Arousability is the capability to wake up from sleep in response to a sensory stimulus. Arousability is a defining behavioral attribute of healthy sleep. Arousal-like events that are brief in time (≤16 s in mouse [Franken et al., 1999]; typically 3—15 s in human [Berry et al., 2018]) and that occur without an overt sensory stimulus are also part of healthy sleep (Halàsz et al., 2004; Lo et al., 2004). The sleeper is often unaware of these brief arousals, although they fragment sleep into bouts of shorter duration. When sleep is excessively fragmented, as is the case in many diverse types of sleep disorders, sleep’s restorative effects are severely compromised (Stepanski et al., 1984) and long-term health risks increase (Van Someren et al., 2015; Grandner, 2020). Estimating sleep’s vulnerability based on the frequency and types of spontaneous arousals is, therefore, essential to assess the severity of a sleep disruption.

The accurate quantification of spontaneous arousals is indeed a fundamental part in the diagnosis of widespread sleep disorders such as sleep apneas (Malhotra and Jordan, 2016). However, this task remains challenging in cases in which arousal origins are complex, such as in insomnias resulting from chronic pain or from psychiatric disorders (Feige et al., 2013; Bjurstrom and Irwin, 2016; van Someren, 2020). Spontaneous arousals seem to occur randomly (Lo et al., 2004; Dvir et al., 2018), and they can manifest as ‘partial’ because only some but not all of the physiological correlates of an arousal appear (cortical activation, heart rate increases, muscle tone), as observed in both human (Bergmann et al., 1987; Parrino et al., 2009) and mice (Franken, 2002; Léna et al., 2004; Fulda, 2011). In humans, arousals additionally occur along a continuum of heart rate changes and cortical activation (Sforza et al., 2000; Azarbarzin et al., 2014), although such graduation has not been described in rodents. Arousals can occur locally in restricted areas of cortex (Nobili et al., 2011; Castelnovo et al., 2018). Such variations impede advances in understanding abnormal arousability as a core component of sleep disruptions in both humans and rodent models. In patients, arousals may go undetected and leave sleep fragmentation underestimated. Supporting this possibility, patients complaining about sleep while their polysomnographic measures appear normal remain a subject of intense study (Bjurstrom and Irwin, 2016; Lecci et al., 2020; van Someren, 2020). In rodents, the diversity of spontaneous arousals is poorly studied, which limits the usefulness of animal models of sleep disorders.

The state of non-rapid eye movement sleep (NREMS) is commonly considered a restorative state with an overall low sensory arousability, but also a dynamic state in which various global measures of neuronal excitability and functional connectivity vary on multiple time scales. Of particular interest for arousability is the infraslow time scale that relates to intervals of tens of seconds (Drew et al., 2008; Watson, 2018). During sleep, the infraslow time scale characterizes variations of brain excitability in epileptic patients (Vanhatalo et al., 2004), the clustering of sleep rhythms called sleep spindles in rodents and humans (Lecci et al., 2017; Lázár et al., 2019), the time course of the human cyclic alternating pattern (Parrino et al., 2012), and the dynamics of resting-state networks obtained in functional resonance imaging studies (Fukunaga et al., 2006; Hiltunen et al., 2014). Infraslow fluctuations in brain excitability and connectivity could render the sleeping brain vulnerable to spontaneous disruptive events, including more frequent arousals in case of sleep disorders.

This study addresses both these aspects in mice. First, we demonstrate in healthy C57Bl/6J mice that infraslow fluctuations determine when spontaneous arousals preferentially occur. Next, we evaluate the case of neuropathic pain, a clinical condition often intractable and of high sensory discomfort in which sleep disruptions are frequent (Bjurstrom and Irwin, 2016; Mathias et al., 2018). Experimental models of peripheral neuropathic pain rely on surgically crushing or lesioning peripheral nerves, such as the sciatic nerve (Jaggi et al., 2011; Challa, 2015). In spite of their limitations, these models have allowed numerous breakthrough understandings on chronic pain mechanisms (Kuner and Kuner, 2020) and avenues for treatment (Finnerup et al., 2021). In contrast, sleep studies on chronic pain models are scarce and have produced variable results (Andersen and Tufik, 2003; Kontinen et al., 2003; Tokunaga et al., 2007; Cardoso-Cruz et al., 2011; Leys et al., 2013).

We find that the animals with spared nerve injury (SNI) in early stages of chronic pain do sleep normally in terms of sleep macrostructure, although physiological signatures of painful sensations persist in both the waking and the sleeping brain. Using infraslow measures of NREMS vulnerability, in particular the clustering of sleep spindles (Lecci et al., 2017), we find that sleep is disrupted in two previously undescribed ways. First, there are more frequent cortico-autonomic local arousals in pain-processing cortical areas while, polysomnographically, sleep remains continuous. Moreover, animals show a higher vulnerability to wake up from NREMS in response to fine vibrational stimuli. The essence of the sleep disturbances in this experimental chronic pain model thus lies in an abnormal arousability, in both its spontaneous and evoked forms.

