How Sleep-Deprived People See and Evaluate Others’ Faces: An Experimental Study

Abstract

Background
Acute sleep loss increases the brain’s reactivity toward positive and negative affective stimuli. Thus, despite well-known reduced attention due to acute sleep loss, we hypothesized that humans would gaze longer on happy, angry, and fearful faces than neutral faces when sleep-deprived. We also examined if facial expressions are differently perceived after acute sleep loss.

Methods
In the present, within-subjects study, 45 young adults participated in one night of total sleep deprivation and one night with an 8-hour sleep opportunity. On the morning after each night, an eye tracker was used to measure participants’ time spent fixating images of happy, angry, fearful, and neutral faces. Participants also evaluated faces’ attractiveness, trustworthiness, and healthiness on a 100-mm visual analog scale.

Results
Following sleep loss, participants struggled more fixating the faces than after sleep. The decrease in total fixation duration ranged from 6.3% to 10.6% after sleep loss (P<0.001). Contrary to our hypothesis, the reduction in total fixation duration occurred irrespective of the displayed emotion (P=0.235 for sleep*emotion interaction) and was also present for the upper (P<0.001) but not the lower part of the faces (except for the lower part of angry faces). Overall, faces were evaluated as less trustworthy (−2.6 mm) and attractive (−3.6 mm) after sleep loss (p<0.05).

Discussion
Facial expressions are crucial for social interactions. Thus, spending less time fixating on faces after acute sleep loss may come along with several problems for social interactions, eg, inaccurate and delayed judgment of the emotional state of others. In addition, more negative social impressions of others may lead to social withdrawal in sleep-deprived humans.

Full text

ORIGINAL RESEARCH
How Sleep-Deprived People See and Evaluate
Others’ Faces: An Experimental Study
Lieve T van Egmond
1,
*, Elisa MS Meth
1,
*, Shervin Bukhari
1
, Joachim Engström
1
, Maria Ilemosoglou
1
,
Jasmin Annica Keller
1
, Shiyang Zhou
1
, Helgi B Schiöth
2
, Christian Benedict
3
1
Department of Surgical Sciences (Sleep Science Laboratory), Uppsala University, Uppsala, Sweden;
2
Department of Surgical Sciences, Functional
Pharmacology, Uppsala University, Uppsala, Sweden;
3
Department of Pharmaceutical Biosciences, Molecular Neuropharmacology (Sleep Science
Laboratory), Uppsala University, Uppsala, 751 24, Sweden
*These authors contributed equally to this work
Correspondence: Lieve T van Egmond, Department of Surgical Sciences (Sleep Science Laboratory), Uppsala University, Husargatan 3, Box 593,
Uppsala, 751 24, Sweden, Tel +46 18-4714123, Email lieve.van.egmond@neuro.uu.se
Background: Acute sleep loss increases the brain’s reactivity toward positive and negative affective stimuli. Thus, despite well-known
reduced attention due to acute sleep loss, we hypothesized that humans would gaze longer on happy, angry, and fearful faces than neutral
faces when sleep-deprived. We also examined if facial expressions are differently perceived after acute sleep loss.
Methods: In the present, within-subjects study, 45 young adults participated in one night of total sleep deprivation and one night with
an 8-hour sleep opportunity. On the morning after each night, an eye tracker was used to measure participants’ time spent xating
images of happy, angry, fearful, and neutral faces. Participants also evaluated faces’ attractiveness, trustworthiness, and healthiness on
a 100-mm visual analog scale.
Results: Following sleep loss, participants struggled more xating the faces than after sleep. The decrease in total xation duration
ranged from 6.3% to 10.6% after sleep loss (P<0.001). Contrary to our hypothesis, the reduction in total xation duration occurred
irrespective of the displayed emotion (P=0.235 for sleep*emotion interaction) and was also present for the upper (P<0.001) but not the
lower part of the faces (except for the lower part of angry faces). Overall, faces were evaluated as less trustworthy (−2.6 mm) and
attractive (−3.6 mm) after sleep loss (p<0.05).
Discussion: Facial expressions are crucial for social interactions. Thus, spending less time xating on faces after acute sleep loss may
come along with several problems for social interactions, eg, inaccurate and delayed judgment of the emotional state of others. In
addition, more negative social impressions of others may lead to social withdrawal in sleep-deprived humans.
Keywords: sleep deprivation, sleep, eye-tracking, gaze, faces, emotions
Introduction
Sleep loss can alter humans’ response to affective stimuli. For example, acute total sleep loss and shorter sleep duration
increase the amygdala activation to images with negative valence.
1,2
This brain region plays a role in reactivity toward
affective stimuli.
3
Relative to sleep, acute sleep deprivation is also associated with enhanced reactivity toward positive
stimuli.
4
For example, acute sleep loss elicits a stronger amygdala activation in response to emotional stimuli judged as
pleasant.
5
The face represents an essential source of information for the observer to identify the other person’s emotional state.
6
In
this context, the amygdala is central to emotional face recognition due to its specic role in stimuli saliency detection.
7
For
example, lesions of the amygdala can impair recognition of fear from facial expressions.
8,9
Similarly, amygdala activation
correlates with the correct identication of angry and fearful faces.
10
The amygdala is also activated in response to other
facial expressions, suggesting that this brain region has a general-purpose function in processing salient information from
faces.
11
Nature and Science of Sleep 2022:14 867–876 867
© 2022 van Egmond et al. This work is published and licensed by Dove Medical Press Limited. The full terms of this license are available at https://www.dovepress.com/
terms.php and incorporate the Creative Commons Attribution – Non Commercial (unported, v3.0) License (http://creativecommons.org/licenses/by-nc/3.0/). By accessing
the work you hereby accept the Terms. Non-commercial uses of the work are permitted without any further permission from Dove Medical Press Limited, provided the work is properly attributed.
For permission for commercial use of this work, please see paragraphs 4.2 and 5 of our Terms (https://www.dovepress.com/terms.php).
Nature and Science of Sleep Dovepress
open access to scientific and medical research
Open Access Full Text Article
Received: 29 January 2022
Accepted: 9 April 2022
Published: 2 May 2022

