Pressbild John Axelsson. Foto: Sara Appelgren

John Axelsson


Visa sidan på svenska
Visiting address Frescati Hagväg 16 A
Room A 128
Postal address Stressforskningsinstitutet 106 91 Stockholm

About me

Sleep is an essential biological phenomonea that is exceptionally complex and dynamic. Still, the function of sleep is one of the largest riddles within biology, and we can look forward to many exciting discoveries in this research field. I am fascinated by sleep, and is presently trying to determine the biolgical and cultrural influences - such as living in the modern 24-hours society with peculiarities such as night work and artificial lighting - on how we sleep and the consequences thereof.


John Axelssons and his research team at the Stress Research Institute aim to increase the knowledge and awareness of how sleep in the modern life affects our biology, cognition and health. Some of the research questions we pursue include:

- How sleep loss affects the brain and why some are vulnarable than others?

- How fast can you trust an awakening brain?

- How do parents to young children sleep and what can be done to support their sleep?

- Kan more sleep speed up recovery from infections?


A selection from Stockholm University publication database
  • 2018. John Axelsson, Julie Lasselin, Mats Lekander. BMJ. British Medical Journal 360
  • 2017. Christina Regenbogen (et al.). Proceedings of the National Academy of Sciences of the United States of America 114 (24), 6400-6405

    Throughout human evolution, infectious diseases have been a primary cause of death. Detection of subtle cues indicating sickness and avoidance of sick conspecifics would therefore be an adaptive way of coping with an environment fraught with pathogens. This study determines how humans perceive and integrate early cues of sickness in conspecifics sampled just hours after the induction of immune system activation, and the underlying neural mechanisms for this detection. In a double-blind placebo-controlled crossover design, the immune system in 22 sample donors was transiently activated with an endotoxin injection [lipopolysaccharide (LPS)]. Facial photographs and body odor samples were taken from the same donors when sick (LPS-injected) and when healthy (saline-injected) and subsequently were presented to a separate group of participants (n = 30) who rated their liking of the presented person during fMRI scanning. Faces were less socially desirable when sick, and sick body odors tended to lower liking of the faces. Sickness status presented by odor and facial photograph resulted in increased neural activation of odor-and faceperception networks, respectively. A superadditive effect of olfactory-visual integration of sickness cues was found in the intraparietal sulcus, which was functionally connected to core areas of multisensory integration in the superior temporal sulcus and orbitofrontal cortex. Taken together, the results outline a disease-avoidance model in which neural mechanisms involved in the detection of disease cues and multisensory integration are vital parts.

  • 2017. Ellen R. Stothard (et al.). Current Biology 27 (4), 508-513

    Reduced exposure to daytime sunlight and increased exposure to electrical lighting at night leads to late circadian and sleep timing [1-3]. We have previously shown that exposure to a natural summer 14 hr 40 min:9 hr 20 min light-dark cycle entrains the human circadian clock to solar time, such that the internal biological night begins near sunset and ends near sunrise [1]. Here we show that the beginning of the biological night and sleep occur earlier after a week's exposure to a natural winter 9 hr 20 min:14 hr 40 min light-dark cycle as compared to the modern electrical lighting environment. Further, we find that the human circadian clock is sensitive to seasonal changes in the natural light-dark cycle, showing an expansion of the biological night in winter compared to summer, akin to that seen in non-humans [4-8]. We also show that circadian and sleep timing occur earlier after spending a weekend camping in a summer 14 hr 39 min:9 hr 21 min natural light-dark cycle compared to a typical weekend in the modern environment. Weekend exposure to natural light was sufficient to achieve similar to 69% of the shift in circadian timing we previously reported after a week's exposure to natural light [1]. These findings provide evidence that the human circadian clock adapts to seasonal changes in the natural light-dark cycle and is timed later in the modern environment in both winter and summer. Further, we demonstrate that earlier circadian timing can be rapidly achieved through natural light exposure during a weekend spent camping.

