Content uploaded by Eti Ben-Simon
Author content
All content in this area was uploaded by Eti Ben-Simon on Jun 19, 2023
Content may be subject to copyright.
Available via license: CC0
Content may be subject to copyright.
Sleep Loss Influences the Interconnected Brain-Body
Regulation of Cardiovascular Function in Humans
Adam J. Krause, PhD, Raphael Vallat, PhD, Eti Ben Simon, PhD, and Matthew P. Walker, PhD
ABSTRACT
Objective: Poor sleep is associated with hypertension, a major risk factor for cardiovascular disease. However, the mechanism(s) through
which sleep loss affects cardiovascular health remains largely unknown, including the brain and body systems that regulate vascular function.
Methods: Sixty-six healthy adults participated in a repeated-measures, crossover, experimental study involving assessmentsof cardiovas-
cular function and brain connectivity after a night of sleep and a night of sleep deprivation.
Results: First, sleep deprivation significantly increased blood pressure—both systolic and diastolic. Interestingly, this change was inde-
pendent of any increase in heart rate, inferring a vasculature-specific rather than direct cardiac pathway. Second, sleep loss compromised
functional brain connectivity within the vascular control network, specifically the insula, anterior cingulate, amygdala, and ventral and me-
dial prefrontal cortices. Third, sleep loss–related changes in brain connectivity and vascular tone were not independent, but significantly
interdependent, with changes within the vascular control brain network predicting the sleep-loss shift toward hypertension.
Conclusions: These findings establish an embodied framework in which sleep loss confers increased risk of cardiovascular disease
through an impact upon central brain control of vascular tone, rather than a direct impact on accelerated heart rate itself.
Key words: sleep, sleep deprivation, functional magnetic resonance imaging, connectivity, cardiovascular, blood pressure.
INTRODUCTION
Poor and insufficient sleep predicts hypertension—a risk factor
for cardiovascular disease and mortality (1,2). Epidemiologi-
cal studies indicate that sleep disturbances are associated with a
greater incidence of hypertension, even when controlling for other
risk factors (2,3). Longitudinal studies have demonstrated that short
or low-quality sleep forecast a higher risk of hypertension (2,4). Ma-
nipulations using partial or total sleep deprivation both causally in-
crease blood pressure (5,6). Conversely, the extension of sleep in
prehypertensive short sleepers reduces blood pressure (7). Together,
such evidence supports the proposed causal link between coinciding
increases in short sleep and increasing rates of cardiovascular dis-
ease in numerous first-world nations (8).
However, the mechanistic pathways through which insufficient
sleep increases blood pressure and thus, cardiovascular disease
risk remain unclear. This is especially true regarding a possible in-
teraction between aberrant central brain and peripheral body sys-
tems that are known to control blood pressure. Indeed, the brain
is recognized to play a causal role in the regulation of peripheral
physiology, including that of blood pressure and heart rate (9).
Core to this control is a discrete set of viscerosensory brain regions
forming a network, including the amygdala, anterior cingulate,
insula, and ventromedial prefrontal cortex (vmPFC) and medial
prefrontal cortex (MPFC), that map ascending and conversely in-
stigate descending changes in vascular regulation.
We sought to test the hypothesis that one pathway through
which insufficient sleep instigates increases in blood pressure is
through an altered relationship between brain vascular control net-
works and the peripheral cardiovascular system. Specifically, we
tested the prediction that elevations in blood pressure caused by
a night of sleep deprivation are associated with changes in the
functional connectivity state of the a priori brain network.
Sixty-six normotensive adult participants were enrolled in a
counterbalanced, repeated-measures design involving two condi-
tions: one night of sleep and one night of sleep loss. In each con-
dition, participants had their cardiovascular state, including blood
pressure and heart rate, assessed at circadian-matched morning
timepoints, and received a resting-state functional magnetic reso-
nance imaging (fMRI) scan after sleep or sleep deprivation.
METHODS
Experimental Model and Participant Details
Sixty-six healthy adults aged 18 to 24 years (mean [standard deviation] =
20.7 [1.7], 52% female) completed a repeated-measures crossover design
(described hereinafter). There was no influence of sex on cardiovascular
outcomes (p> .64). Participants abstained from caffeine and alcohol for
Supplemental Digital Content
From the Center for Human Sleep Science, Department of Psychology (Krause, Vallat, Simon, Walker), and Helen Wills Neuroscience Institute (Walker),
University of California, Berkeley, California.
Address correspondence to Matthew P. Walker, PhD, University of California, Berkeley, Berkeley, CA 94720-1650. E-mail: mpwalker@berkeley.edu
ORCID IDs: 0000-0001-5529-1296 (A.J.K.), 0000-0003-1779-7653 (R.V.), 0000-0002-3862-8288 (E.B.S.), 0000-0002-7839-6389 (M.P.W.).
Received for publication January 31, 2022; revision received July 26, 2022.
DOI: 10.1097/PSY.0000000000001150
Copyright © 2022 by the American Psychosomatic Society
fMRI = functional magnetic resonance imaging, MPFC = medial
prefrontal cortex, ROI = region of interest, vmPFC = ventromedial
prefrontal cortex
ORIGINAL ARTICLE
Psychosomatic Medicine, V 85 •34-41 34 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
72 hours before and during the entire course of the study and kept a normal
sleep-wake rhythm (7–9 hours of sleep per night) with sleep onset before
1:00 AM andrisenolaterthan9:00AM for the three nights before the study
participation, as verified by sleep logs and actigraphy.
Exclusion criteria, assessed using a prescreening semistructured inter-
view, were as follows: a history of previously diagnosed sleep disorders,
neurological disorders, closed head injury, Axis 1 psychiatric disorders
according to the Diagnostic and Statistical Manual of Mental Disorders
(Fourth Edition, Text Revision) criteria (encompassing mental disorders,
including depression, anxiety disorders, bipolar disorder, attention deficit
disorder, and schizophrenia), history of drug abuse and current use of anti-
depressant or hypnotic medication, nicotine use, consumption of more than
five alcoholic drinks per week, crossing of time zones in the 3 months be-
fore the study, and general contraindications to MRI. Participants also had
blood pressures within normotensive ranges, confirmed with a blood pressure
reading of less than 120/80 mm Hg at the time of consenting (10). Participants
who reported sleeping <7 hours per night or consuming three or more daily
caffeine-containing drinks were also excluded from entering the study. The Pitts-
burgh Sleep Quality Index (11) was used to determine the quality of recent sleep
history, and participants considered poor sleepers (global score >5) were ex-
cluded. The study was approved by the UC Berkeley Committee for the Pro-
tection of Human Subjects, with all participants providing written consent.