Several studies have been published regarding how sleep-deprived people respond to emotional and neutral faces,
including impaired accuracy and slower speed identifying emotional expression in faces.
12,13
In addition, patients
suffering from insomnia exhibit a visual attentional bias toward tired faces and misrecognize angry faces as fearful
faces.
14–16
Whether faces with negative or positive emotional valence are more visually salient and differently evaluated
regarding social appeal when the observer is sleep-deprived is, however, not well-researched.
Thus, in the present within-subjects experiment, 45 young adults spent one night asleep and one night awake in our
laboratory. In the morning after each night, we used eye-tracking technology to examine possible differences in dwell time on
standardized images of happy, neutral, angry, and fearful faces. Acute sleep loss causes reductions in sustained attention.
17
Thus, we expected that participants would spent less time xating on the faces after sleep loss than after sleep. Given the bias
toward negative and positive affect following sleep loss,
1,5
we nonetheless expected that the time spent xating on emotional
faces would be longer than for neutral faces after sleep loss. Acute sleep loss may reduce the drive to interact with others
socially. For example, using a standardized social distance task, a previous study showed that participants enforce greater
social separation from others following one night of sleep deprivation.
18
Thus, we also examined whether participants would
rate the attractiveness, trustworthiness, and healthiness of faces differently when sleep-deprived.
Materials and Methods
Participants
We recruited study participants via local billboards and an online recruitment website. A total of 508 potential
participants lled out an online survey to assess their general health status, including sleep habits. The following
exclusion criteria were applied:
●When participants indicated that their overall sleep quality was either bad or very bad
●Presence of extreme chronotype (Morningness-Eveningness Questionnaire score ≤30 or ≥70; ref.
19
)
●Self-reported habitual daily sleep duration of fewer than 7 hours
●Habitual bedtime on working days before 22:00 or later than 24:00
●Participants answered with yes to the question “Do you have any physical and/or mental illnesses (including chronic
diseases) ?”
●Regular medication use (eg, medicine used to treat allergies)
●Drugs or nicotine usage
●More than ve standard units of alcohol or caffeine beverages per day
●Self-reported uncorrected vision problems (eg, myopia)
●Planned travel across time zones within three months of the study period or between sessions.
Of the 508 people who lled in the online survey, 81% were excluded, as they met at least one of the exclusion
criteria. The remaining 96 potential participants were invited for an onsite screening session. Subsequently, we
excluded 49 additional potential participants due to discrepancies in their online screening and onsite interview
responses. Following these exclusions, 47 participants remained for inclusion into the study (mean ± SD, age: 25
± 3 years; 20 women). All women included in the study were on hormonal contraceptives to minimize potential
confound by the menstrual cycle on facial exploration and evaluation.
20,21
Due to unexpected hardware and
software issues with the eye tracker, seven eye-tracker recordings were not available for analysis. One participant
only slept 02h:45min according to polysomnography (PSG), ie, the participant slept half shorter than the intended
8 hours. Thus, we excluded all data collected in the sleep condition of this participant. The nal analytical sample
consisted of 45 participants (including 20 women).
The study was conducted following the Declaration of Helsinki, and the regional ethics board in Uppsala
approved experimental procedures (DNR2017/560). All subjects gave written informed consent before experi-
ments and were compensated for participation in the study. The experiment described herein is part of a more
extensive study investigating the possible health consequences of sleep loss. No other results have been
https://doi.org/10.2147/NSS.S360433
DovePress
Nature and Science of Sleep 2022:14
868
van Egmond et al Dovepress
Powered by TCPDF (www.tcpdf.org)