  • 2017. Benjamin C. Holding (et al.). Sleep 40 (11)

    Objectives: Insufficient sleep has been associated with impaired recognition of facial emotions. However, previous studies have found inconsistent results, potentially stemming from the type of static picture task used. We therefore examined whether insufficient sleep was associated with decreased emotion recognition ability in two separate studies using a dynamic multimodal task.

    Methods: Study 1 used a cross-sectional design consisting of 291 participants with questionnaire measures assessing sleep duration and self-reported sleep quality for the previous night. Study 2 used an experimental design involving 181 participants where individuals were quasi-randomized into either a sleep-deprivation (N = 90) or a sleep-control (N = 91) condition. All participants from both studies were tested on the same forced-choice multimodal test of emotion recognition to assess the accuracy of emotion categorization.

    Results: Sleep duration, self-reported sleep quality (study 1), and sleep deprivation (study 2) did not predict overall emotion recognition accuracy or speed. Similarly, the responses to each of the twelve emotions tested showed no evidence of impaired recognition ability, apart from one positive association suggesting that greater self-reported sleep quality could predict more accurate recognition of disgust (study 1).

    Conclusions: The studies presented here involve considerably larger samples than previous studies and the results support the null hypotheses. Therefore, we suggest that the ability to accurately categorize the emotions of others is not associated with short-term sleep duration or sleep quality and is resilient to acute periods of insufficient sleep.

  • 2017. Tina Sundelin (et al.). Royal Society Open Science 4 (5)

    The importance of assessing evolutionarily relevant social cues suggests that humans should be sensitive to others' sleep history, as this may indicate something about their health as well as their capacity for social interaction. Recent findings show that acute sleep deprivation and looking tired are related to decreased attractiveness and health, as perceived by others. This suggests that one might also avoid contact with sleep-deprived, or sleepy-looking, individuals, as a strategy to reduce health risk and poor interactions. In this study, 25 participants (14 females, age range 18-47 years) were photographed after 2 days of sleep restriction and after normal sleep, in a balanced design. The photographs were rated by 122 raters (65 females, age range 18-65 years) on how much they would like to socialize with the participants. They also rated participants' attractiveness, health, sleepiness and trustworthiness. The results show that raters were less inclined to socialize with individuals who had gotten insufficient sleep. Furthermore, when sleep-restricted, participants were perceived as less attractive, less healthy and more sleepy. There was no difference in perceived trustworthiness. These findings suggest that naturalistic sleep loss can be detected in a face and that people are less inclined to interact with a sleep-deprived individual.

  • 2017. Alexandre Marraffa (et al.). Physiology and Behavior 182, 27-33

    Yawning has been proposed to serve both physiological and social functions, the latter likely to have developed later in its evolution. A central hypothesis is that yawning cools the brain but whether yawning is a thermoregulatory mechanism that is activated during hyperthermia (i.e., thermoregulatory failure) or is activated in any instance of brain temperature increase (e.g., also during fever) is unclear and experimental assessments of yawning during fever are lacking. In this study, we determined the effect of experimentally induced fever on yawning frequency. We also explored alternative predictors of yawning during sickness (sleepiness, autonomic nervous system indexes and sickness symptoms). Twenty-two healthy human subjects participated in a randomized, placebo-controlled, cross-over study, where the subjects received an injection of the bacterial endotoxin lipopolysaccharide (LPS) at a dose of 2 ng/kg body weight in one condition and placebo in the other. Yawning was scored from video recordings from 30 min before to 4 h after the injection. Body temperature was measured frequently, alongside with heart rate, blood pressure, nausea and overall sickness symptoms. Yawning frequency was found to significantly increase over time during experimentally induced sickness, but not in the placebo condition. In particular, yawning frequency was increased during the rising phase of body temperature induced by LPS administration, although no significant correlation was found between body temperature increase and yawning frequency. In addition, exploratory analyses showed that a higher yawning frequency was associated with less increase in sickness symptoms and nausea intensity. While the current study adds to previous research showing significant increase in yawning frequency during hyperthermia, further studies are needed if we are to properly characterize the brain cooling role of yawning in humans. The investigation of other functions, such as being a vasovagal inhibitory, may shed stronger light on the functions of yawning.

