleep duration and IL-6 during Test condition relative to implicit

baseline.

4. Discussion

The current study investigated whether and how individual differences in

peripheral immune markers (namely, TNF-α and IL-6) related to neural response to

stress in adolescents and whether these immune-brain associations were moderated by

adolescents’ sleep habits (sleep duration). Based on previous research demonstrating

associations between peripheral inflammation and corticolimbic function, and the

interactive effects of sleep and stress on inflammatory processes, we hypothesized that

adolescent who have higher levels of inflammatory markers in the periphery would

evince brain activation patterns that reflected heightened neural response to stress

(e.g., greater dACC, anterior insula, amygdala activation or deactivation during

stress/test condition relative to non-stress/practice condition), and that this effect would

be stronger among adolescents with shorter sleep duration. We conducted a priori ROI

analyses using dACC, bilateral anterior insula, bilateral amygdala, and bilateral

hippocampus as ROIs, followed by exploratory whole-brain analyses.

After correcting for multiple comparisons, there were no significant associations

between TNF-α, IL-6, or their interactions with sleep duration on neural response to
stress. However, uncorrected for multiple comparisons, we found that sleep duration

moderated the association between TNF-α and neural response to stress in the left

amygdala. Follow-up tests probing the nature of the interaction showed that, among

adolescents who reported short sleep duration, greater levels of TNF-α were associated

with greater deactivation in the left amygdala during stress. This effect was attenuated

among adolescents who reported average to long sleep duration. A similar pattern of

results was observed in whole-brain analyses for the sleep duration x TNF-α interaction,

but in the left MPFC such that, among adolescents who reported short sleep duration,

greater levels of TNF-α were associated with greater deactivation in left MPFC during

stress. In contrast, among adolescents who reported long sleep duration, greater levels

of TNF-α were associated with less deactivation in left MPFC during stress.

These findings are consistent with previous research demonstrating associations

between peripheral inflammation and amygdala function and connectivity in adolescents

(Miller et al., 2020; Nusslock et al., 2019; Swartz et al., 2021). Deactivation in limbic

regions (e.g., amygdala, hippocampus, ventral striatum) has been commonly observed

in previous studies that have utilized the MIST in adults and adolescents (Berretz et al.,

2021; Corr et al., 2021; Pruessner et al., 2008). Furthermore, deactivation in limbic

regions (e.g., hippocampus) has been associated with HPA activity (Corr et al., 2021;

Pruessner et al., 2008). The current study extends from previous research to

demonstrate that deactivation in corticolimbic regions during stress (i.e., on the MIST)

may also be associated with peripheral inflammation in adolescents. These findings

suggest that short/insufficient sleep might exacerbate the associations between

inflammation and heightened activation (or deactivation) in emotion/stress-related

regions during stress, which has implications for increased anxiety and potentiated

stress reactivity. Indeed, while we did not find associations between amygdala or MPFC

deactivation and anxiety in the immune subsample, in the full larger sample, we found

that greater deactivation in the amygdala during stress was associated with greater

test/stress-related anxiety.

We did not observe any significant associations between IL-6 or its interaction

with sleep duration on adolescents’ neural response to stress in corticolimbic regions.

While this differs from some studies in adults (e.g., Eisenberger et al., 2009) and

adolescents (e.g., Nusslock et al., 2019; Miller et al., 2020) that observed links between

IL-6 and brain function, it is consistent with another study in adolescents that also did

not find significant associations between IL-6 and brain function, but found effects

between TNF-α and brain function (Swarz et al.,2021). As Swarz et al. (2021)

suggested, one reason why we may not have observed any effects with IL-6 could be

because we covaried for BMI in our analysis, which may confound associations

between inflammation and brain function, but was not done in Nusslock et al. (2019) or

Miller et al. (2020). Indeed, IL-6 and BMI were highly correlated in our sample, which

may explain why we did not see significant effects between IL-6 and brain function after

covarying for BMI. In addition to methodological differences, another reason could be

due to the relatively low levels of IL-6 in our sample: a majority of adolescents in our

sample had relatively low levels of IL-6, causing a positive skew, whereas the

distribution of the TNF-α levels (before log-transformation) was relatively normal,

indicative of greater variability. Moreover, adolescents tend to have relatively intact

immune systems that keep inflammatory activity from fostering a chronic inflammatory
state, therefore having relatively low levels of inflammation (Miller & Chen, 2010). As a

result, our sample may have had restricted range for discovery of substantial

mind/brain-body associations.

