Work Description

Date: 23 March, 2023

Dataset title: The temporal structure of REM sleep shows minute-scale fluctuations across brain and body in mice and humans

Dataset creators: L.S. Bueno-Junior, M.S. Ruckstuhl, M.M. Lim, & B.O. Watson

Dataset contacts: Lezio S. Bueno-Junior soaresbl@umich.edu
Brendon O. Watson brendonw@umich.edu

Funding: National Institutes of Health (MH107662), Pritzker Neuropsychiatric Research Consortium, and University of Michigan Neuroscience Scholars Fund to B.O.W.; Portland Veterans Affairs Research Foundation to M.M.L.

Abstract:
Rapid eye movement sleep (REM) is believed to have a binary temporal structure with "phasic" and "tonic" microstates, characterized by motoric activity versus quiescence, respectively. However, we observed in mice that the frequency of theta activity (a marker of rodent REM) fluctuates in a non-binary fashion, with the extremes of that fluctuation correlating with phasic-type and tonic-type facial motricity. Thus, phasic and tonic REM may instead represent ends of a continuum. These cycles of brain physiology and facial movement occurred at 0.01-0.06 Hz, or infraslow frequencies, and affected cross-frequency coupling and neuronal activity in the neocortex, suggesting network functional impact. We then analyzed human data and observed that humans also demonstrate non-binary phasic/tonic microstates, with continuous 0.01-0.04 Hz respiratory rate cycles matching the incidence of eye movements. These fundamental properties of REM can yield new insights into our understanding of sleep health.

Significance statement:
Rapid eye movement sleep (REM) was originally identified in relation to dreaming and has various functions in neurodevelopment, synaptogenesis, and memory consolidation. REM is traditionally defined by stereotypical neurophysiology, including wake-like brain electrical activity, muscle atonia, and rapid eye movements. In this study, we identify a new temporal pattern defining REM physiology consisting of brain activity in conjunction with respiratory rate fluctuating at the very slow timescale of several seconds, with the frequency of such fluctuations matching the incidence of rapid eye movements. With these findings, we identify a new metric with which to study the timing of REM-dependent processes.

Methodology:
Methods are described in detail in the related article (see Supplementary Information Appendix). For an overview, methods are summarized as follows.

1. The mouse portion of this study involved chronic electrophysiology integrated with facial videography for recording of eye and whisker movements during REM sleep.
2. Mice were implanted with 64-channel silicon probes (one shank) perpendicularly to barrel cortex layers, with deeper channels reaching into the lateral portion of hippocampal CA.
3. Mouse implants were made with somatotopic correspondence between implanted barrels and un-trimmed whiskers, as determined by functional imaging of brain surface during whisker stimulation (prior to main recordings).
4. Mouse implants were made so that each mouse could be transferred between freely-moving home-cage recordings (electrophysiology only) and head-fixed recordings (electrophysiology combined with eye and whisker videography).
5. This study did not involve experimental designs other than just recording mice for 6-9 h continuously, with the goal of capturing spontaneously occurring REM episodes. No drugs or stimuli were delivered during the main recordings.
6. The human portion of this study involved a polysomnography (PSG) dataset with nine de-identified participants, all of which classified as healthy in terms of breathing, limb movements and medical history.
7. Data were analyzed in Matlab. See relevant code repositories below.

https://github.com/brendonw1/WatsonLabCode/tree/master/REMinfraslow (created for this study)
https://github.com/buzsakilab/buzcode/tree/dev (buzcode)

Folders contained here (ZIP format):

Mouse folders are named under the convention MOUSE_YYMMDD_HHMMSS_ followed by extra descriptions (e.g., LFP_SortedSpikes). For example, in "LB_13_190609_084409", "LB_13" is mouse, "190609" is day of recording and "084409" is recording start time.

Mouse folders containing _Videos indicate presence of eye movement and whisker movement videos from head-fixed mice. Mouse folders without videos are from home-cage recordings.

Mouse folders containing _SortedSpikes indicate presence of spike timestamps under buzcode conventions, after spike sorting using kilosort2 and phy (see article).

Mouse folders containing _LFP indicate presence of LFP and sleep scoring data under buzcode conventions.

The few mouse folders with just _LFP correspond to a mouse with no spikes and redundant LFP channels due to implantation and/or recording issues.

All human PSG data were compressed into a single ZIP file (nine participants). Per Data Use Agreement, the human dataset contained here is a reduced/derived version of the original PSG dataset. Specifically, the human data contained here are only from a subset of PSG channels (see article) and were epoched into REM, NREM and WAKE.

A preservation copy of the code repository created for this study (REMinfraslowCode). Code under development can be obtained via GitHub (see link above).

Files within folders:

Most files follow buzcode conventions for session metadata, sleep scoring and spikes, except files containing "Timestamps" (video frame timestamps), "REMints" or "WAKEints" (lists of epoch intervals in seconds after epoch trimming, see article).

Mouse REM videos (eye and whisker movements) were extracted based on initial sleep scoring and therefore contain NREM and WAKE margins. These margins were not included in the analysis of rapid movements (see article).

Certain mouse videos contain issues, such as hidden pupil. See Supplementary Information Appendix.

Human PSG data are organized in Matlab tables, with self-explanatory variable names. The tables contain cell arrays, each cell representing an epoch. Individual epochs contain either a column vector (single channel) or a two-column numerical array (two channels). When two channels are present, they represent left and right sides (electrooculography and legs) or two mastoid-referenced EEG montages (each of these from frontal, central or occipital recordings).

Related publication:
L. S. Bueno-Junior, M. S. Ruckstuhl, M. M. Lim, B. O. Watson, The temporal structure of REM sleep shows minute-scale fluctuations across brain and body in mice and humans. Proc. Natl. Acad. Sci. U. S. A. In press (2023).

Use and access:
This data set is made available under CC BY-NC 4.0

To cite data:
L. S. Bueno-Junior, M. S. Ruckstuhl, M. M. Lim, B. O. Watson, The temporal structure of REM sleep shows minute-scale fluctuations across brain and body in mice and humans [Data set]. University of Michigan - Deep Blue (2023). https://doi.org/10.7302/xxgv-ws73