  • 2016. Göran Kecklund, John Axelsson. BMJ. British Medical Journal 355

    This review summarises the literature on shift work and its relation to insufficient sleep, chronic diseases, and accidents. It is based on 38 meta-analyses and 24 systematic reviews, with additional narrative reviews and articles used for outlining possible mechanisms by which shift work may cause accidents and adverse health. Evidence shows that the effect of shift work on sleep mainly concerns acute sleep loss in connection with night shifts and early morning shifts. A link also exists between shift work and accidents, type 2 diabetes (relative risk range 1.09-1.40), weight gain, coronary heart disease (relative risk 1.23), stroke (relative risk 1.05), and cancer (relative risk range 1.01-1.32), although the original studies showed mixed results. The relations of shift work to cardiometabolic diseases and accidents mimic those with insufficient sleep. Laboratory studies indicate that cardiometabolic stress and cognitive impairments are increased by shift work, as well as by sleep loss. Given that the health and safety consequences of shift work and insufficient sleep are very similar, they are likely to share common mechanisms. However, additional research is needed to determine whether insufficient sleep is a causal pathway for the adverse health effects associated with shift work.

  • 2016. Sean N. Talamas (et al.). Journal of experimental psychology. General 145 (5), 603-620

    Impression formation is profoundly influenced by facial attractiveness, but the existence of facial cues which affect judgments beyond such an attractiveness halo may be underestimated. Because depression and tiredness adversely affect cognitive capacity, we reasoned that facial cues to mood (mouth curvature) and alertness (eyelid-openness) affect impressions of intellectual capacity. Over 4 studies we investigated the influence of these malleable facial cues on first impressions of intelligence. In Studies 1 and 2 we scrutinize the perceived intelligence and attractiveness ratings of images of 100 adults (aged 18-33) and 90 school-age children (aged 5-17), respectively. Intelligence impression was partially mediated by attractiveness, but independent effects of eyelid-openness and subtle smiling were found that enhanced intelligence ratings independent of attractiveness. In Study 3 we digitally manipulated stimuli to have altered eyelid-openness or mouth curvature and found that each independent manipulation had an influence on perceptions of intelligence. In a final set of stimuli (Study 4) we explored changes in these cues before and after sleep restriction, to examine whether natural variations in these cues according to sleep condition can influence perceptions. In Studies 3 and 4 variations with increased eyelid-openness and mouth curvature were found to relate positively to intelligence ratings. These findings suggest potential overgeneralizations based on subtle facial cues that indicate mood and tiredness, both of which alter cognitive ability. These findings also have important implications for students who are directly influenced by expectations of ability and teachers who may form expectations based on initial perceptions of intelligence.

  • 2016. Mats Lekander (et al.). Brain, behavior, and immunity 56, 34-41

    Task-based fMRI has been used to study the effects of experimental inflammation on the human brain, but it remains unknown whether intrinsic connectivity in the brain at rest changes during a sickness response. Here, we investigated the effect of experimental inflammation on connectivity between areas relevant for monitoring of bodily states, motivation, and subjective symptoms of sickness. In a double blind randomized controlled trial, 52 healthy volunteers were injected with 0.6 ng/kg LPS (lipopolysaccharide) or placebo, and participated in a resting state fMRI experiment after approximately 2h 45 minutes. Resting state fMRI data were available from 48 participants, of which 28 received LPS and 20 received placebo. Bilateral anterior and bilateral posterior insula sections were used as seed regions and connectivity with bilateral orbitofrontal and cingulate (anterior and middle) cortices was investigated. Back pain, headache and global sickness increased significantly after as compared to before LPS, while a non-significant trend was shown for increased nausea. Compared to placebo, LPS was followed by increased connectivity between left anterior insula and left midcingulate cortex. This connectivity was significantly correlated to increase in back pain after LPS and tended to be related to increased global sickness, but was not related to increased headache or nausea. LPS did not affect the connectivity from other insular seeds. In conclusion, the finding of increased functional connectivity between left anterior insula and middle cingulate cortex suggests a potential neurophysiological mechanism that can be further tested to understand the subjective feeling of malaise and discomfort during a sickness response.