The current study has several limitations to note. First, the correlational design of

the study precludes drawing any conclusions about the directionality of the relations

among the sleep measures, immune markers, and brain function. Second, while we

found significant effects for the interactions between sleep duration and TNF-α on brain

function, the sample size for those analyses was very small, so it remains to be tested

whether these effects would replicate in larger samples. Third, sleep duration was

determined via self-report from the adolescents and also at the end of the day rather

than when they wake up that day, which may not be as accurate as more objective

measures of daily sleep such as actigraphy. Additionally, sleep duration was measured

for only one week and it is unknown whether this one week captured an average week

in the adolescents’ lives. It could be possible that sleep behavior during the measured

week may not represent some adolescents’ average week (e.g., adolescents could be

on break from school, traveling, having a particularly challenging week, etc.).

Unfortunately, typicality of the week was not assessed. Fourth, the age-range of our

adolescents was restricted to 14-15 years of age, which precludes generalization of our

findings to younger or older adolescents. There also may have been self-selection bias

of subjects, as immune data were only available from those who opted in for the blood

draw. Future studies using longitudinal assessments of sleep (subjectively and

objectively measured), multiple immune markers, and a variety of brain measures

across adolescence in a diverse sample of adolescents (including those experiencing

adversity or clinical disorders) would be well-positioned to circumvent the limitations of

the current study.

Despite limitations, the current study makes novel contributions to the literature

on sleep, peripheral immune markers, and brain function in adolescents. It provides

preliminary evidence that even in a relatively healthy sample of adolescents, variability

in sleep patterns and immune markers relate to variability in neural response to stress in

corticolimbic circuitry in a sample of older adolescents, suggesting that sleep is an

important factor to consider when studying neuroimmune processes during

adolescence.

Acknowledgments

This research was supported by the National Science Foundation (BSC 1551952

to AJF, NIE, AG), the National Institute of Child Health and Human Development

(1R01HD093823-01 to AJF, NIE, and AG), an American Psychological Association

Dissertation Research Award to JPU, and a UC Consortium of the Science of

Adolescence Seed Grant to JPU.

 References

Anderson, C., & Platten, C. R. (2011). Sleep deprivation lowers inhibition and enhances

impulsivity to negative stimuli. Behavioural Brain Research, 217(2), 463–466.

https://doi.org/10.1016/j.bbr.2010.09.020

Baker, A. E., Tashjian, S. M., Goldenberg, D., & Galván, A. (2020). Neural activity

moderates the association between sleep and risky driving behaviors in

adolescence. Developmental Cognitive Neuroscience, 43(November 2019), 0–8.

https://doi.org/10.1016/j.dcn.2020.100790

Berretz, G., Packheiser, J., Kumsta, R., Wolf, O. T., & Ocklenburg, S. (2021). The brain

under stress—A systematic review and activation likelihood estimation meta-

analysis of changes in BOLD signal associated with acute stress exposure.

Neuroscience and Biobehavioral Reviews, 124(September 2020), 89–99.

https://doi.org/10.1016/j.neubiorev.2021.01.001

Carskadon, M. A., Vieira, C., & Acebo, C. (1993). Association between puberty and

delayed phase preference. Sleep, 16(3), 258–262.

https://doi.org/10.1093/sleep/16.3.258

Carskadon, M. A., Wolfson, A. R., Acebo, C., Tzischinsky, O., & Seifer, R. (1998).

Adolescent sleep patterns, circadian timing, and sleepiness at a transition to early

school days. Sleep, 21(8), 871–881.

