The role of sleep in emotional processing and social functioning
Disturbed sleep is common and causes increased risks for psychiatric morbidity as well as worse subjective emotional experience. Emotional processing and regulation is important for maintaining adequate function at work and in private life. Brain mechanisms for emotional processing and social functioning is increasingly well known thanks to functional brain imaging.
Our experience of others emotional states appear to be represented hierarchically in the brain. The simplest level of representation is emotional mimicry. On a higher level, empathy is a representation where one is aware that the other s emotional state is primary. Cognitive regulation and reappraisal represents the highest level of regulation. Using functional magnetic resonance imaging (fMRI), we can identify outcomes pertaining to processing in the brain on these three levels.
This project encompasses three experimental studies with functional brain imaging. In study 1, we will investigate the effect of sleep loss on emotional processing. In study 2, we will investigate the effect of sleep loss on brain activity as well as social behaviour. In study 3, we will investigate the role of sleep quality in recovery by suppressing specific sleep stages.
The project is novel and important in that it connects sleep loss to subjective and objective measures not only of social emotions but also highly relevant measures of social behaviour.
Our experience of others emotional states appear to be represented hierarchically in the brain. The simplest level of representation is emotional mimicry. On a higher level, empathy is a representation where one is aware that the other s emotional state is primary. Cognitive regulation and reappraisal represents the highest level of regulation. Using functional magnetic resonance imaging (fMRI), we can identify outcomes pertaining to processing in the brain on these three levels.
This project encompasses three experimental studies with functional brain imaging. In study 1, we will investigate the effect of sleep loss on emotional processing. In study 2, we will investigate the effect of sleep loss on brain activity as well as social behaviour. In study 3, we will investigate the role of sleep quality in recovery by suppressing specific sleep stages.
The project is novel and important in that it connects sleep loss to subjective and objective measures not only of social emotions but also highly relevant measures of social behaviour.
Final report
Purposes, execution and development of the project
This project aimed first to conduct analyses and report findings from the first wave of the Sleepy Brain project, where participants underwent partial sleep deprivation (3 h sleep) and then completed emotion-related tasks during functional magnetic resonance imaging (fMRI). Next, the project aimed to plan, pilot test, and run a second wave of participants based on our experiences and new research questions from the first run. We have finished reporting of results from Sleepy Brain 1, then conducted 2 rounds of pilot testing with 8 and 9 participants, respectively, and a full new data collection, involving total sleep deprivation, with 38 participants. In the course of piloting, we have developed and tested new paradigms for empathy for pain.
Main results
- Sleep deprivation increased negative emotion and impaired emotional regulation
One of the strongest effects of sleep deprivation we detected was that participants reported more negative emotion and less positive emotion, e.g. on the positive and negative affect scale (PANAS). This effect was stronger that effects on reaction times and cognition. Consistent with this finding, in Sleepy Brain 1 data, participants reported less success in cognitively reappraising unpleasant images. This finding suggests a mechanism that may explain how sleep deprivation contributes to affective disorders and reduced socioemotional daytime functioning. Sleep deprivation did not cause decreased ratings of unpleasantness to empathy-related stimuli. On brain imaging outcomes, we found that cognitive reappraisal was associated with activation in bilateral orbitofrontal cortices, as expected, and empathy-related stimuli of others in pain was associated with activation in bilateral insulae and in anterior/middle cingulate cortex, also as expected. Sleep deprivation however did not affect these measures, and we hypothesized that one explanation could be the limited dose of sleep deprivation in the first wave. In the second wave, with total sleep deprivation, we have preliminarily found that sleep deprivation did reduce rated unpleasantness to empathy-related stimuli, which we interpret as a reduction in empathic responding.
- Sleep deprivation caused increased variability of the global fMRI signal
In the first wave of data, we found that sleep deprivation caused increased global signal variability, i.e. larger variation of the average signal from the gray matter in the brain. We propose that this is due to wake-state instability, i.e. global or local episodes of falling asleep during scanning, leading to changes in neural activity as well as correlated changes in respiration, heart rate, and head motion, possibly connected with wake-state instability. During piloting, we have made extensive attempts to monitor wakefulness in the scanner with electroencephalography (EEG) and electrooculography (EOG), however not reaching satisfactory signal quality. In the second wave, we preliminarily confirm increased global signal variability, and we will carefully investigate contributions of heart rate, respiratory rate, and the association to subjective sleepiness.
- Sleep deprivation caused reduced functional connectivity in intrinsic networks
Intrinsic connectivity refers to correlations between blood flow in different brain areas during rest. Typically, canonical networks can be observed using independent component analysis or similar methods. We have found results consistent with decreased connectivity in the default mode network, one such canonical network, following sleep deprivation, and decreased anticorrelation between the default mode network and the typical anticorrelated network. In an ongoing systematic review of investigations of resting state connectivity following sleep deprivation, we have found that default mode connectivity is the only metric consistently investigated (in 10 of 22 reports from 14 datasets). All ten reports show a similar direction of the effect, but only in five reports does it reach statistical significance. This suggests that many studies are underpowered. Nonetheless, taken together, these studies provide strong evidence for an effect. However, all other proposed effects of sleep deprivation on resting state connectivity have been investigated only in small numbers of studies and with inconsistent results.
Conclusions and new questions generated by the project
Our data help clarify currently open questions on the associations between sleep and emotion, suggesting that sleep deprivation causes a negativity bias, i.e. a tendency to generally interpret stimuli more negatively, and to report more negative emotions. Additionally, we have found support for inhibited emotional regulation and inhibited empathic responding following sleep deprivation.