Experimental Design and Statistical Analyses
To test the experimental hypotheses, participants entered a repeated-measures
study design, including two sessions—one after a normal night of sleep and
one after 24 hours of total sleep deprivation. The two sessions were separated
by at least 7 days (mean [standard deviation] = 9.8 [3.8] days), with the order of
the sleep-rested and deprived conditions counterbalanced across participants.
In the sleep-deprived session, participants arrived at the laboratory at
10:00 PM and were continuously monitored throughout the enforced wak-
ing period. Activities during the sleep deprivation condition were limited
to use of the Internet, email, short walks, reading, movies of low emotion-
ality, and playing board games, thereby providing a standardized regimen
of waking activity without undue stress. Participants also avoided exercise
during the experimental sessions. Although participants in the sleep-rested
session were supine for a longer period than in the sleep-deprived session,
the limiting of physical activity meant the sleep-deprived session was con-
ducted with the participant in a relaxed sitting position for much of the ex-
perimental night. Furthermore, cardiovascular assessments wereperformed
with an identical upright sitting posture for both experimental conditions.
In the sleep-rested session, participants came to the laboratory at 8:00 PM
and were prepared for an 8-hour night of sleep (~11:00 PM to 7:30 AM ±
30 minutes). The following morning, participants’cardiovascular state
was measured followed by fMRI scanning. During resting-state fMRI
scanning, participants were instructed to keep eyes open and fixated, with-
out thinking of anything in particular.
Cardiovascular Assessments
In each condition, cardiovascular state was assessed to determine blood pres-
sure, heart rate, and heart rate variability. These assessments were performed
approximately 1 hour after awakening, and at circadian-matched times in the
sleep rested and sleep deprivation conditions for each participant. Participants
did not eat or drink before the cardiovascular assessments.
Blood pressure measurement was performed with Omron BP742 blood
pressure monitor (Omron Healthcare Europe BV, Hoofddorp, the Netherlands),
fitted to the participants’nondominant arm over the brachial artery. After a rest
period of at least 10 minutes, participants were instructed to remain motionless
with open eyes while sitting with feet flat on the floor during the recording with
rested cuffed arm at heart height, consistent with standard blood pressure
recording guidelines (12). Two consecutive measurements were taken,
and the mean of these two measurements was used for analyses.
Immediately after blood pressure assessment, participants were fitted
with electrodes for bipolar electrocardiography, and cardiac activity was
recorded for 5 minutes. Electrodes were applied in a lead II format with the
reference electrode placed below the right clavicle parallel to the right shoul-
der and a second electrode placed on the torso at the fourth intercostal space
on the left side parallel to the left hip. Similar to the blood pressure recording,
participants sat motionless with open eyes while sitting with feet flat on the
floor. All data were screened for measurement artifacts and beat-to-beat inter-
vals extracted using ARTiiFACT software (13), and all flagged interbeat in-
tervals were visually checked. If confirmed as artifactual, they were deleted
and substituted by means of cubic spline interpolation of neighboring inter-
vals. The time-domain measure pNN50 was calculated for each participant,
which is the percentage of differences between adjacent interbeat intervals
(NN) that are greater than 50 milliseconds and represent a recommended
measure of vagally mediated heart rate variability (14).
fMRI Networks and Processing
A set of cortical and subcortical brain regions was selected to define the
central cardiovascular control network, based on three convergent criteria.
First, regions were selected from previous reports associating cardiovascu-
lar changes with fMRI patterns. Second, regions were selected if they had
anatomical connections allowing for top-down cardiovascular modulation
or bottom-up visceral representation. Last, regions known to be involved
in the mapping and modulation of autonomic outflow were considered. From
these criteria, a collection of bilateral regions formed the a priori cardiovascu-
lar network used in the current analyses, including the insula, MPFC, anterior
cingulate cortex, vmPFC, and amygdala.
Of this collection of regions, the insular cortex and MPFC are of special im-
portance in the regulation of vascular state. Both regions are involved in the cen-
tral brain regulation of autonomic state, including the mapping of current auto-
nomic balance and modulation autonomic outflow (9,15,16). Activity in the
insula and MPFC have further been linked with increased vasoconstriction in
individuals with a history of coronary artery disease (17), and ischemic injury
to the insula induces endothelial dysfunction and inflammation (18). Similarly,
lesioning of the MPFC increases plasma levels of adrenocorticotropic hormone
and corticosterone (19), both important stress-response factors that can affect
cardiovascular function (20). The insula and MPFC are unique in the regulation
of cardiovascular function through direct connections to brainstem areas
(21,22). Finally, both regions can further affect vascular function through their
influence on immune function because vascular endothelial cells are sensitive to
inflammatory processes (23,24).
Preprocessing was performed using fMRIPrep 1.2.5 (RRID:SCR_016216).
The T1-weighted image was corrected for intensity nonuniformity using
N4BiasFieldCorrection and used as a reference throughout the analysis.
Spatial normalization to the ICBM 152 Nonlinear Asymmetrical template
was performed through nonlinear registration with antsRegistration using
brain-extracted version of both T1-weighted volume and template (ANTs
2.2.0 RRID:SCR_004757). Brain tissue segmentation of cerebrospinal
fluid, white matter, and gray matter was performed on the brain-extracted
T1-weighted image using fast (FSL 5.0.10, RRID:SCR_002823).
For fMRI data, the mean BOLD (blood-oxygen-level-dependent)
image was co-registered to the T1-weighted reference using flirt.