Chronic pain is a widespread and complex condition compromising sleep. As poor sleep further aggravates pain, and sleep disturbances impact quality of life, therapeutic approaches to target the sites of interaction between sleep and pain are of particular interest. Here, in efforts to tease apart pain-sleep associations, we focused on abnormal arousability at moments when NREMS is most vulnerable. This study progresses on the sleep-pain association in four essential ways. First, we show that brain and autonomic signatures of SNI during the day intrude in a persistent manner into sleep. Second, sleep appeared nevertheless preserved in architecture, dominant spectral band power, and homeostatic regulation. Third, a previously undescribed spontaneous arousal during NREMS, showing local cortical activation with concomitant heart rate increases, appeared more frequently after SNI. Fourth, we also demonstrate that mechanovibrational stimuli of the body triggered brief wake-ups from NREMS more easily in the SNI group of mice. In summary, an experimental model of chronic pain impacts on NREMS in terms of arousability, more specifically on the probability that NREMS transits towards diverse levels of wakefulness, either spontaneously, or with external stimuli.

Accumulating evidence indicates that infraslow time scales relate to variations in sensory arousability (Lecci et al., 2017; Yüzgeç et al., 2018). We found here that spontaneous MAs without obvious sensory input occur clustered at moments when sensory arousability is high (Lecci et al., 2017). Therefore, when mouse NREMS is vulnerable to sensory stimuli, it is also vulnerable to spontaneous arousals. Sensory arousals and MAs may have common mechanistic origins, in line with evidence that both depend on activity in wake-promoting brain areas (Léna et al., 2004; Huang et al., 2006; Dvir et al., 2018; Mátyás et al., 2018). Furthermore, the result suggests that infraslow time scales determine the vulnerability of NREMS for both weak (in case of spontaneous arousals) and more intense (in case of evoked arousals) levels of wake-promoting input. The increased delta power and decreased AI in S1HL in fragility periods continuing into NREMS seem at odds with this finding because they signal momentarily increased sleep depth. These dynamics are likely a consequence of an incompatibility between sigma and delta power in sensory thalamocortical loops (Fernandez et al., 2018) that are not, however, strong enough to override the vulnerability of NREMS identified here. Intriguingly similar dynamics are also observed for CAP, in which periodic increases in delta alone are part of the arousal pattern (Parrino et al., 2012). This motivates studies testing the usefulness of fragility periods as time raster in the search for diverse, even subtle, arousal-like events that could be relevant to model pathological conditions of human patients.

The major portion of this study aimed to obtain a proof of concept for this possibility. The SNI model appeared particularly appropriate because it presents a specific pain hypersensitivity that could be investigated with respect to its impact on both, abnormal spontaneous and evoked arousability, during NREMS. The availability of fragility periods as time raster guided the identification of arousals that occurred locally while NREMS continued uninterrupted and that were more pronounced and more frequent in SNI. Without the raster provided by the fragility periods, the phasic differences in AI amidst the tonically elevated high-frequency power would easily have gone undetected. To further ascertain that these detected events constituted true arousals rather than accidental spectral fluctuations, we sought for independent physiological correlates. Heart rate increases were consistently higher when calculated for fragility periods with AI peak than for the ones without AI peak. Moreover, their distribution across the resting phase was similar to the one found for MAs and they were present in both S1HL and PrL. This result supports our interpretation that we have identified here a novel cortical-autonomic arousal subject to similar time-of-day-dependent regulatory mechanisms as conventional MAs. It also concurs with human NREMS in which heart rate increases occur even for very weak cortical arousals (Azarbarzin et al., 2014). Still, we cannot exclude that arousal subtypes outside the fragility periods went undetected that would require further characterization. We also remark here clearly that, aside from more frequent cortico-autonomic arousals, the SNI group of mice suffered from a tonically elevated high-frequency power in S1HL that likely underlay the more elevated sensory arousability throughout continuity and fragility periods.

We found that the cortical regions in which arousals appeared coincided with sites known to be affected by neuropathic pain. The continued presence of high-frequency activity in S1HL is reminiscent of the cortical oscillatory activity present during withdrawal from acute painful stimuli (Tan et al., 2019), suggesting that, after SNI, nociceptive input continues to arrive in cortex during NREMS to generate excessive excitation. Indeed, it has been suggested that the SNI model does show spontaneous ectopic electrical activity in peripheral sensory neurons as a result of nerve injury that can increase nociceptive inputs (Wall and Devor, 1983; Devor, 2009). NREMS is thought to protect relatively weakly from nociceptive inputs (Claude et al., 2015), therefore possibly allowing continued processing of spontaneous nociceptive activity that could explain the cortical spectral changes we detected. Besides peripheral nociceptive activity barrages, numerous changes within pain-processing pathways could be involved. We mention that optogenetic stimulation of the thalamic reticular nucleus (TRN) alleviates SNI-related pain (LeBlanc et al., 2017). As TRN is involved in the regulation of low-frequency power during NREMS (Fernandez et al., 2018), the attenuated delta dynamics observed in NREMS after SNI could result also from altered activity in this nucleus. Pain induces maladaptive plasticity processes with large alteration of brain circuitry, and it is clear further cellular studies will be necessary to understand the origins of elevated arousal frequency in NREMS after SNI before pointing more selectively to one or another mechanism.