CDC. (2011). Unhealthy sleep-related behaviors--12 States, 2009. MMWR. Morbidity

and Mortality Weekly Report, 60(8), 233–238.

Chiang, J. J., Cole, S. W., Bower, J. E., Irwin, M. R., Taylor, S. E., Arevalo, J., & Fuligni,

A. J. (2019). Daily interpersonal stress , sleep duration , and gene regulation during
 late adolescence. Psychoneuroendocrinology, 103, 147–155.

https://doi.org/10.1016/j.psyneuen.2018.11.026

Corr, R., Pelletier-Baldelli, A., Glier, S., Bizzell, J., Campbell, A., & Belger, A. (2021).

Neural mechanisms of acute stress and trait anxiety in adolescents. NeuroImage:

Clinical, 29, 102543. https://doi.org/10.1016/j.nicl.2020.102543

Dedovic, K., Renwick, R., Mahani, N. K., Engert, V., Lupien, S. J., & Pruessner, J. C.

(2005). The Montreal Imaging Stress Task: Using functional imaging to investigate

the effects of perceiving and processing psychosocial stress in the human brain.

Journal of Psychiatry and Neuroscience, 30(5), 319–325.

Dedovic, K., Slavich, G. M., Muscatell, K. A., Irwin, M. R., & Eisenberger, N. I. (2016).

Dorsal anterior cingulate cortex responses to repeated social evaluative feedback

in young women with and without a history of depression. Frontiers in Behavioral

Neuroscience, 10(MAR), 1–13. https://doi.org/10.3389/fnbeh.2016.00064

Eisenberger, N. I., Inagaki, T. K., Mashal, N. M., & Irwin, M. R. (2010). Inflammation and

social experience: An inflammatory challenge induces feelings of social

disconnection in addition to depressed mood. Brain, Behavior, and Immunity, 24(4),

558–563. https://doi.org/10.1016/j.bbi.2009.12.009

Eisenberger, N. I., Inagaki, T. K., Rameson, L. T., Mashal, N. M., & Irwin, M. R. (2009).

An fMRI study of cytokine-induced depressed mood and social pain: The role of

sex differences. NeuroImage, 47(3), 881–890.

https://doi.org/10.1016/j.neuroimage.2009.04.040

Eisenberger, N. I., Moieni, M., Inagaki, T. K., Muscatell, K. A., & Irwin, M. R. (2017). In

Sickness and in health: The co-regulation of inflammation and social behavior.

 Neuropsychopharmacology, 42(1), 242–253. https://doi.org/10.1038/npp.2016.141

Harrison, N. A., Brydon, L., Walker, C., Gray, M. A., Steptoe, A., & Critchley, H. D.

(2009). Inflammation Causes Mood Changes Through Alterations in Subgenual

Cingulate Activity and Mesolimbic Connectivity. Biological Psychiatry, 66(5), 407–

414. https://doi.org/10.1016/j.biopsych.2009.03.015

Inagaki, T. K., Bryne Haltom, K. E., Suzuki, S., Jevtic, I., Hornstein, E., Bower, J. E., &

Eisenberger, N. I. (2016). The neurobiology of giving versus receiving support: The

role of stress-related and social reward-related neural activity. Psychosomatic

Medicine, 78(4), 443–453. https://doi.org/10.1016/j.joca.2015.05.020.Osteoarthritic

Inagaki, T. K., Muscatell, K. A., Irwin, M. R., Cole, S. W., & Eisenberger, N. I. (2012).