Head-motion parameters with respect to the BOLD reference (transforma-
tion matrices, and six corresponding rotation and translation parameters)
were estimated before any spatiotemporal filtering using mcflirt. BOLD
runs were slice-time corrected using 3dTshift from AFNI (RRID:
SCR_005927), and the BOLD time series (including slice-timing cor-
rection when applied) were resampled to MNI152NLin2009cAsym stan-
dard space. Known time-series factors requiring covariate accommodation
were calculated based on the preprocessed BOLD: framewise displacement,
DVARS (spatial standard deviation of successive difference images), and
three region-wise global signals. The three global signals are extracted within
the cerebrospinal fluid, the white matter, and the whole-brain masks. In
addition, a set of physiological regressors were extracted to allow for
component-based noise correction (CompCor (25)). Finally, regression
was used to estimate and remove the whole-brain global signal time series,
Sleep Loss/Cardiovascular Function Regulation
Psychosomatic Medicine, V 85 •34-41 35 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
and functional images were smoothed using a 6-mm full-width at half-maximum
Gaussian kernel.
Statistical and fMRI Analysis
Connectivity analyses were performed using the CONN toolbox version
19b (www.nitrc.org/projects/conn, RRID:SCR_009550). First, denoising
was performed by regressing the six motion parameters, framewise dis-
placement, and physiological noise components. The resulting BOLD time
series was band-pass filtered (0.008–0.09 Hz), followed by ROI analysis,
specifically ROI-to-ROI connectivity analysis.
Based on the experimental hypotheses, analyses focused on an a priori
set of cortical and subcortical ROIs known to be involved central brain car-
diovascular control (9,15,16,26–29). Although the brainstem is known to
be directly involved in blood pressure control, its location close to arteries
and fluid-filled spaces makes recording reliable BOLD signals difficult and is
not included in the analysis (30). These ROIs were drawn from anatomical
atlases (Harvard-Oxford Cortical and Subcortical Atlas) included in the CONN
toolbox and were extracted using an ICA (Independent Components Analysis)
approach. For each ROI, the connectivity to all other ROIs was calculated for
each participant in each condition separately, generating a symmetric correla-
tion matrix wherein each element is the correlation between two ROIs in each
condition and each participant.
In addition to ROI-to-ROI connectivity analyses, control analyses for
specificity examined connectivity within and between additional prototypical
brain networks of the default-mode network, dorsal attention network, ventral
salience network, and somatomotor network (31). Within-network connec-
tivity was calculated for each network by extracting the mean resting-state
BOLD time series from each voxel within a region of a network and calcu-
lating the mean Pearson correlation with all other regions of the network.
Between-network connectivity was similarly calculated for each network
by extracting the mean resting-state BOLD time series from every voxel
within a network, and then calculating the correlation with the mean time
series of all other networks.
The resulting correlation coefficients from these fMRI analyses across
participants in each condition were then statistically compared (two-tailed
paired ttest) to determine the effectof sleep loss on brain connectivity. Con-
sistent with imaging recommendations to correct for multiple statistical
comparisons among the eight ROIs and four networks, the Benjamini-Hochberg
procedure for controlling FDR was applied, with q= 0.05 (32). This procedure
addresses the problem of simultaneously performing multiple significance tests
by using an adaptive stepwise procedure to control the expected ratio of errone-
ous rejections to the number of total rejections. Only connectivity changes that
survived the FDR criterion after correction were considered statistically signif-
icant and used in analyses. Pearson correlations were used to test the relation-
ships between sleep loss–related changes in brain connectivity and sleep
loss–related changes in cardiovascular state. All statistical analyses
were performed in Python using the Pingouin package (33). Data and
code used in this study will be shared by application request from a
qualified investigator at an academic institute, subject to the negotiation
of a university and data use agreement.
RESULTS
Sleep Loss and Cardiovascular Function
We first tested the hypothesis that sleep deprivation significantly
alters key vascular, cardiac, and autonomic metrics linked with
cardiovascular disease (34): systolic and diastolic blood pressure,
heart rate, and heart rate variability.
Supportive of the hypothesis and prior findings (1–4), sleep depri-
vation significantly increased systolic blood pressure, relative to rested
conditions (Figure 1A; systolic: sleep-rested mean = 107.6 mm Hg,
sleep-deprived mean = 112.1 mm Hg, t(60) = −3.04, p= .003, 95%
confidence interval [CI] to −7.04 to −1.45). Similarly, diastolic blood
pressure also increased significantly after sleep loss (Figure 1A;
diastolic: sleep-rested mean = 71.1 mm Hg, sleep-deprived
mean = 73.7 mm Hg, t(61) = −2.24, p= .03, 95% CI = −5.04 to −0.28]).
Importantly, these changes in blood pressure occurred in the
absence of an increase in heart rate. Despite the increase in blood
pressure, there was an overall reduction of resting heart rate under
conditions of sleep loss (Figure 1B; sleep-rested mean = 67.5 beats/min,
sleep-deprived mean = 62.9 beats/min, t(63) = 3.73, p< .001, 95%
CI = 2.12 to 6.99]), a finding consistent with previous reports
(5,35–38) (cf. (6)). Accordingly, heart rate variability increased un-
der conditions of sleep deprivation (Figure 1B; pNN50: sleep-rested
mean = 27.5%, sleep-deprived mean = 42.0%, t(61) = −5.93,
p< .0001, 95% CI = −19.11 to −9.47). Because pNN50 reflects vagal
modulation of the heart (14), this finding is consistent with increased
parasympathetic input to the heart associated with sleep loss (35,37).
Further advancing the dissociation between increased periph-
eral blood pressure in the absence of accelerated heart rate, there
was no significant association between the changes in blood pressure
and heart rate between conditions ([sleep-rested] −[sleep depriva-
tion]) for either systolic or diastolic measures (systolic: r(60) = 0.07,
p= .61, 95% CI = −0.19 to 0.32]; diastolic: r(61) = 0.03, p= .84,
95% CI = −0.23 to 0.28). A similar lack of association was ob-
served between the sleep-loss changes in heart rate variability
and blood pressure (systolic: r(60) = −0.13, p= .32, 95% CI =
−0.38 to 0.13; diastolic: r(61) = −0.08, p= .55, 95% CI = −0.33
to 0.18). In contrast, there was a significant association between
the sleep loss–related reduction in heart rate and increased heart
rate variability (r(62) = −0.48, p< .0001, 95% CI = −0.65 to
−0.26). Heart rate variability is known to have a positive relation-
ship with the beat-to-beat period, and it has been argued that eval-
uation of experimental effects on heart rate variability should con-
trol for heart rate (39). Thus, we used an analysis of covariance to
test whether heart rate moderated the effect of condition on heart
rate variability. After controlling for heart rate, heart rate variabil-
ity remained significantly increased under conditions of sleep dep-
rivation (F(1,126) = 9.9, p= .002).