With respect to brain connectivity, we have demonstrated that the global signal has a clear physiological correlate; thus it is not merely a nuisance variable. We have additionally demonstrated that sleep deprivation causes reduced functional connectivity, putatively through wake-state instability.
Future studies should use larger samples to obtain reliable estimates with high statistical power, possibly through collaboration with pooling of data from different centers. We have established a framework for collaboration with several Nordic and European partners and applied for funds for future work along these lines.
Dissemination
This project has resulted in a number of publications. Additional publications from wave 2 are forthcoming. All publications have been and will be published with open access, using open-access journal publication as well as parallel publishing. Additionally, we have made considerable efforts to openly publish data, code, and stimulus materials resulting from this project. Thus, we are ensuring that not only publications, but also other research outputs are openly published so that the scientific community may make use of the full value of this work. One tangible result of our data sharing is a collaboration with the research group of Uri Hasson in Italy, on fMRI correlates of eye movements, which has resulted in a preprint (no. 3 in the publication list).
This project aimed first to conduct analyses and report findings from the first wave of the Sleepy Brain project, where participants underwent partial sleep deprivation (3 h sleep) and then completed emotion-related tasks during functional magnetic resonance imaging (fMRI). Next, the project aimed to plan, pilot test, and run a second wave of participants based on our experiences and new research questions from the first run. We have finished reporting of results from Sleepy Brain 1, then conducted 2 rounds of pilot testing with 8 and 9 participants, respectively, and a full new data collection, involving total sleep deprivation, with 38 participants. In the course of piloting, we have developed and tested new paradigms for empathy for pain.
Main results
- Sleep deprivation increased negative emotion and impaired emotional regulation
One of the strongest effects of sleep deprivation we detected was that participants reported more negative emotion and less positive emotion, e.g. on the positive and negative affect scale (PANAS). This effect was stronger that effects on reaction times and cognition. Consistent with this finding, in Sleepy Brain 1 data, participants reported less success in cognitively reappraising unpleasant images. This finding suggests a mechanism that may explain how sleep deprivation contributes to affective disorders and reduced socioemotional daytime functioning. Sleep deprivation did not cause decreased ratings of unpleasantness to empathy-related stimuli. On brain imaging outcomes, we found that cognitive reappraisal was associated with activation in bilateral orbitofrontal cortices, as expected, and empathy-related stimuli of others in pain was associated with activation in bilateral insulae and in anterior/middle cingulate cortex, also as expected. Sleep deprivation however did not affect these measures, and we hypothesized that one explanation could be the limited dose of sleep deprivation in the first wave. In the second wave, with total sleep deprivation, we have preliminarily found that sleep deprivation did reduce rated unpleasantness to empathy-related stimuli, which we interpret as a reduction in empathic responding.
- Sleep deprivation caused increased variability of the global fMRI signal
In the first wave of data, we found that sleep deprivation caused increased global signal variability, i.e. larger variation of the average signal from the gray matter in the brain. We propose that this is due to wake-state instability, i.e. global or local episodes of falling asleep during scanning, leading to changes in neural activity as well as correlated changes in respiration, heart rate, and head motion, possibly connected with wake-state instability. During piloting, we have made extensive attempts to monitor wakefulness in the scanner with electroencephalography (EEG) and electrooculography (EOG), however not reaching satisfactory signal quality. In the second wave, we preliminarily confirm increased global signal variability, and we will carefully investigate contributions of heart rate, respiratory rate, and the association to subjective sleepiness.
- Sleep deprivation caused reduced functional connectivity in intrinsic networks
Intrinsic connectivity refers to correlations between blood flow in different brain areas during rest. Typically, canonical networks can be observed using independent component analysis or similar methods. We have found results consistent with decreased connectivity in the default mode network, one such canonical network, following sleep deprivation, and decreased anticorrelation between the default mode network and the typical anticorrelated network. In an ongoing systematic review of investigations of resting state connectivity following sleep deprivation, we have found that default mode connectivity is the only metric consistently investigated (in 10 of 22 reports from 14 datasets). All ten reports show a similar direction of the effect, but only in five reports does it reach statistical significance. This suggests that many studies are underpowered. Nonetheless, taken together, these studies provide strong evidence for an effect. However, all other proposed effects of sleep deprivation on resting state connectivity have been investigated only in small numbers of studies and with inconsistent results.
Conclusions and new questions generated by the project
Our data help clarify currently open questions on the associations between sleep and emotion, suggesting that sleep deprivation causes a negativity bias, i.e. a tendency to generally interpret stimuli more negatively, and to report more negative emotions. Additionally, we have found support for inhibited emotional regulation and inhibited empathic responding following sleep deprivation.
With respect to brain connectivity, we have demonstrated that the global signal has a clear physiological correlate; thus it is not merely a nuisance variable. We have additionally demonstrated that sleep deprivation causes reduced functional connectivity, putatively through wake-state instability.
Future studies should use larger samples to obtain reliable estimates with high statistical power, possibly through collaboration with pooling of data from different centers. We have established a framework for collaboration with several Nordic and European partners and applied for funds for future work along these lines.
Dissemination
This project has resulted in a number of publications. Additional publications from wave 2 are forthcoming. All publications have been and will be published with open access, using open-access journal publication as well as parallel publishing. Additionally, we have made considerable efforts to openly publish data, code, and stimulus materials resulting from this project. Thus, we are ensuring that not only publications, but also other research outputs are openly published so that the scientific community may make use of the full value of this work. One tangible result of our data sharing is a collaboration with the research group of Uri Hasson in Italy, on fMRI correlates of eye movements, which has resulted in a preprint (no. 3 in the publication list).