published at the time of this publication. Experimental sessions took place between March 2018 and
November 2020.
Study Design and Procedure
A scheme of the study can be found in Figure 1. The present crossover study consisted of two experimental conditions
separated by at least one week: one night of total sleep deprivation vs one night with an 8-hour sleep opportunity. The
order of experimental conditions was almost perfectly counterbalanced across participants. Study sessions of female
participants were scheduled outside of their menstrual phase. Within the week before the rst experimental session,
participants visited the lab for an adaptation night. The adaptation night served to adjust subjects to the experimental
setting and was conducted under the conditions of the experiment, including the attachment of electrodes for sleep
recordings.
On experimental nights, subjects were served a standardized dinner upon arrival at the laboratory at 19:00. In the sleep
condition, sleep was allowed between ~23:00 and 07:00. PSG recordings were performed using SOMNO HD (10–20
system; SOMNOmedics GmbH, Randersacker, Germany). Sleep stage scoring was performed following the American
Academy of Sleep Medicine criteria.
22
Participants’ slept on average seven hours and twelve minutes in the sleep condition
(for more details, see Supplementary Results and Supplementary Table S1). In the sleep loss session, an experimenter
continuously monitored participants, who were allowed to spend their time between 23:00 and 07:00 with a selection of
movies, games, and books under normal indoor light conditions (~500 lux); however, no food intake, including caffeinated
beverages was allowed. On the morning after sleep and sleep loss, participants performed the eye-tracking task at ~08:00.
Eye-Tracking Paradigm
The eye-tracking paradigm included high-resolution and standardized images of adult male and female actors (18–40
years). All pictures were derived from the Chicago Face Database.
23
For each face category (happy, fearful, angry, and
neutral), six images (three male, three female) were randomly selected from the Caucasian subgroup of the Chicago Face
Database, resulting in 24 pictures of 24 different individuals (can be retrieved upon request from the correspondence
Figure 1 Experimental scheme. (A) Order of experimental conditions. (B) Overview of the eye-tracking paradigm.
Notes: Reprinted from Ma DS, Correll J, Wittenbrink B. The Chicago face database: a free stimulus set of faces and norming data. Behav Res Methods. 2015;47(4):1122–1135
with persmission from the Chicago Face Database
23
with graphic images created with BioRender.com.
Abbreviations: ET, eye-tracking; PSG, polysomnography.
Nature and Science of Sleep 2022:14 https://doi.org/10.2147/NSS.S360433
DovePress 869
Dovepress van Egmond et al
Powered by TCPDF (www.tcpdf.org)

author L.v.E.). We used the same 24 face images in both experimental conditions to avoid variability in dwell time on and
evaluation of faces due to different sets of photographed actors. For the experiment, we created two versions of the face
paradigm using a random number generator to ensure that participants were not presented the faces in the same order in
both conditions. The two versions were almost perfectly counterbalanced across participants’ experimental nights.
Participants sat between 60 and 70 cm in front of the eye tracker screen. We used a headrest to minimize head
movement. Participants looked at a xation cross in the middle of the screen for three seconds before they could freely
explore each stimulus for 5 seconds. Then participants rated the actor’s healthiness (“How healthy do you deem this
person?”), attractiveness (“How attractive do you deem this person?”), and trustworthiness (“How trustworthy do you
deem this person?”) on a 100-mm visual analog scale (0= not at all; 100= very much). After a practice run to ensure that
the subject understood the instructions, the eye-tracking paradigm started. We did not ask the participants to identify the
emotion communicated by each displayed face, as this would have caused unwanted priming effects.
Eye Tracker Hardware
Participants’ visual gaze was assessed using a Tobii Pro Spectrum eye tracker sampling at 600Hz with binocular tracking
and Tobii Pro Lab software (Tobii© Technology AB, Stockholm, Sweden). The screen showing the stimuli measured
23.8” and 1920×1080 pixels.
As summarized in Figure 2, the same procedure was manually applied to each face to determine the area of interest
(AOI). For the AOI for the upper face region, including the eyes, we rst dened the middle point between both pupils of
the eyes. We then drew a vertical line from this midpoint to the tip of the nose (eye-nose line). We inserted a horizontal
line at 50% of the eye-nose line in the next step (lower bound). To catch the eyebrow, we vertically ipped the eye-nose
line (ie, in the direction of the forehead). We then inserted a horizontal line at 75% of the ipped eye-nose line (eyebrow
bound). The root of the helix of each ear was used as a vertical boundary. To catch forehead wrinkles that can affect
observers’ interpretation of facial expressions,
24
the resulting area between the lower and eyebrow bounds was then
duplicated and ipped at the horizontal level of the eyebrow bound and was dened as the AOI for the upper face region,
Figure 2 Facial locations used to dene the areas of interest for analysis.
Notes: Facial image reproduced from Ma DS, Correll J, Wittenbrink B. The Chicago face database: a free stimulus set of faces and norming data. Behav Res Methods. 2015;47
(4):1122–1135 with persmission from the Chicago Face Database.
23
https://doi.org/10.2147/NSS.S360433
DovePress
Nature and Science of Sleep 2022:14
870
van Egmond et al Dovepress
Powered by TCPDF (www.tcpdf.org)