Sleep Loss and Resting Brain Connectivity
Next, we assessed functional connectivity within the a priori network
known to be associated with cardiovascular control in mammals, in-
cluding the amygdala, anterior cingulate, insula, and vmPFC, and
MPFC (9,15,16,26–28). Within this network, 12 connection paths
were significantly altered under conditions of sleep deprivation, rela-
tive to the sleep rested state (false discovery rate [FDR] correction for
multiple tests) (32). These included connectivity between the insula
and anterior cingulate, the insula and MPFC, the anterior cingulate
and the MPFC, the anterior cingulate and the amygdala, the
vmPFC and MPFC, and MPFC and amygdala (Supplemental Dig-
ital Content, Table S1, http://links.lww.com/PSYMED/A880).
Eleven of these region-of-interest (ROI) pairs demonstrated sig-
nificant increases in their connectivity (Supplemental Digital Con-
tent, Table S1, http://links.lww.com/PSYMED/A880), with the con-
nectivity changes all sharing at least one of two nodes: the insula
and MPFC, whereas connectivity between the vmPFC and MPFC
showed a converse decrease in connectivity after sleep deprivation
(t(65) = 3.2, p= .002, 95% CI = 0.03 to 0.15). In the former set,
the insula had significantly increased connectivity with the anterior
cingulate cortex (t(65) = −2.6, p= .012, 95% CI = −0.15 to −0.02)
and the MPFC (t(65) = −4.66, p< .001, 95% CI = −0.23 to −0.09).
ORIGINAL ARTICLE
Psychosomatic Medicine, V 85 •34-41 36 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
The MPFC further demonstrated increases in connectivity with both
the anterior cingulate in the right hemisphere (t(65) = −4.38,
p< .001, 95% CI = −0.17 to −0.06) and amygdala in both hemi-
spheres (right: t(65) = −2.98, p= .004, 95% CI = −0.13 to −0.03;
left: t(65) = −2.36, p= .021, 95% CI = −0.13 to −0.01). Connectivity
changes are described in full in Supplemental Digital Content,
http://links.lww.com/PSYMED/A880).
Having established changes in brain network connectivity caused
by sleep deprivation, we next tested whether these changes were
associated with the sleep loss–related increases in blood pressure,
as an embodied explanatory model would predict.
Consistent with the prediction, increases in insula connectivity
were significantly associated with a rise in systolic blood pressure,
especially between the insula and the MPFC (r(60) = 0.25,
p= .050, 95% CI = 0.01 to 0.47; Figure 2C). Second, increases
in MPFC connectivity with the amygdala were significantly and
negatively related to increases in diastolic blood pressure (r(61)
=−0.29, p= .024, 95% CI = −0.5 to −0.04; Figure 2D). Therefore,
sleep-loss changes in systolic and diastolic blood pressure were
predicted by alterations in brain connectivity, with a clustering
around changes in insular and MPFC connectivity, respectively.
No associations between changes in heart rate or heart rate
FIGURE 1. Cardiovascular outcomes. A, After sleep deprivation, there was an increase in both systolic and diastolic blood pressure, relative
to the sleep rested condition. B, Sleep deprivation resulted in a reduction in heart rate, relative to the sleep-rested condition. Sleep deprivation
resulted in an increase in heart rate variability (pNN50), relative to the sleep-rested condition. C–F, Scatterplots represent
nonsignificant correlations between sleep-loss changes (deprived −rested) in cardiovascular outcomes, including (C) heart rate
and systolic blood pressure, (D) heart rate and diastolic blood pressure, (E) heart rate variability (pNN50) and systolic blood pressure,
and (F) heart rate variability (pNN50) and diastolic blood pressure. Error bars indicate SEM. *p< .05. pNN50 = percentage of
differences between adjacent internet intervals exceeding 50ms; SEM = standard error of the mean. Color image is available only in
online version (www.psychosomaticmedicine.org).
Sleep Loss/Cardiovascular Function Regulation
Psychosomatic Medicine, V 85 •34-41 37 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
variability and connectivity were observed (all p> .05). Further-
more, the significant associations between increased blood pres-
sure and insula and MPFC connectivity remained significant when
controlling for either heart rate or heart rate variability (all p<.05).
To understand the specificity of brain and cardiovascular changes,
post hoc analyses examined the associations between sleep loss–
related alterations in four, non–a-priori, larger-scale resting-state brain
networks: the default-mode network, dorsal attention network, Ventral
salience network, and somatomotor network (31). Changes in con-
nectivity neither within nor between these large-scale networks
showed significant associations with the sleep loss–related changes
in blood pressure, heart rate, or heart rate variability (all p>.05).
Therefore, local changes expressly within the specific a priori cardio-
vascular control network (9,15,16,26–28), rather than any nonspe-
cific, global changes in larger-scale brain networks, best accounted
for changes in peripheral blood pressure. Finally, post hoc sensitivity
analyses were conducted to test whether any of the identified rela-
tionships were moderated by condition order. None of the associa-
tions reported previously were significantly moderated by the order
of the sleep rested and sleep-deprived conditions (all p>.22).
FIGURE 2. Brain connectivity and cardiovascular changes. Connections shown are significant after FDR correction (q= 0.05). A, Connections
that are significantly increased in the sleep deprived condition relative to the sleep-rested condition. B, Connections that are significantly
decreased in the sleep-deprived condition relative to the sleep rested condition. C, Scatterplot represents a significant positive correlation
between sleep loss–related increases in insula and MPFC connectivity and sleep loss–related increases in systolic blood pressure. D,
Scatterplot represents significant negative correlation between sleep-loss increases in MPFC and amygdala connectivity and sleep loss–
related increases in diastolic blood pressure. FDR = false discovery rate; MPFC = medial prefrontal cortex; ACC = anterior cingulate
cortex; vmPFC = ventromedial prefrontal cortex. Color image is available only in online version (www.psychosomaticmedicine.org).
