A cognitive process occurring during sleep is revealed by rapid eye movements

Abstract

Since the discovery of rapid eye movement (REM) sleep, the nature of the eye movements that characterize this sleep phase has remained elusive. Do they reveal gaze shifts in the virtual environment of dreams or simply reflect random brainstem activity? We harnessed the head direction (HD) system of the mouse thalamus, a neuronal population whose activity reports, in awake mice, their actual HD as they explore their environment and, in sleeping mice, their virtual HD. We discovered that the direction and amplitude of rapid eye movements during REM sleep reveal the direction and amplitude of the ongoing changes in virtual HD. Thus, rapid eye movements disclose gaze shifts in the virtual world of REM sleep, thereby providing a window into the cognitive processes of the sleeping brain.

Fig. 1.. Saccade direction predicts internal representation of head turns.

(A) Schematic of experimental configuration illustrating a chronic electrophysiological recording from the anterodorsal nucleus of the thalamus (ADN) while eye and head movements are monitored with cameras. (B) Schematic illustration of the open field arena with a mouse carrying head-mounted eye cameras and silicon probes. The heading of the animal is monitored with a top-view camera. (C) Top left: Example frame from the top-view camera. Top right: Schematic illustration of frame on the left. Middle: Example frames showing a counterclockwise (CCW) turn and, bottom, a clockwise (CW) turn. Arrows indicate the heading of the animal in each frame. (D) The tuning curves of eleven example head direction (HD) cells recorded from the ADN of an awake mouse. The arrow on the right indicates the preferred head direction of each HD cell. Peak firing rates for HD cells are shown on the right. (E) Top: Raster plot of the firing of the eleven example HD cells shown in (D). Middle traces: Actual heading of the animal (gray) and heading decoded from the population activity of HD cells (green). Triangles mark the timing of example frames for CCW turn (cyan) and CW turn (magenta) shown in (C). Bottom traces: Horizontal position of the two eyes (red and blue). The vertical dotted lines indicate the onset of saccades. (F) Top: Two snapshots of the right eye taken before and after a CCW saccade. The pupil is delineated with a black circle. A dotted circle in the right image labels the pupil’s original position (N: nasal commissure; T: temporal commissure). The red trace illustrates the right eye’s horizontal position in time. Saccade amplitude was defined as the change in the horizontal eye position upon a saccade. The green trace illustrates the decoded heading. The head turn amplitude is the change in the decoded heading between 200ms before and 200ms after saccade onset. Note the CCW shift in decoded heading concomitant with the CCW saccade. (G) Summary scatter plot of the amplitude of conjugated saccades during the exploration in the open field arena (n=13778 from 6 mice). The amplitude and sign of the decoded head turn during the saccade is color coded. Note that the direction of the saccades matches the direction of the decoded head turn. In this and the rest of the figures the upper right and lower left quadrants of the scatter plots represent CW and CCW movements, respectively. (H) Correlation between the amplitude of conjugated saccades (averaged over both eyes) with the decoded head turn amplitude. Saccade amplitude predicted the amplitude of decoded head turns for each individual mouse (gray lines) as well as for all mice (red line, vertical lines for standard deviation).

Fig. 2.. Rapid eye movements during REM sleep

(A) Schematic of experimental configuration illustrating a chronic electrophysiological recording from the anterodorsal nucleus of the thalamus (ADN) while rapid eye movements are monitored with cameras in a sleeping mouse. (B) Snapshots of the right (top) and left eye (bottom) before (left) and after (right) a CCW rapid eye movement. The pupil is delineated with a black circle. A dotted circle in the right image labels the pupil’s original position (N: nasal commissure; T: temporal commissure). Note that both eyes move CCW. (C) Top: Example spectrogram of LFP recorded in the ADN during non-REM sleep (NREM), REM sleep and wakefulness (Wake). Middle: EMG power. Bottom: Horizontal position of the two eyes (red and blue). Note the increase in eye movements at the onset of REM sleep. (D) The shaded time interval in (C) shown on an expanded time scale. Note that for most rapid eye movements both eyes moved to the same direction. Dotted black and gray vertical lines indicate the onset of leading (not preceded by a rapid eye movement for at least 400 ms) and follower rapid eye movements. (E) Scatter plots of the amplitude of right versus left rapid eye movements during REM sleep for all mice (n=6689 from 6 mice). Note that most data points are in the lower-left or upper-right quadrants, indicating CCW and CW movements of both eyes. (F) Distribution of intervals between rapid eye movements during REM sleep for all mice (n = 6689 from 6 mice). Leading eye movements are in black and followers in gray.