including the eyes. The AOI for the lower face region, including the mouth, included the area between the lower bound,
the chin-top, and the vertical boundaries.
Eye-Tracking Outcome Variables
Participants were free to explore each face. However, they were instructed to always look at the xation cross in the
center of the screen during the 3-seconds inter-stimulus interval to ensure that they would start exploring each face from
the same screen position. Thus, the rst 300 ms of each 5-seconds long face stimulus presentation was discarded.
Fixation duration was dened as the time during which participants’ eyes rested on the face using the classications
from the Tobii I–VT Fixation lter, as described elsewhere.
25
For the analysis, the total xation duration (TFD),
expressed as a percentage of the maximum of 4.7 seconds, was used as the primary outcome. TFD is a cumulative
measure of the time participants’ eyes xate a stimulus. Additionally, we measured the total number of xation points
(TFP) during the 4.7 seconds. TFP indicates the number of true xation points participants’ exhibit when visually
exploring a stimulus (here: faces).
To explore possible spatial differences in eye-gaze patterns between sleep loss and sleep, we also compared TFD and
TFP for the lower and upper parts of the face. Single probes in either condition were excluded if the eye-tracker did not
detect eyes during the presentation of the face.
Statistical Analysis
All analyses were performed using IBM SPSS Statistics 26 (SPSS Inc. Chicago, IL, USA). Generalized linear mixed
models (GLMM) were used to compare the main effects of the within-subjects factor SLEEP (ie, sleep vs sleep
deprivation) and EMOTION (happy vs neutral vs angry vs fearful). We also modeled possible interactions of SLEEP
with EMOTION. Post-hoc testing was performed using pairwise t–tests with LSD adjustment for multiple comparisons.
We also tested whether the effects of SLEEP on visual exploration and ratings may vary by individual faces within each
Table 1 Total Fixation Duration (Expressed as % from the Maximum Dwell Time of 4.7 Seconds) for Faces
with Different Emotional Expressions Following Sleep Loss and Sleep
Facial Area Facial Expression Sleep Loss, % Sleep, %
Whole face Happy 72.0 [69.3, 74.7]*** 81.4 [79.6, 83.2]
Fearful 72.7 [69.9, 75.5]*** 79.9 [77.9, 82.0]
Angry 70.6 [67.6, 73.5]*** 81.2 [79.4, 83.0]
Neutral 74.9 [72.4, 77.4]*** 81.2 [79.4, 83.0]
Upper part of the face, including the eyes Happy 46.7 [43.6, 49.8]*** 56.6 [54.0, 59.1]
Fearful 47.8 [44.8, 50.8]*** 55.1 [52.6, 57.7]
Angry 51.8 [48.7, 54.8]*** 58.9 [56.3, 61.6]
Neutral 50.8 [47.8, 53.7]*** 59.2 [56.7, 61.7]
Lower part of the face, including the mouth Happy 24.2 [21.6, 26.9] 24.4 [22.4, 26.4]
Fearful 24.2 [21.7, 26.7] 24.8 [22.8, 26.8]
Angry 18.4 [16.2, 20.6]* 22.0 [20.1, 23.9]
Neutral 24.0 [21.3, 26.6] 21.7 [19.8, 23.6]
Notes: Values are shown as estimated marginal means [95%-CI]. Post hoc testing was performed using pairwise comparisons using
Student’s t–tests with LSD adjustment for multiple comparisons. *p<0.05; ***p<0.001.
Nature and Science of Sleep 2022:14 https://doi.org/10.2147/NSS.S360433
DovePress 871
Dovepress van Egmond et al
Powered by TCPDF (www.tcpdf.org)