ORIGINAL ARTICLE
Psychosomatic Medicine, V 85 •34-41 38 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
DISCUSSION
The current study establishes that a) experimental sleep loss selec-
tively increases systolic and diastolic blood pressure, independent
of heart rate; b) sleep loss compromises functional brain connec-
tivity within cardiovascular-regulating regions; and c) the changes
in connectivity and vascular tone are interdependent.
Although blood pressure and heart rate can change in parallel
in responseto varied demands, each isregulated through separable
mechanisms, including at the levels of the autonomic nervous, en-
docrine, immune, and endothelial systems. For example, sleep loss
increases inflammatory markers of endothelial dysfunction that are
linked with increases in peripheral blood pressure (40). Indeed, un-
coordinated or directionally opposite changes in cardiac and vas-
cular function occur under conditions of excess allostatic load, a
sign of impaired homeostasis (41,42).
Our findings comport well to a model of impaired, uncoordi-
nated whole-organism (brain and body) homeostasis caused by a
lack of sleep (42). Specifically, there was a significant increase in both
systolic and diastolic blood pressure, yet no concomitant acceleration
(rather, a deceleration) of heart rate contractility. Indeed, the sleep
loss–related changes in blood pressure and heart rate showed no evi-
dence of being significantly correlated with one another, further fitting
a state of allostatic distress and physiological discoordination (41,42).
Chronic sleep disturbances, including sleep disorders such as
insomnia and sleep apnea, are recognized risk factors for hyperten-
sion and cardiovascular disease (1–4). Moreover, experimental
sleep loss causally increases blood pressure (5,6), whereas length-
ening sleep in chronic short sleepers lowers blood pressure (7). To-
gether with the current findings, these results support a model in
which insufficient sleep increases blood pressure and the risk of
cardiovascular disease. However, some epidemiological studies
have reported U-shaped relationships between sleep duration and
hypertension (43–45). Because the current results do not provide
clarification regarding potentially increased risks of hypertension
associated with long sleep durations,future experiments will be re-
quired to address possible parabolic relationships between sleep
duration and hypertension that use experimental sleep extension
along with objective sleep measurements, to account for known
biases in self-reports (46,47).
The heart is known to receive both sympathetic and parasym-
pathetic input, yet corporeally, most vessels (arteries and veins)
only receive sympathetic innervation (48). Based on the dissocia-
tion observed in the current study, one parsimonious mechanism
capable of accounting for the dissociation between blood pressure
and heart rate is that increases in blood pressure caused by sleep
deprivation are triggered by an increase in sympathetic tone, whereas
the decrease in heart rate reflects a separable shift in cardiac auto-
nomic balance toward parasympathetic dominance (49). That is,
an autonomic dissociation wherein increased sympathetic output
dominates at the level of peripheral vascular control, yet an opposite,
parasympathetic influence dominates the heart. This possibility
is supported by sleep-loss increases in heart rate variability—a
measure reflecting vagal (parasympathetic) modulation of the heart
(14). Indeed, consistent with the current findings, increased sleep
debt resulting from insufficient sleep increases parasympathetic
drive and, as a consequence, slows heart rate (35,37). Notably, the
currently reported increase in heart rate variability replicates previ-
ous reports of the effects of sleep loss on next-day heart rate variabil-
ity (37,50,51). However, others have either reported no impact or
that sleep deprivation caused a decrease in heart rate variability
(52–54). There are several possible reasons that may account for
such inconsistent findings, including varied sleep manipulations
(chronic versus acute sleep deprivation). The cardiac autonomic ef-
fects of chronic sleep loss are more consistently reported than acute
sleep deprivation (55). In addition, there have been reports of sex
differences in the effects of sleep loss on heart rate variability (55),
but the current results showed no differences between male and fe-
male individuals for any cardiovascular measure. Future studies will
be required to address these inconsistencies, including additional
control of circadian timing (morning versus afternoon) and method
(active versus resting) of the cardiovascular assessments.
Combining these prior reports with findings from the current
study, one speculative candidate explaining the dissociable impact
of sleep loss on vascular and cardiac function involves three inde-
pendent mechanistic pathways, with each, or their combination, ac-
counting for our findings. The first is a reduction in sympathetic
signalling to the heart, combined with an increase parasympathetic
drive. Thisis evidenced by the increase in heart rate variability, re-
sulting in a lowering of contraction rate. The second is endothelial
dysfunction resulting from secretion of inflammatory mediators
and/or increased sympathetic input to the endothelium, thereby in-
creasing blood pressure. The third and nonmutually exclusive
pathway involves changes in the activation of adrenergic receptors
in the vasculature. Because α-adrenergic receptors are largely
expressed in vascular smooth muscle, increased sympathetic acti-
vation due to sleep loss can elicit vasoconstriction (48). Neverthe-
less, these possibilities do not discount nonautonomic mechanisms
that may further contribute or modulate peripheral systems, in-
cluding central brain regulation.
The aforementioned cardiovascular changes observed in the
current study were accompanied by significant functional connec-
tivity changes within the brain. Specifically, the shift toward hy-
pertension caused by sleep loss was matched by hyperconnectivity
between nodes of the vascular control network, all pivoting around
two hub nodes: the insula and the MPFC. The insula and MPFC
are sensitive to sleep status (56–61) and are key in the representa-
tion and control of cardiovascular states (9,15,16,28,62,63). Con-
sistent with such a privileged locus of control, both the MPFC
and insula are interconnected with cortical, subcortical, and
brainstem regions that dictate autonomic activity (9,15,16,28).
In contrast to the dominant profile of increased connectivity, one
pathway showed a decrease in connectivity: the link between the
MPFC and vmPFC, with both regions associated with sympathetic
nervous system control (29). Furthermore, reduced activation in the
MPFC has been associated with stress-induced vasoconstriction
(17), and experiments in rodent models have further established that
MPFC lesions alter baroreflex function (64). Our findings therefore
support a model in which bidirectional sleep loss changes in nodal
connectivity result in downstream impairments in the coordination
and thus peripheral control of cardiovascular physiology.
