Eye Movements During Sleep May Represent Head Movement in the Virtual Dream World
Post by Lani Cupo
The takeaway
In mice, researchers discovered that eye movements during sleep reflect changes in virtual head direction, suggesting that in dreaming humans’ eye movement may reflect changes in the virtual environment.
What's the science?
It is well known that dreaming occurs during the so-called Rapid Eye Movement “REM” phase of sleep, however, whether the eye movements that give the phase its name represent random muscular activity, or whether they correlate with the virtual environment of dreams is still an open question. This week in Science, Senzai and colleagues used mice to demonstrate that the activity of cells that fire, when the head is in a particular direction, correlates with eye movements during sleep.
How did they do it?
Head direction cells are a population of neurons that fire when the mouse’s head orients in a particular direction. While they can be found in different brain regions, the authors examined them in the thalamus. The authors embedded silicone probes in the thalamus to record the cellular activity while recording eye movements with head-mounted cameras. Mice were allowed to explore an arena and to fall asleep naturally (no anesthesia), with their sleep identified as REM or non-REM. First, the authors examined the relationship between internal representation of head direction (from the cellular activity) with saccade-like eye movements, where the eyes move quickly between two fixation points in the visual field. Second, the authors identified the properties of eye movement during REM sleep. Finally, they examined the relationship between internal representation of head direction and eye movements during REM sleep.
What did they find?
As expected, during wakefulness, clockwise and counterclockwise saccades correlate with head movements clockwise and counterclockwise respectively. The authors trained a statistical model only with data from the head direction cells and found that the model could accurately determine the mouse’s head direction in the environment. To their first question, the vast majority of saccades occurred simultaneously in the same direction as the head movement. Using the model they had trained on head direction cells during wakefulness, in the sleeping phase the authors decoded “virtual heading”, or the head direction indicated by the head direction cells. Changes in virtual head direction occurred with similar patterns to real turns in wakefulness. To their second question, the authors categorized saccades into two groups: “leading” eye movements, which were not preceded by any eye movement for 400 ms, or “followers”, which occurred less than 400 ms after another saccade. To their third question, the authors found that leading saccades predicted not only the direction of virtual head movements but also the amplitude, as represented by the firing cells. In contrast with leading movements, followers tended to recenter the eye, moving in the opposite direction of the leading movement. Thus, most followers tended to be opposite the direction of changes in the virtual head direction, with the exception of those with the largest amplitudes.
What's the impact?
This study provides evidence that eye movements are tightly coupled with internal representations of head direction. Therefore, eye movements during sleep are likely an external readout of internal processes. Ultimately, eye movements may allow a deeper understanding of the neurophysiological mechanisms coordinating the experience of dreaming.