emotional face category. Unless otherwise specied, data are reported as estimated mean [95%-CI], as derived from the
GLMM. P values<0.05 were considered signicant.
Results
Eye-Tracking results
Following sleep loss, the TFD decreased by about 8% compared to sleep (sleep loss vs sleep: 72.7% [71.3%, 74.0%] vs
81.0% [80.0%, 81.9%]; p<0.001 for SLEEP). In contrast, no main effect for EMOTION was found (happy vs fearful vs
angry vs neutral faces: 76.9% [75.4%, 78.4%] vs 76.2% [74.5%, 77.8%] vs 76.4% [74.9%, 78.0%] vs 77.7% [76.3%,
79.2%], p=0.497). Also, the SLEEP* EMOTION interaction did not reach signicance (p=0.235; differences in TFD
between the conditions ranged from 6.3% to 10.6%, Table 1).
When using the TFD for the upper face, we found a main effect for SLEEP (sleep loss vs sleep: 49.3% [47.8%,
50.8%] vs 57.5% [56.2%, 58.7%]; p<0.001) and EMOTION (happy vs fearful vs angry vs neutral faces: 51.8% [49.8%,
53.8%] vs 51.4% [49.5%, 53.3%] vs 55.3% [53.3%, 57.3%] vs 55.0% [53.1%, 56.9%], p=0.005). As suggested by the
GLMM, the SLEEP effect on the TFD for the upper face did not vary by EMOTION (p=0.770 for SLEEP* EMOTION;
differences in TFD-upper face between the conditions ranged from −9.9% to −7.2%, Table 1).
When analyzing the TFD for the lower part of the face, we found no main effect for SLEEP (sleep loss vs sleep:
22.5% [21.3%, 23.8%] vs 23.2% [22.3%, 24.2%]; p=0.381) but EMOTION (happy vs fearful vs angry vs neutral faces:
24.2% [22.7%, 25.8%] vs 24.5% [22.9%, 26.0%] vs 20.4% [19.0%, 20.9%] vs 22.4% [20.8%, 23.9%], p<0.001).
Furthermore, there was a trend for an interaction of SLEEP with EMOTION (P=0.065). Subsequent post-hoc compar-
isons showed that participants in the sleep loss condition spent about 3.6% less time xating on the lower part of the face
than after sleep (p=0.015; Table 1). The difference in TFD for the lower part of the face did not reach signicance for the
other face categories (p≥0.171; Table 1).
Figure 3 Face ratings following one night of sleep and sleep loss. Ratings collected during one sleep loss condition were not analyzed since the participant misinterpreted
the instructions. Values shown as mean [95%-CI]. **p<0.01; *p<0.05.
Abbreviations: ns, non-signicant; VAS, visual analogue scale.
https://doi.org/10.2147/NSS.S360433
DovePress
Nature and Science of Sleep 2022:14
872
van Egmond et al Dovepress
Powered by TCPDF (www.tcpdf.org)

The effects of SLEEP on the eye-tracking outcomes did not vary by the individual face when running a GLMM for
each face category separately (p≥0.610 for all interactions). Results when using the TFP as a dependent variable are
summarized in the Supplementary Results and Supplementary Table S2.
Ratings
As summarized in Figure 3, participants evaluated the faces as less attractive and trustworthy after sleep loss than sleep
(the mean difference between sleep loss and sleep for attractiveness: −3.6 mm [−5.6 mm, −1.7 mm]; and trustworthiness:
−2.6 mm [−4.4 mm, −0.8 mm]; p≤0.004 for SLEEP). Ratings regarding healthiness did not differ signicantly between
sleep loss and sleep (the mean difference between sleep loss and sleep: −1.2 mm [−3.0 mm, 0.6 mm]; p=0.180 for
SLEEP; Figure 3). We also found a main effect of EMOTION for attractiveness (happy vs fearful vs angry vs neutral
faces: 54.7 mm [52.5 mm, 56.8 mm] vs 49.9 mm [39.1 mm, 42.6 mm] vs 45.1 mm [43.0 mm, 47.2 mm] vs 39.0 mm
[37.2 mm, 40.8 mm]; p<0.001), trustworthiness (happy vs fearful vs angry vs neutral faces: 63.3 mm [61.4 mm,
65.3 mm] vs 45.3 mm [43.7 mm, 46.9 mm] vs 39.3 mm [37.5 mm, 41.0 mm] vs 45.2 mm [43.3 mm, 47.1 mm];
p<0.001), and healthiness (happy vs fearful vs angry vs neutral faces: 68.2 mm [66.6 mm, 69.8 mm] vs 53.3 mm
[51.6 mm, 55.1 mm] vs 57.7 mm [55.9 mm, 59.4 mm] vs 52.6 mm [50.7 mm, 54.5 mm]; p<0.001).
Post-hoc comparisons revealed that images of angry faces were rated as less trustworthy (−4.9 mm, P=0.005) and
healthy (−4.2 mm, P=0.020), whereas fearful and neutral faces were rated as less attractive after sleep loss (−3.9 mm and
−5.5 mm; p≤0.030; Table 2). In contrast, ratings for happy faces did not differ between sleep loss and sleep (Table 2).
However, the SLEEP* EMOTION interaction was not signicant for any of the ratings (p≥0.187). Finally, the effects of
SLEEP on ratings did not vary by the individual face when running a GLMM for each face category separately (p≥0.658
for attractiveness; p≥0.594 for trustworthiness; and p≥0.838 for healthiness).
Discussion
Previous human studies showed that images with negative or positive affective valence more signicantly activate brain
regions involved in processing relevant stimuli when the observer is sleep-deprived.
1,2,5
With these results in mind, we
Table 2 Social Appeal Ratings for Faces with Different Emotional Expressions Following Sleep Loss and
Sleep
Question Facial Expression Sleep Loss, mm Sleep, mm
How attractive do you deem this person?
0 mm, not at all; 100 mm, very much
Happy 54.8 [51.7, 58.0] 55.0 [52.1, 57.8]
Fearful 38.9 [36.4, 41.5]* 42.8 [40.4, 45.2]
Angry 43.5 [40.4, 46.5] 46.8 [44.0, 49.7]
Neutral 36.3 [33.7, 38.9]** 41.8 [39.1, 44.4]
How trustworthy do you deem this person?
0 mm, not at all; 100 mm, very much
Happy 63.2 [60.5, 66.0] 63.7 [61.0, 66.4]
Fearful 43.8 [41.4, 46.2] 46.8 [44.5, 49.1]
Angry 36.7 [34.3, 39.2]** 41.6 [39.2, 44.1]
Neutral 44.8 [42.1, 47.4] 45.6 [43.0, 48.3]
How healthy do you deem this person?
0 mm, not at all; 100 mm, very much
Happy 68.8 [66.4, 71.2] 67.8 [65.6, 69.9]
Fearful 52.6 [50.0, 55.2] 54.1 [51.6, 56.5]
Angry 55.6 [53.0, 58.1]* 59.7 [57.4, 60.1]
Neutral 52.4 [49.7, 55.1] 52.8 [50.1, 55.5]
Notes: Values are shown as estimated marginal means [95%-CI]. Post hoc testing was performed using pairwise comparisons using
Student’s t–tests with LSD adjustment for multiple comparisons.*p<0.05; **p<0.01.
Nature and Science of Sleep 2022:14 https://doi.org/10.2147/NSS.S360433
DovePress 873
Dovepress van Egmond et al
Powered by TCPDF (www.tcpdf.org)