In addition to establishing changes in blood pressure and brain
network connectivity, our final analyses established that changes
in the brain and body were not simply co-occurring, but signifi-
cantly interrelated. First, the increase in systolic blood pressure
was predicted by the increase in connectivity between the insula
and MPFC. Indeed, the insula and MPFC are connected to
preautonomic nuclei, including the periaqueductal gray and para-
brachial nucleus, which modulate both the heart and vasculature
Sleep Loss/Cardiovascular Function Regulation
Psychosomatic Medicine, V 85 •34-41 39 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
(65,66). Moreover, increased resting activity in the insula and
MPFC predicts greater evoked blood pressure, further suggesting
increased sympathetic-mediated blood pressure reactivity influ-
enced by the connectivity changes in the insula and MPFC (62).
Second, a different part of the cardiovascular-control network
predicted elevations in diastolic blood pressure. Specifically, the
magnitude of increased diastolic blood pressure was negatively as-
sociated with increased connectivity between the MPFC and amyg-
dala. The amygdala is instrumental in coordinating behavioral and
physiological adjustments through connections to cortical, hypotha-
lamic, and brainstem nuclei involved in autonomic-cardiovascular
control (62). Thus, sleep deprivation caused an increase in connec-
tivity between the MPFC and amygdala, but the magnitude of this
increase was associated with a smaller increase in diastolic blood
pressure. This may suggest that increased top-down input from the
MPFC to the amygdala served a protective or compensatory mech-
anism that attempted to return blood pressure to its sleep-rested set
point, albeit unsuccessfully.
The vascular system is subject to behavioral, autonomic, and en-
docrine regulation (67), all of which converge within the brain. From
this has emerged evidence that altered blood pressure control, al-
though multifaceted, can be caused by a failure of central brain
control (68). Notable in the current study, neither the increase in
blood pressure nor changes in functional brain connectivity were
associated with the decrease in heart rate. This dissociation not
only suggests that brain network dysfunction impairs cardiovascu-
lar regulation and increases blood pressure, but further indicates
that sleep loss potentially interferes in the coordination of comple-
mentary vascular and cardiac responses—a sign of disorganized
homeostatic control.
More generally, these findings indicate that the state of insuffi-
cient sleep triggers a broad state of biological distress caused by
allostatic overload, a model within which our data are best under-
stood (42). Pragmatically, these results highlight inadequate sleep
throughout society as a risk factor and modifiable intervention tar-
get, determining cardiovascular disease risk through regulation of
cardiovascular-control brain mechanisms.
Study Considerations
Our findings must be appreciated within the context of key limita-
tions. First, because of the sample size, CIs for this relationship
were medium to large and consistent with effect sizes ranging from
small to moderate. Second, the design of the study means that the
afferent-efferent direction of influence between brain and body
cannot be determined. Third, the experiment was performed on
healthy young adults, by design, to investigate any shifts from a
healthy to an unhealthy cardiovascular profile. It remains to be de-
termined whether these findings translate to those with preexisting
hypertension. Fourth, although heavy users of caffeine were ex-
cluded from this study, participants abstained from caffeine use
for the duration of the study, so potential effects of acute caffeine
withdrawal cannot be ruled out. Fifth, the choice of three nights
of sleep stabilization before each experimental condition was se-
lected as a pragmatic balance between participant compliance
and reducing any accumulated sleep debt but may not have been
sufficient to fully eliminate prior sleep debt. Finally, it should be
noted that we assessed cardiovascular function under quiescent
conditions, as common in standardized blood pressure assess-
ments. How our findings would translate to an active cardiovascu-
lar challenge is therefore to be determined. Finally, whether similar
changes in brain-body interrelationships would be observed under
conditions of chronic sleep loss also remains to be determined.
We thank Olivia G. Murillo and Mark Reed for their assis-
tance in running the study.
Source of Funding and Conflicts of Interest: Dr. Walker serves
as a consultant for and has equity interest in Bryte, Oura, Shuni,
and StimScience.
Author Contributions: A.J.K., R.V., E.B.S., and M.P.W. designed
research. A.J.K., R.V., and E.B.S. performed research. A.J.K. ana-
lyzed data. A.J.K. and M.P.W. wrote the article with contributions
from all authors.
REFERENCES
1. Stokes J 3rd,Kannel WB, Wolf PA,D’Agostino RB, Cupples LA. Bloodpressure
as a risk factor for cardiovascular disease. The Framingham Study—30 years of
follow-up. Hypertension 1989;13(5 Suppl):I13–8.
2. Gangwisch JE. A review of evidence for the link between sleep duration and hy-
pertension. Am J Hypertens 2014;27:1235–42.
3. Pepin J-L, Borel A-L, Tamisier R, Baguet J-P, Levy P, Dauvilliers Y. Hyperten-
sion and sleep: overview ofa tight relationship. Sleep Med Rev 2014;18:509–19.
4. Jackowska M, Steptoe A. Sleep and future cardiovascular risk: prospective anal-
ysis fromthe English Longitudinal Study of Ageing. Sleep Med 2015;16:768–74.
5. Kato M, Phillips BG, Sigurdsson G, Narkiewicz K, Pesek CA, Somers VK. Effects
of sleep deprivation on neural circulatory control. Hypertension 2000;35:1173–5.
6. Lusardi P, Zoppi A,Preti P, Pesce RM, Piazza E, Fogari R. Effects of insufficient
sleep on blood pressure in hypertensive patients: a 24-h study. Am J Hypertens
1999;12:63–8.
7. Haack M, Serrador J, Cohen D, Simpson N, Meier-Ewert H, Mullington JM. In-
creasing sleep duration to lower beat-to-beat blood pressure: a pilot study. J Sleep
Res 2013;22:295–304.
8. Wolk R, Gami AS, Garcia-Touchard A, Somers VK. Sleep and cardiovascular
disease. Curr Probl Cardiol 2005;30:625–62.
9. Critchley HD, Mathias CJ, Dolan RJ. Neuroanatomical basis for first- and
second-order representations of bodily states. Nat Neurosci 2001;4:207–12.