Inflammation Selectively Enhances Amygdala Activity to Socially Threatening

Images. NeuroImage, 59(4), 3222–3226.

https://doi.org/10.1016/j.neuroimage.2011.10.090.Inflammation

Irwin, M. R. (2015). Why Sleep Is Important for Health: A Psychoneuroimmunology

Perspective. Annual Review of Psychology, 66(1), 143–172.

https://doi.org/10.1146/annurev-psych-010213-115205

Irwin, M. R. (2019). Sleep and inflammation : partners in sickness and in health. Nature

Reviews Immunology, 19(November), 702–715. https://doi.org/10.1038/s41577-

019-0190-z

Irwin, M. R., Olmstead, R., & Carroll, J. E. (2016). Sleep disturbance, sleep duration,

and inflammation: A systematic review and meta-analysis of cohort studies and

experimental sleep deprivation. Biological Psychiatry, 80(1), 40–52.

https://doi.org/10.1016/j.biopsych.2015.05.014
Jenni, O. G., Achermann, P., & Carskadon, M. A. (2005). Homeostatic Sleep Regulation

in Adolescents. Sleep, 28(11), 1446–1454.

Kann, L., Kinchen, S., Shanklin, S. L., Flint, K. H., Kawkins, J., Harris, W. A., Lowry, R.,

Olsen, E. O., McManus, T., Chyen, D., Whittle, L., Taylor, E., Demissie, Z., Brener,

N., Thornton, J., Moore, J., & Zaza, S. (2014). Youth risk behavior surveillance--

United States, 2013. Morbidity and Mortality Weekly Report. Surveillance

Summaries (Washington, D.C. : 2002), 63 Suppl 4(4), 1–168.

https://doi.org/ss6304a1 [pii]

Killgore, W. D. S., Kamimori, G. H., & Balkin, T. J. (2011). Caffeine protects against

increased risk-taking propensity during severe sleep deprivation. Journal of Sleep

Research, 20(January 2016), 395–403. https://doi.org/10.1111/j.1365-

2869.2010.00893.x

Kraynak, T. E., Marsland, A. L., Wager, T. D., & Gianaros, P. J. (2018). Functional

neuroanatomy of peripheral inflammatory physiology: A meta-analysis of human

neuroimaging studies. Neuroscience and Biobehavioral Reviews, 94(July), 76–92.

https://doi.org/10.1016/j.neubiorev.2018.07.013

Lucas-Thompson, R. G., Crain, T. L., & Brossoit, R. M. (2021). Measuring sleep

duration in adolescence: Comparing subjective and objective daily methods. Sleep

Health, 7(1), 79–82. https://doi.org/10.1016/j.sleh.2020.06.005

Marsland, A. L., Gianaros, P. J., Abramowitch, S. M., Manuck, S. B., & Hariri, A. R.

(2008). Interleukin-6 Covaries Inversely with Hippocampal Grey Matter Volume in

Middle-Aged Adults. Biological Psychiatry, 64(6), 484–490.

https://doi.org/10.1016/j.biopsych.2008.04.016
Miller, G. E., & Chen, E. (2010). Harsh family climate in early life presages the

emergence of a proinflammatory phenotype in adolescence. Psychological

Science, 21(6), 848–856. https://doi.org/10.1177/0956797610370161

Miller, G. E., White, S. F., Chen, E., & Nusslock, R. (2020). Association of Inflammatory

Activity With Larger Neural Responses to Threat and Reward Among Children

Living in Poverty. American Journal of Psychiatry, 11, appi.ajp.2020.2.

https://doi.org/10.1176/appi.ajp.2020.20050635

Minkel, J., Moreta, M., Muto, J., Htaik, O., Jones, C., Basner, M., & Dinges, D. (2014).

Sleep deprivation potentiates HPA axis stress reactivity in healthy adults. Health

Psychology, 33(11), 1430–1434. https://doi.org/10.1037/a0034219

Moieni, M., Irwin, M. R., Jevtic, I., Olmstead, R., Breen, E. C., & Eisenberger, N. I.

(2015). Sex Differences in Depressive and Socioemotional Responses to an

Inflammatory Challenge: Implications for Sex Differences in Depression.