hypothesized that happy, angry, and fearful faces are more visually salient than neutral faces under conditions of acute
sleep loss, ie, they are more likely to catch the observer’s gaze.
We found that participants spent less time xating on the faces after acute total sleep loss than after sleep. Several central
nervous mechanisms could explain the decrease in participants’ total xation duration after sleep loss. For example, sleep
loss causes reductions in the speed of processing and attention.
17,26
Furthermore, sleepiness, inevitably caused by overnight
wakefulness, may outcompete other goal-directed behaviors.
27
Thus, the sleep-deprived brain may be less capable of
exerting efforts to explore the face visually. Reduced attention and sleepiness may also explain why participants, contrary to
our hypothesis, did not spend more time xating on emotional faces than they did on neutral faces.
We also examined whether participants’ total xation duration for the upper and lower parts of the face differed
between sleep loss and sleep. Acute sleep loss diminished gaze duration for the upper part of the faces, irrespective of the
expressed emotion. The time spent xating the lower part of the faces did not differ between sleep loss and sleep, except
for reduced gaze duration of angry faces. As suggested by previous work, recognizing human emotional facial
expressions mainly relies on the eyes and the mouth regions.
28
A reduced dwell time on the eyes (and mouth for
angry faces) seen herein could explain why others have shown that sleep-deprived healthy adults are signicantly less
accurate at identifying happy, sad, and angry faces.
12,13
When sleep-deprived, humans are also slower to recognize
emotional facial expressions.
13
A diminished or delayed ability to identify the emotional state of others can have negative
consequences for the sleep-deprived, eg, in occupations where workers who are routinely restricted of sleep are also
responsible for making judgments about other people’s emotional state (eg, police, security, and teachers). Of note, these
recognition decits are likely to disappear after recovery sleep.
12
The present study found that angry faces were deemed less trustworthy and healthy after sleep loss. Intriguingly, as
suggested by results from a previous meta-analysis, the greater the amygdala activation – a brain activity pattern seen in
sleep-deprived humans viewing images with negative valence after acute sleep loss
1
–, the more likely are observers to rate
faces as untrustworthy.
29
We also found that neutral and fearful faces were rated less attractive after sleep loss. In contrast,
the rating of happy faces, the most socially appealing face category in the present study, remained unaffected by sleep loss.
Similar ndings for the effects of acute sleep loss on social impressions have been found after chronic sleep loss.
Specically, when undergoing a chronic sleep loss intervention, humans deem individuals with less positive and more
negative facial features as more dangerous. However, these biases are not present when participants are rested.
30
More
negative social impressions of others could explain why acute sleep loss may promote social withdrawal and loneliness.
18
Several strengths apply to our study. The experiment was performed under highly controlled in-laboratory conditions.
Moreover, women used hormonal contraceptives to reduce possible confounding by the menstrual cycle on facial
exploration and evaluation.
20,21
However, several limitations apply to our study. First, we chose only face images of
Caucasian actors. A second limitation is that our participants were young adults. Our study focused on the effects of
acute sleep loss on happy, neutral, angry, and fearful facial expression. Thus, our data cannot be extrapolated to other
facial expressions. Finally, our results are based on non-dynamic photographs. Future studies should therefore examine if
dynamic face representations, eg, using real actors or pre-recorded videos reveal similar results.
Conclusion
Among non-verbal signals, facial expressions are crucial for social interactions.
31–33
Thus, spending less time xating on
faces after acute sleep loss may come along with several problems for social interactions, eg, inaccurate and delayed
judgment of the emotional state of others. In addition, more negative social impressions of others, as shown herein, may
increase the risk of the sleep-deprived becoming socially isolated.
Acknowledgments
C.B.’s research is nancially supported by the Novo Nordisk Foundation [NNF19OC0056777] and the Swedish Brain
Research Foundation [FO2020-0044]. H.B.S. was supported by Uppsala University for the purchase of the equipment.
The funding sources had no input in the design and conduct of this study, in the collection, analysis, and interpretation of
the data, or the manuscript’s preparation, review, or approval. We would like to thank the anonymous academic peer
reviewers who provided helpful and detailed comments on earlier drafts of this review manuscript.
https://doi.org/10.2147/NSS.S360433
DovePress
Nature and Science of Sleep 2022:14
874
van Egmond et al Dovepress
Powered by TCPDF (www.tcpdf.org)