10. Whelton PK, Carey RM, Aronow WS, Casey DE, Collins KJ, Dennison Himmelfarb
C, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/
PCNA guideline for the prevention, detection, evaluation, and management of
high blood pressure in adults: a report of the American College of Cardiology/
American Heart Association Task Force on Clinical Practice Guidelines. J Am
Coll Cardiol 2018;71:e127–248.
11. Buysse DJ, Reynolds CF III, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh
Sleep Quality Index: a new instrument for psychiatric practice and research. Psy-
chiatry Res 1989;28:193–213.
12. Muntner P, Shimbo D, Carey RM, Charleston JB, Gaillard T, Misra S, et al. Mea-
surement of blood pressure in humans: a scientific statement from the American
Heart Association. Hypertension 2019;73:e35–66.
13. Kaufmann T, Sütterlin S, Schulz SM, Vögele C. ARTiiFACT: a tool for heart rate
artifact processing and heart rate variability analysis. Behav Res Methods 2011;
43:1161–70.
14. Heart rate variability: standards of measurement, physiological interpretation, and
clinical use. Task Force of the European Society of Cardiology and the North
American Society of Pacing and Electrophysiology. Circulation 1996;93:1043–65.
15. Yasui Y, Breder CD, Safer CB, Cechetto DF. Autonomic responses and efferent
pathways from the insular cortex in the rat. J Comp Neurol 1991;303:355–74.
16. Gianaros PJ, Sheu LK, Matthews KA, Jennings JR, Manuck SB, Hariri AR.
Individual differences in stressor-evoked blood pressure reactivity vary with
activation, volume, and functional connectivity of the amygdala. J Neurosci
2008;28:990–9.
17. Shah A, Chen C, Campanella C, Kasher N, Evans S, Reiff C, et al. Brain cor re-
lates of stress-induced peripheral vasoconstriction in patients with cardiovascular
disease. Psychophysiology 2019;56:e13291.
18. Balint B, Jaremek V, Thorburn V, Whitehead SN, Sposato LA. Left atrial micro-
vascular endothelial dysfunction, myocardial inflammation and fibrosis after se-
lective insular cortex ischemic stroke. Int J Cardiol 2019;292:148–55.
19. Diorio D, Viau V, MeaneyMJ. The role of the medial prefrontal cortex (cingulate
gyrus) in the regulation of hypothalamic-pituitary-adrenal responses to stress. J
Neurosci 1993;13:3839–47.
20. Burford NG, Webster NA, Cruz-Topete D. Hypothalamic-pituitary-adrenal axis
modulation of glucocorticoids in the cardiovascular system. Int J Mol Sci 2017;
18:2150.
21. Frysztak RJ, Neafsey EJ. The effect of medial frontal cortex lesions on car-
diovascular conditioned emotional responses in the rat. Brain Res 1994;
643:181–93.
ORIGINAL ARTICLE
Psychosomatic Medicine, V 85 •34-41 40 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
22. An X, Bandler R, ÖngürD, Price J. Prefrontal cortical projections to longitudinal
columns in the midbrain periaqueductal gray in macaque monkeys. J Comp
Neurol 1998;401:455–79.
23. Chen P, Chen F, Chen G, Zhong S, Gong J, Zhong H, et al. Inflammationis asso-
ciated with decreased functional connectivity of insula in unmedicated bipolar
disorder. Brain Behav Immun 2020;89:615–22.
24. Vecchiarelli HA, Gandhi CP, Gray JM, Morena M, Hassan KI, Hill MN. Di-
vergent responses of inflammatory mediators within the amygdala and me-
dial prefrontal cortex to acute psychological stress. Brain Behav Immun
2016;51:70–91.
25. Behzadi Y, Restom K, Liau J, Liu TT. A component based noise correction method
(CompCor) for BOLD and perfusion based fMRI. Neuroimage 2007;37:90–101.
26. Dampney RA. Functional organization of central pathways regulating the cardio-
vascular system. Physiol Rev 1994;74:323–64.
27. Pool JL, Ransohoff J. Autonomic effects on stimulating rostral portion of cingu-
late gyri in man. J Neurophysiol 1949;12:385–92.
28. Oppenheimer SM, Gelb A, Girvin JP, Hachinski VC. Cardiovascular effects of
human insular cortex stimulation. Neurology 1992;42:1727–32.
29. Valenza G, Sclocco R, Duggento A, Passamonti L, Napadow V, Barbieri R, et al.
The central autonomic network at rest: uncovering functional MRI correlates of
time-varying autonomic outflow. Neuroimage 2019;197:383–90.
30. Brooks JCW, Faull OK, Pattinson KT, Jenkinson M. Physiological noise in
brainstem FMRI. Front Hum Neurosci 2013;7:623.
31. YeoBT, KrienenFM, Sepulcre J, Sabuncu MR, LashkariD, Hollinshead M, et al.
The organization of the human cerebral cortex estimated by intrinsic functional
connectivity. J Neurophysiol 2011;106:1125–65.
32. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and pow-
erful approach to multiple testing. J R Stat Soc B Methodol 1995;57:289–300.
33. Vallat R. Pingouin: statistics in Python. J Open Source Softw 2018;3:1026.
34. Benetos A, Rudnichi A, Thomas F, Safar M, Guize L. Influence of heart rate on
mortality in a French population: role of age, gender, and blood pressure. Hyper-
tension 1999;33:44–52.
35. Holmes AL,Burgess HJ, DawsonD. Effects of sleep pressure on endogenous car-
diac autonomic activity and bodytemperature. J Appl Physiol2002;92:2578–84.
36. Franzen PL,Gianaros PJ, MarslandAL, Hall MH, Siegle GJ, Dahl RE, et al. Car-
diovascular reactivity to acute psychological stress following sleep deprivation.
Psychosom Med 2011;73:679–82.
37. Vaara J, KyröläinenH, Koivu M, Tulppo M, FinniT. The effect of 60-h sleepdep-
rivation on cardiovascular regulation and body temperature. Eur J Appl Physiol
2009;105:439–44.
38. Zhong X, Hilton HJ, Gates GJ,Jelic S, Stern Y, BartelsMN, et al. Increased sym-
pathetic and decreased parasympathetic cardiovascular modulation in normal
humans with acute sleep deprivation. J Appl Physiol 2005;98:2024–32.