Neuropsychopharmacology : Official Publication of the American College of

Neuropsychopharmacology, 40(7), 1–8. https://doi.org/10.1038/npp.2015.17

Muscatell, K. A., Moieni, M., Inagaki, T. K., Dutcher, J. M., Jevtic, I., Breen, E. C., Irwin,

M. R., & Eisenberger, N. I. (2016). Exposure to an Inflammatory Challenge

Enhances Neural Sensitivity to Negative and Positive Social Feedback. Brain,

Behavior, and Immunity, 57, 21–29. https://doi.org/10.1016/j.bbi.2016.03.022

Nusslock, R., Brody, G. H., Armstrong, C. C., Carroll, A. L., Sweet, L. H., Yu, T., Barton,

A. W., Hallowell, E. S., Chen, E., Higgins, J. P., Parrish, T. B., Wang, L., & Miller,

G. E. (2019). Higher Peripheral Inflammatory Signaling Associated With Lower

Resting-State Functional Brain Connectivity in Emotion Regulation and Central

 Executive Networks. Biological Psychiatry, 1–10.

https://doi.org/10.1016/j.biopsych.2019.03.968

Park, H., Chiang, J. J., Bower, J. E., Irwin, M. R., Almeida, D. M., Seeman, T. E.,

McCreath, H., & Fuligni, A. J. (2020). Sleep and Inflammation During Adolescents’

Transition to Young Adulthood. Journal of Adolescent Health, 67(6), 821–828.

https://doi.org/10.1016/j.jadohealth.2020.04.015

Park, H., Tsai, K. M., Dahl, R. E., Irwin, M. R., McCreath, H., Seeman, T. E., & Fuligni,

A. J. (2016). Sleep and inflammation during adolescence. Psychosomatic Medicine,

78(6), 677–685. https://doi.org/10.1097/PSY.0000000000000340

Pruessner, J. C., Dedovic, K., Khalili-Mahani, N., Engert, V., Pruessner, M., Buss, C.,

Renwick, R., Dagher, A., Meaney, M. J., & Lupien, S. (2008). Deactivation of the

Limbic System During Acute Psychosocial Stress: Evidence from Positron

Emission Tomography and Functional Magnetic Resonance Imaging Studies.

Biological Psychiatry, 63(2), 234–240.

https://doi.org/10.1016/j.biopsych.2007.04.041

Reichenberg, A., Yirmiya, R., Schuld, A., Kraus, T., Haack, M., Morag, A., &

Pollmächer, T. (2001). Cytokine-Associated Emotional and Cognitive Disturbances

in Humans. Archives of General Psychiatry, 58(5), 445.

https://doi.org/10.1001/archpsyc.58.5.445

Slavich, G. M., Way, B. M., Eisenberger, N. I., & Taylor, S. E. (2010). Neural sensitivity

to social rejection is associated with inflammatory responses to social stress.

Proceedings of the National Academy of Sciences, 107(33), 14817–14822.

https://doi.org/10.1073/pnas.1009164107
Swartz, J. R., Carranza, A. F., Tully, L. M., Knodt, A. R., Jiang, J., Irwin, M. R., &

Hostinar, C. E. (2021). Associations between peripheral inflammation and resting

state functional connectivity in adolescents. Brain, Behavior, and Immunity,

95(June 2020), 96–105. https://doi.org/10.1016/j.bbi.2021.02.018

Telzer, E. H., Fuligni, A. J., Lieberman, M. D., & Galván, A. (2013). The effects of poor

quality sleep on brain function and risk taking in adolescence. NeuroImage, 71,

275–283. https://doi.org/10.1016/j.neuroimage.2013.01.025

Vgontzas, A. N., Zoumakis, E., Bixler, E. O., Lin, H.-M. M., Follett, H., Kales, A., &

Chrousos, G. P. (2004). Adverse Effects of Modest Sleep Restriction on

Sleepiness, Performance, and Inflammatory Cytokines. Journal of Clinical

Endocrinology and Metabolism, 89(5), 2119–2126. https://doi.org/10.1210/jc.2003-

031562

Walker, M. P., & Stickgold, R. (2006). Sleep, memory, and plasticity. Annual Review of

Psychology, 57, 139–166.

https://doi.org/10.1146/annurev.psych.56.091103.070307