Disclosure
Between 2020 and 2021, Dr Christian Benedict was part of the scientic advisory board for Repha GmBH. The activity
was unrelated to the submitted work. The authors report no other conicts of interest in this work.
References
1. Yoo SS, Gujar N, Hu P, Jolesz FA, Walker MP. The human emotional brain without sleep - a prefrontal amygdala disconnect. Curr Biol.2007;17
(20):R877–R878. doi:10.1016/j.cub.2007.08.007
2. Reidy BL, Hamann S, Inman C, Johnson KC, Brennan PA. Decreased sleep duration is associated with increased fMRI responses to emotional
faces in children. Neuropsychologia.2016;84:54–62. doi:10.1016/j.neuropsychologia.2016.01.028
3. Murray EA. The amygdala, reward and emotion. Trends Cogn Sci.2007;11(11):489–497. doi:10.1016/j.tics.2007.08.013
4. Ben Simon E, Vallat R, Barnes CM, Walker MP. Sleep loss and the socio-emotional brain. Trends Cogn Sci.2020;24(6):435–450. doi:10.1016/j.
tics.2020.02.003
5. Gujar N, Yoo SS, Hu P, Walker MP. Sleep deprivation amplies reactivity of brain reward networks, biasing the appraisal of positive emotional
experiences. J Neurosci.2011;31(12):4466–4474. doi:10.1523/JNEUROSCI.3220-10.2011
6. Ekman P, Oster H. Facial expressions of emotion. Annu Rev Psychol.1979;30:527–554. doi:10.1146/annurev.ps.30.020179.002523
7. Adolphs R. Fear, faces, and the human amygdala. Curr Opin Neurobiol.2008;18(2):166–172. doi:10.1016/j.conb.2008.06.006
8. Adolphs R, Tranel D, Damasio H, Damasio AR. Fear and the human amygdala. J Neurosci.1995;15(9):5879–5891. doi:10.1523/jneurosci.15-09-
05879.1995
9. Graham R, Devinsky O, LaBar KS. Quantifying decits in the perception of fear and anger in morphed facial expressions after bilateral amygdala
damage. Neuropsychologia.2007;45(1):42–54. doi:10.1016/j.neuropsychologia.2006.04.021
10. Loughead J, Gur RC, Elliott M, Gur RE. Neural circuitry for accurate identication of facial emotions. Brain Res.2008;1194:37–44. doi:10.1016/j.
brainres.2007.10.105
11. Fitzgerald DA, Angstadt M, Jelsone LM, Nathan PJ, Phan KL. Beyond threat: amygdala reactivity across multiple expressions of facial affect.
Neuroimage.2006;30(4):1441–1448. doi:10.1016/j.neuroimage.2005.11.003
12. Van Der Helm E, Gujar N, Walker MP. Sleep deprivation impairs the accurate recognition of human emotions. Sleep.2010;33(3):335–342.
doi:10.1093/sleep/33.3.335
13. Cote KA, Mondloch CJ, Sergeeva V, Taylor M, Semplonius T. Impact of total sleep deprivation on behavioural neural processing of emotionally
expressive faces. Exp Brain Res.2014;232(5):1429–1442. doi:10.1007/s00221-013-3780-1
14. Akram U, Ellis JG, Myachykov A, Barclay NL. Preferential attention towards the eye-region amongst individuals with insomnia. J Sleep Res.
2017;26(1):84–91. doi:10.1111/jsr.12456
15. Akram U, Robson A, Ypsilanti A. Sleep-related attentional bias for faces depicting tiredness in insomnia: evidence from an eye-tracking study.
J Clin Sleep Med.2018;14(6):959–965. doi:10.5664/jcsm.7160
16. Akram U. Categorization decit of facially expressed anger in insomnia (Commentary on Zhang et al. Individuals with insomnia misrecognize
angry faces as fearful faces while missing the eyes: an eye-tracking study). Sleep.2019;42(2). doi:10.1093/sleep/zsz012
17. Hudson AN, Van Dongen HPA, Honn KA. Sleep deprivation, vigilant attention, and brain function: a review. Neuropsychopharmacology.2020;45
(1):21–30. doi:10.1038/s41386-019-0432-6
18. Ben Simon E, Walker MP. Sleep loss causes social withdrawal and loneliness. Nat Commun.2018;9(1):3146. doi:10.1038/s41467-018-05377-0
19. Horne JA, Ostberg O. A self assessment questionnaire to determine morningness eveningness in human circadian rhythms. Int J Chronobiol.