39. de Geus EJC, Gianaros PJ, Brindle RC, Jennings JR, Berntson GG. Should heart
rate variability be “corrected”for heart rate?Biological, quantitative, andinterpre-
tive considerations. Psychophysiology 2019;56:e13287.
40. Słomko J, Zawadka-Kunikowska M, Kozakiewicz M, Klawe JJ, Tafil-Klawe M,
Newton JL, et al. Hemodynamic, autonomic, and vascular function changes after sleep
deprivation for 24, 28, and 32 hours in healthy men. Yonsei Med J 2018;59:1138–42.
41. McEwen BS. Stress, adaptation,and disease: allostasisand allostatic load. Ann N
Y Acad Sci 1998;840:33–44.
42. McEwen BS. Sleep deprivation as a neurobiologic and physiologic stressor:
allostasis and allostatic load. Metabolism 2006;55:S20–3.
43. Guo X, Zheng L, Wang J, Zhang X, Zhang X, Li J, et al. Epidemiological evi-
dence for the link between sleep duration and high blood pressure: a systematic
review and meta-analysis. Sleep Med 2013;14:324–32.
44. Wang Y, Mei H, Jiang Y-R, Sun W-Q, Song Y-J, Liu S-J, et al. Relationship be-
tween duration of sleep and hypertension inadults: a meta-analysis. J Clin Sleep
Med 2015;11:1047–56.
45. Bock JM, Vungarala S, Covassin N, Somers VK. Sleep duration and hyperten-
sion: epidemiological evidence and underlying mechanisms. Am J Hypertens
2022;35:3–11.
46. Grandner M, Mullington JM, Hashmi SD, Redeker NS, Watson NF, Morgenthaler
TI. Sleep duration and hypertension: analysisof >700,000 adults by age and sex.
J Clin Sleep Med 2018;14:1031–9.
47. Bliwise DL, Young TB. The parable of parabola: what the U-shaped curve can
and cannot tell us about sleep. Sleep 2007;30:1614–5.
48. Gordan R, Gwathmey JK, Xie L-H. Autonomic and endocrine control of cardio-
vascular function. World J Cardiol 2015;7:204–14.
49. Paton J, Boscan P, Pickering A, NalivaikoE. The yin and yang of cardiac autonomic
control: vago-sympathetic interactions revisited. Brain Res Rev 2005;49:555–65.
50. Chua EC-P, Tan W-Q, Yeo S-C, LauP, Lee I, Mien IH,et al. Heart rate variability
can be used to estimate sleepiness-related decrements in psychomotor vigilance
during total sleep deprivation. Sleep 2012;35:325–34.
51. Wehrens SM, Hampton SM, Skene DJ. Heart rate variability and endothelial
function after sleep deprivation and recovery sleep among male shift and
non-shift workers. Scand J Work Environ Health 2012;38:171–81.
52. Pagani M, Pizzinelli P, Pavy-Le Traon A, Ferreri C, Beltrami S, Bareille M-P,
et al. Hemodynamic, autonomic and baroreflex changesafter one night sleepdep-
rivation in healthy volunteers. Auton Neurosci 2009;145:76–80.
53. Takase B, Akima T, Satomura K, Mastui T, Ishihara M, Kurita A. Effects of
chronic sleep deprivation on autonomic activity by examining heart rate variability,
plasma catecholamine, and intracellular magnesium levels. Biomed Pharmacother
2004;58:S35–9.
54. Nam KC, Kwon MK, Kim DW. Effects of posture and acutesleep deprivation on
heart rate variability. Yonsei Med J 2011;52:569–73.
55. Stein PK, Pu Y. Heart rate variability, sleep and sleep disorders. Sleep Med Rev
2012;16:47–66.
56. Goldstein AN, Greer SM, Saletin JM, Harvey AG, Nitschke JB, Walker MP.
Tired and apprehensive: anxiety amplifies the impact of sleep loss on aversive
brain anticipation. J Neurosci 2013;33:10607–15.
57. Sämann PG, Tully C, Spoormaker VI, Wetter TC, Holsboer F, Wehrle R, et al. Increased
sleep pressure reduces resting state functional connectivity. MAGMA 2010;23:375–89.
58. Jiang B, He D, Guo Z, Gao Z. Effect-size seed-based d mapping of resting-state
fMRI for persistent insomnia disorder. Sleep Breath 2019;24:653–9.
59. Greer SM, Goldstein AN, Walker MP. The impact of sleep deprivation on food
desire in the human brain. Nat Commun 2013;4:2259.
60. Goldstein-Piekarski AN, Greer SM, Saletin JM, Walker MP. Sleep deprivation
impairs the human central and peripheral nervous system discrimination of social
threat. J Neurosci 2015;35:10135–45.
61. Simon EB, Rossi A, Harvey AG, Walker MP. Overanxious and underslept. Nat
Hum Behav 2020;4:100–10.
62. Gianaros PJ, Sheu LK. A review of neuroimaging studies of stressor-evoked
blood pressure reactivity: emerging evidence for a brain-body pathway to coro-
nary heart disease risk. Neuroimage 2009;47:922–36.
63. CritchleyHD. Neural mechanisms of autonomic, affective, and cognitive integra-
tion. J Comp Neurol 2005;493:154–66.
64. Verberne AJ, Lewis SJ, Worland PJ, Beart PM, Jarrott B, Christie MJ, et al. Me-
dial prefrontal corticallesions modulate baroreflex sensitivity in the rat. Brain Res
1987;426:243–9.
65. Augustine JR. Circuitry and functional aspects of the insular lobe in primates in-
cluding humans. Brain Res Rev 1996;22:229–44.
66. Öngür D, Price JL. The organization of networks within the orbital and medial
prefrontal cortex of rats, monkeys and humans. Cereb Cortex 2000;10:206–19.
67. Oparil S, Zaman MA, Calhoun DA. Pathogenesis of hypertension. Ann Intern
Med 2003;139:761–76.
68. Jennings JR, Zanstra Y. Is the brain the essential in hypertension? Neuroimage
2009;47:914–21.
Sleep Loss/Cardiovascular Function Regulation
Psychosomatic Medicine, V 85 •34-41 41 January 2023
Copyright © 2022 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