Fig. 3.. Leading rapid eye movements predict decoded head turns during REM sleep

(A) Top: Schematic of experimental configuration (same as Fig. 2A) (B-F) Same mouse as in Fig. 1C–E. (B) Top: Two snapshots of the right eye during REM sleep taken before and after a CCW rapid eye movement. The red trace illustrates the right eye’s horizontal position in time. Bottom: Raster plot of the firing of three example HD cells (out of the eleven HD cells illustrated in Fig. 1D; same color code). Note the CCW shift in heading representation concomitant with the CCW rapid eye movement. (C) Four example episodes illustrating a concomitant shift in decoded heading and eye position during REM sleep. Top traces: Horizontal position of the two eyes (red and blue) and decoded heading (green). The vertical dotted lines indicate the onset of leading rapid eye movements. Bottom: Raster plot of the firing of the eleven example HD cells (same as in Fig. 1D). (D) Top: Average position of right (red) and left (blue) eye for CW (left) and CCW (right) leading eye movements. Bottom: Average decoded heading (green). Shaded areas are standard error of the mean (average of 34 traces for CW and 47 traces for CCW leading eye movements: same mouse as in B and C). (E) Scatter plot of the amplitude of leading right versus left eye movements during REM sleep. The amplitude and sign of the decoded head turns during the eye movements is color coded. Note the good match between the direction of the rapid eye movements and the direction of the decoded head turn (same mouse as in B and C). (F) Correlation between the amplitude of leading eye movements during REM sleep (averaged over both eyes) and the decoded head turn amplitude (same mouse as in B and C). (G) Top: Mean traces of horizontal eye position averaged across both eyes for CW (left) and CCW (right) leading eye movements (n = 6 mice). The dataset was separated into quintiles based on the amplitude. Bottom: Mean traces of the decoded heading in each quantile based on the amplitude of leading rapid eye movements. Note that leading rapid eye movements with larger amplitude coincide with larger decoded head turns. (H) Summary scatter plot of the amplitude of leading right versus left eye movements during REM sleep (n = 330 from 6 mice). The amplitude and sign of the decoded head turns during the eye movements is color coded. (I) Correlation between the amplitude of leading eye movements during REM sleep and the decoded head turn amplitude (n = 330 from 6 mice). Leading rapid eye movements predicted the direction and amplitude of decoded head turns in each individual mouse (gray lines) as well as for all mice (red line, vertical lines for standard deviation).

Fig. 4.. Small-amplitude follower eye movements represents recentering eye movements

(A) Top: Example traces of horizontal eye position (averaged across both eyes) aligned to the onset of first follower rapid eye movement after a CW (left) and CCW (right) leading eye movement. Middle: Average horizontal eye position (mean of both eyes) for first follower eye movements after CW (left) and CCW (right) leading eye movement (average over 132X traces for CW and 107 traces for CCW from 6 mice). Bottom: Average traces of concomitantly decoded heading. Note that, on average, first followers occur in the opposite direction as compared to the preceding leading eye movement and to the decoded head turn. Shaded area represents standard error of the mean. The time scale of the top panels covers a larger interval to include the preceding leading eye movement. (B) Correlation (heatmap) between the decoded head turn amplitude and amplitude of rapid eye movement (in bins; averaged for both eyes; y-axis). First columns from left: Leading eye movements (n = 330 from 6 mice); second column: first followers (n = 239 from 6 mice), third column: rest of the followers (n = 970 from 6 mice). Note the increased positive correlation with increasing eye movement amplitude.