Tracking Eye Movements During Sleep in Mice

Abstract

Eye movement is not only for adjusting the visual field and maintaining the stability of visual information on the retina, but also provides an external manifestation of the cognitive status of the brain. Recent studies showed similarity in eye movement patterns between wakefulness and rapid eye movement (REM) sleep, indicating that the brain status of REM sleep likely resembles that of awake status. REM sleep in humans could be divided into phasic REM and tonic REM sleep according to the difference in eye movement frequencies. Mice are the most commonly used animal model for studying neuronal and molecular mechanisms underlying sleep. However, there was a lack of details for eye movement patterns during REM sleep, hence it remains unknown whether REM sleep can be further divided into different stages in mice. Here we developed a device combining electroencephalogram (EEG), electromyogram (EMG) as well as eye movements recording in mice to study the eye movement patterns during sleep. We implanted a magnet beneath the conjunctiva of eye and tracked eye movements using a magnetic sensor. The magnetic signals showed strong correlation with video-oculography in head-fixed mice, indicating that the magnetic signals reflect the direction and magnitude of eye movement. We also found that the magnet implanted beneath the conjunctiva exhibited good biocompatibility. Finally, we examined eye movement in sleep–wake cycle, and discriminated tonic REM and phasic REM according to the frequency of eye movements, finding that compared to tonic REM, phasic REM exhibited higher oscillation power at 0.50 Hz, and lower oscillation power at 1.50–7.25 Hz and 9.50–12.00 Hz. Our device allowed to simultaneously record EEG, EMG, and eye movements during sleep and wakefulness, providing a convenient and high temporal-spatial resolution tool for studying eye movements in sleep and other researches in mice.

Pupil Dynamics-derived Sleep Stage Classification of a Head-fixed Mouse Using a Recurrent Neural Network

“…63 A recent report noted that the oscillatory population-level activity of dorsal raphe serotonin neurons gradually increased in amplitude during NREM sleep, 64 which implies a relation between deeper NREM sleep and larger PD fluctuations. During REM sleep, two microstates can be defined using eye movement frequencies 65,66 : phasic REM sleep accompanies bursts of eye movements, and tonic REM sleep has fewer eye movements. This is another example of a subdivision of a brain state suggested by pupil dynamics.…”

REM sleep has minute-scale rhythms in mice and humans: A non-binary continuum between phasic and tonic microstates

“…Figure 2B illustrates the alignment of a theta-band spectrogram and EMG tone with video-derived measures (see STAR methods). Here we focused on saccading, blinking and whisking, which exhibited spike-like events, similar to previous studies using non-videographic recording of rapid eye movements in rodents (Massot et al, 2019; Meng et al, 2021; Sánchez-López and Escudero, 2011, 2015). Relevant to intra-REM rhythms, facial movements occurred in bursts, which we captured using peak envelopes with 2 s window (Figure 2B).…”

Pupil dynamics-derived sleep stage classification of a head-fixed mouse using a recurrent neural network

“…During REM sleep, two microstates had been defined by eye movement frequencies. 64,65 Phasic REM sleep accompanied bursts of eye movements, and tonic REM sleep had less eye movement, which are another example of a subdivision of a brain state suggested by pupil dynamics. Additionally, we noticed rapid dilation of the pupil diameter (mydriasis) at the end of REM sleep when a transition from REM to WAKE states occurred.…”