1976;4(2):97–110.
20. Derntl B, Windischberger C, Robinson S, et al. Facial emotion recognition and amygdala activation are associated with menstrual cycle phase.
Psychoneuroendocrinology.2008;33(8):1031–1040. doi:10.1016/j.psyneuen.2008.04.014
21. DeBruine L, Jones BC, Frederick DA, Haselton MG, Penton-Voak IS, Perrett DI. Evidence for menstrual cycle shifts in women’s preferences for
masculinity: a response to Harris (in press) “menstrual cycle and facial preferences reconsidered.”. Evol Psychol.2010;8(4):147470491000800.
doi:10.1177/147470491000800416
22. Berry RB, Brooks R, Gamaldo CE, et al. The AASM Manual for the Scoring of Sleep and Associated Events. Rules, Terminology and Technical
Specications Version 2.2. Darien, Illinois: American Academy of Sleep Medicine; 2015
23. Ma DS, Correll J, Wittenbrink B. The Chicago face database: a free stimulus set of faces and norming data. Behav Res Methods.2015;47
(4):1122–1135. doi:10.3758/s13428-014-0532-5
24. Hess U, Adams RB, Simard A, Stevenson MT, Kleck RE. Smiling and sad wrinkles: age-related changes in the face and the perception of emotions
and intentions. J Exp Soc Psychol.2012;48(6):1377–1380. doi:10.1016/j.jesp.2012.05.018
25. Olsen A. The Tobii I-VT xation lter. Tobii Technol.2012;21:45.
26. Durmer JS, Dinges DF. Neurocognitive consequences of sleep deprivation. Semin Neurol.2005;25(1):117–129. doi:10.1055/s-2005-867080
27. Axelsson J, Ingre M, Kecklund G, Lekander M, Wright KP, Sundelin T. Sleepiness as motivation: a potential mechanism for how sleep deprivation
affects behavior. Sleep.2020;43(6):1–6. doi:10.1093/sleep/zsz291
28. Eisenbarth H, Alpers GW. Happy mouth and sad eyes: scanning emotional facial expressions. Emotion.2011;11(4):860–865. doi:10.1037/a0022758
29. Santos S, Almeida I, Oliveiros B, Castelo-Branco M. The role of the amygdala in facial trustworthiness processing: a systematic review and
meta-analyses of fMRI studies. PLoS One.2016;11:11. doi:10.1371/journal.pone.0167276
30. Alkozei A, Haack M, Skalamera J, et al. Chronic sleep restriction affects the association between implicit bias and explicit social decision making.
Sleep Heal.2018;4(5):456–462. doi:10.1016/j.sleh.2018.07.003
31. Frith C. Role of facial expressions in social interactions. Philos Trans R Soc B Biol Sci.2009;364(1535):3453–3458. doi:10.1098/rstb.2009.0142
32. De Gelder B. Why bodies? Twelve reasons for including bodily expressions in affective neuroscience. Philos Trans R Soc B Biol Sci.2009;364
(1535):3475–3484. doi:10.1098/rstb.2009.0190
33. Adolphs R. Social cognition and the human brain. Trends Cogn Sci.1999;3(12):469–479. doi:10.1016/S1364-6613(99)01399-6
Nature and Science of Sleep 2022:14 https://doi.org/10.2147/NSS.S360433
DovePress 875
Dovepress van Egmond et al
Powered by TCPDF (www.tcpdf.org)

Nature and Science of Sleep Dovepress
Publish your work in this journal
Nature and Science of Sleep is an international, peer-reviewed, open access journal covering all aspects of sleep science and sleep medicine,
including the neurophysiology and functions of sleep, the genetics of sleep, sleep and society, biological rhythms, dreaming, sleep disorders
and therapy, and strategies to optimize healthy sleep. The manuscript management system is completely online and includes a very quick and fair
peer-review system, which is all easy to use. Visit http://www.dovepress.com/testimonials.php to read real quotes from published authors.
Submit your manuscript here: https://www.dovepress.com/nature-and-science-of-sleep-journal
DovePress Nature and Science of Sleep 2022:14
876
van Egmond et al Dovepress
Powered by TCPDF (www.tcpdf.org)