  • 2016. Göran Kecklund, Mikael Sallinen, John Axelsson. Principles and Practice of Sleep Medicine, 742-749
  • 2016. John Axelsson, Göran Kecklund. FYSS 2017: fysisk aktivitet i sjukdomsprevention och sjukdomsbehandling, 171-183
  • 2015. John Axelsson, Vladyslav V. Vyazovskiy. Sleep 38 (12), 1843-1845
  • 2015. Tina Sundelin (et al.). Brain, behavior, and immunity 48, 53-56

    An ability to detect subtle signs of sickness in others would be highly beneficial, as it would allow for behaviors that help us avoid contagious pathogens. Recent findings suggest that both animals and humans are able to detect distinctive odor signals of individuals with activated innate immune responses. This study tested whether an innate immune response affects a person's walking speed and whether other people perceive that person as less healthy. 43 subjects watched films of persons who were experiencing experimental immune activation, and rated the walking individuals in the films with respect to health, tiredness, and sadness. Furthermore, the walking speed in the films was analyzed. After LPS injections, participants walked more slowly and were perceived as less healthy and more tired as compared to when injected with placebo. There was also a trend for the subjects to look sadder after LPS injection than after placebo. Furthermore, there were strong associations between walking speed and the appearance of health, tiredness, and sadness. These findings support the notion that walking speed is affected by an activated immune response, and that humans may be able to detect very early signs of sickness in others by merely observing their gait. This ability is likely to aid both a "behavioral immune system", by providing more opportunities for adaptive behaviors such as avoidance, and the anticipatory priming of biochemical immune responses.

  • 2014. B. Karshikoff (et al.). Brain, behavior, and immunity 46, 35-43

    Systemic inflammation can induce pain hypersensitivity in animal and human experimental models, and has been proposed to be central in clinical pain conditions. Women are overrepresented in many chronic pain conditions, but experimental studies on sex differences in pain regulation during systemic inflammation are still scarce. In two randomized and double blind placebo controlled experiments, we used low doses of lipopolysaccharide (LPS) as an experimental model of systemic inflammation. The first study employed 0.8ng/kg LPS in a within-subject design of 8 individuals (1 woman), and the second study 0.6ng/kg LPS in a between-subject design of 52 participants (29 women). We investigated the effect on (a) pressure, heat, and cold pain thresholds, (b) suprathreshold noxious heat and cold sensitivity, and (c) conditioned pain modulation (CPM), and differences between men and women. LPS induced significantly lower pressure pain thresholds as compared to placebo (mean change with the 0.8ng/kg dose being -64±30kPa P=.04; with the 0.6ng/kg dose -58±55kPa, P<.01, compared to before injection), whereas heat and cold pain thresholds remained unaffected (P's>.70). Suprathreshold noxious pain was not affected by LPS in men (P's⩾.15). However, LPS made women rated suprathreshold noxious heat stimuli as more painful (P=.01), and showed a tendency to rate noxious cold pain as more painful (P=.06) as compared to placebo. Furthermore, LPS impaired conditioned pain modulation, a measure of endogenous pain inhibition, but this effect was also restricted to women (P<.01, for men P=.27). Pain sensitivity correlated positively with plasma IL-6 and IL-8 levels. The results show that inflammation more strongly affects deep pain, rather than cutaneous pain, and suggest that women's pain perception and modulation is more sensitive to immune activation than men's.

Show all publications by John Axelsson at Stockholm University

Last updated: November 27, 2018

Bookmark and share Tell a friend