Despite the fact that we spend about a third of our lives sleeping, our understanding of the physiological processes that occur during sleep remains surprisingly limited. However, although we do not know much about the physiological mechanisms, it is clear that sleep is important for our health, well-being and cognitive function. We probably all know the immediate consequences for our well-being when we have had too little sleep for a period of time, and it is also very well documented that practically all our cognitive functions deteriorate if we do not get enough sleep. Our reaction time increases, our memory and learning abilities deteriorate, our capacity to plan and manage complex situations becomes worse, our abilities to control our emotions decline, and so on.
There is also a strong link between our sleep and health. Insufficient or disrupted sleep increases the risk of conditions like cardiovascular diseases, obesity, and diabetes. Furthermore, vulnerability to a variety of neurodegenerative and psychiatric disorders increases. Although the precise causal mechanism underlying the connection between insufficient or poor sleep and these diseases remains unclear, it is indisputable that there is a link.
When we sleep, we alternate between four different sleep stages: the transition from wakefulness to sleep is called N1, light sleep is called N2, deep sleep is called N3, and finally we have REM sleep, which we also sometimes call dream sleep. During the course of a night, we alternate between the different stages of sleep in a characteristic pattern: from wakefulness we briefly go through N1 to N2, then there is a period of N3, and finally we enter REM sleep. Such a sequence is called a sleep cycle. During a night, we typically go through 4 to 6 of these sleep cycles. A graphical representation of how we alternate between the stages of sleep during a night is called a hypnogram. Figure 1 shows a typical example of a hypnogram. As evident, in practice our sleep is not quite as systematic as described above. For instance, it is quite normal to have several short awakenings throughout the night. When we analyze the physiological aspects of sleep, it involves, among other factors, a description of the hypnogram pattern.
Figure 1. A hypnogram shows the transitions between different sleep stages throughout the night. The figure shows a typical example of a hypnogram from a normal night of sleep.
Most astronauts experience poor sleep during their stay in space. However, our knowledge of their physiological sleep is quite limited. Somewhat surprisingly, there is a very low correlation between our subjective sleep experiences and our physiological sleep. In other words, the subjectively experienced sleep does not reveal much information about our physiological sleep.
Given the foundational significance of sleep for both our health and cognitive capacities, it becomes crucial to investigate if there are systematic changes in the astronauts' physiological sleep during their missions in space. Inadequate sleep, or disruptions in their sleep, can have immediate negative consequences for their cognitive functions, and thus compromise safety and the quality of the work they do. In addition, delving into their physiological sleep will allow us to assess potential long-term consequences for their health – an aspect that will become even more important if we send astronauts on extended space missions in the years to come.
The purpose of the Sleep-in-Orbit project is to investigate whether there are physiological differences between the way astronauts sleep in space and on Earth. To investigate this, the astronaut's sleep is measured before the mission to space, during the stay on the international space station (ISS), and again when the astronaut is back on Earth. Because there are relatively large variations in our sleep from night to night, it is necessary to measure a number of nights both before, during, and after the stay on the space station. Thereby, we aim to get a good picture of their sleep on the ground and in space, and thus investigate whether there are systematic differences.
The Sleep-in-Orbit experiment is an exploratory study, where the primary research question is whether there are systematic physiological differences, and secondly characterize these differences.
There is both a night-to-night variation within the individual subject, but also a variation across subjects. Therefore, to characterize the differences as accurately as possible within subjects, the aim is to measure 10 nights before, 20 nights during, and 10 nights after the stay on the international space station. Moreover, it is our goal to measure a larger number of astronauts, to investigate whether the differences found within a single astronaut generalize across astronauts. The first participant in the study is the Danish ESA astronaut Andreas Mogensen, and currently additional three other astronauts are planned.
Ear-EEG is a technique where electrical signals from the brain are measured via electrodes placed on earpieces. The Sleep-in-Orbit experiment is based on years of research and development of the ear-EEG technology, and investigation of ear-EEG based sleep monitoring in several research projects.
In the Sleep-in-Orbit project the ear-EEG earpieces are custom made to fit the individual astronaut’s ears. The earpieces are made by first taking an impression of the ear, then the ear impressions are 3D scanned, the earpieces are modeled in a CAD program, and molds for the earpieces are 3D printed. Finally, the earpieces are cast in the molds using a soft silicone material.
Figure 2 shows the earplug in the first astronaut's ear.
Usually, electrophysiological signals, such as EEG, are measured using a gel between the electrodes and the body. The gel ensures a good electrical connection between the electrodes and the body. However, in ear-EEG, for ease of use, it is preferred to measure with dry electrodes – that is, without gel. This makes it much more challenging to measure the EEG signals. To meet these challenges, we have developed a customized amplifier and datalogger unit, which are better suited than conventional biopotential amplifiers for measuring with dry electrodes. However, the Sleep-in-Orbit project necessitated a redesign of the hardware to comply with the requirements for equipment on the ISS. In particular, the numerous safety requirements proved to be quite challenging. The earpieces with electrodes and the datalogger unit are shown in Figure 3.
In normal clinical practice, sleep is scored manually by sleep clinicians who visually inspect the measurements in 30 second intervals. However, machine learning algorithms can characterize sleep on par with experienced sleep clinicians. Moreover, in previous research projects we have shown that machine learning based sleep scoring of ear-EEG is almost on par with sleep scoring based on conventional EEG measurements.
For each astronaut, the machine learning based automatic sleep staging is validated based on two nights of reference recordings, where the EEG is measured concurrently both with a conventional EEG setup and ear-EEG.
The primary advantages of ear-EEG are that it is easy to use, causes a minimum of inconvenience, and can be used while engaging in normal activities. That is why ear-EEG is well suited for sleep studies in space.
Figure 3. The Sleep-in-Orbit datalogger used for the sleep measurements on the ISS
Figure 4. Exploded view of the Sleep-in-Orbit datalogger unit.3
Figure 5. The 3 primary researchers in the project together with the first astronaut Andreas Mogensen (from left to right: Preben Kidmose, Simon Kappel, Andreas Mogensen, Kaare Mikkelsen). The picture was taken in connection with the reference measurements that were made prior to the experiment itself, where 2 nights of sleep were measured with both ear EEG and a conventional sleep monitoring setup (so-called polysomnography).
The project is funded through the Center for Ear-EEG by T&W Engineering, William Demant Foundation, and Department of Electrical and Computer Engineering, Aarhus University.