The lack of major sleep disruptions after SNI was initially unexpected but seemed in line with other studies. First, we analyzed these animals at a time point when pain from the incision and associated inflammations are largely over (Guida et al., 2020), both of which have been shown to disrupt sleep (Landis et al., 1989; Silva et al., 2008; Leys et al., 2013). Moreover, the animals showed a preserved time spent in REMS, suggesting that they did not suffer from chronic mild stress-inflicted sleep disruptions (Nollet et al., 2019). However, we did find alterations in a broad range of frequencies including theta power in quiet wakefulness, as reported for other animal models of chronic pain (LeBlanc et al., 2014) or for humans with severe neurogenic pain (Sarnthein et al., 2006; Stern et al., 2006; Ploner et al., 2017). Other studies on experimental models of neuropathic pain also report diverse moderate effects on sleep (Kontinen et al., 2003; Leys et al., 2013). One study on rats 2 and 10 days after SNI suggested that brain states intermediate between NREMS and wakefulness during the resting phase exist (Cardoso-Cruz et al., 2011), which could in part reflect our observations. Analysis of sleep disruptions at later stages will help decide whether distinct phases of sleep disruptions mark distinct phases of pain chronicity when anxiety- and depression-related behaviors appear more strongly (Guida et al., 2020).

Do animals with SNI suffer from disturbed sleep? Our measures of NREMS’s spectral composition point to regionally restricted but tonic imbalances in the contribution of low vs. higher frequencies. Patients with insomnia show such imbalances over widespread brain regions that include sensorimotor areas (Lecci et al., 2020). Furthermore, higher power in the beta frequencies has been related to the patients remaining hypervigilant or excessively ruminating at sleep onset (Perlis et al., 2001), preventing the deactivation of cortical processes required for the loss of consciousness. Although insomnia also needs subjective assessments that are not possible in animals, this comparison suggests that SNI might suffer from similar experiences due to the tonically enhanced high-frequency oscillations. This interpretation is supported by the elevated wake-up rates in response to mild vibrational stimuli throughout the infraslow cycles, suggesting hyperalertness to environmental disturbance. On top of these tonic changes, there were more frequent cortico-autonomic arousals. Although these do not seem to elevate daytime sleepiness based on the mostly unchanged delta power dynamics across time of day, frequent increases in heart rate during the night could augment cardiovascular risk in the long term (Silvani, 2019). To further analyze the animal’s conditions during daytime, tests on their cognitive abilities in memory-dependent tasks while locally manipulating sleep in the affected hindlimb area could be considered. Deficits in working and declarative memories in rodents with SNI have been documented from early periods of chronic pain (Guida et al., 2020). Chemogenetic manipulation of neuronal populations proposed to be responsible for the gamma activity in chronic pain, restricted to sleep periods (Tan et al., 2019), seems a feasible approach to specifically suppress abnormal pain-related activity during sleep while testing performance in such tasks during wakefulness.

We provide here a guideline to identify arousals in mouse NREMS for which we see the potential to move forward both, animal models of sleep disorders, and advances in the understanding of human sleep disorders. Key for this was the recognition that infraslow electrical variations in the sleeping brain serve as an index for the vulnerability of NREMS to undergo arousal. SNI in mouse qualifies as a rodent model replicating physiological correlates of sleep disorders that are hidden behind a comparatively normal sleep architecture, which is of interest in the field of insomnia (Lecci et al., 2020; van Someren, 2020). The location of these arousals within the primary cortical sights of pain makes it obvious that the cortical pain axis is causal to sleep disruptions. It could thus be of interest to target these sites specifically to attenuate abnormal activity and to normalize sleep. Addressing the localization and the intensity of local arousals in chronic pain patients could be a first step in the definition of disease subtypes and the severity of their sleep disruption (Babiloni et al., 2021). Non-invasive brain stimulation techniques applied locally over sensory or motor cortices have been proposed as therapeutic strategy (Lefaucheur et al., 2020). The insights gained from this rodent model of neuropathic pain could thus refine diagnosis and management strategies in the human condition, in which more frequent awakenings are amongst the most consistent effects reported.

All processed data generated or analyzed during this study are included in the manuscript and supporting files. Source data files are provided for all figures. Matlab codes for major analyses are provided.

1. 1. Alexandre C
   2. Latremoliere A
   3. Ferreira A
   4. Miracca G
   5. Yamamoto M
   6. Scammell TE
   7. Woolf CJ

   (2017) Decreased alertness due to sleep loss increases pain sensitivity in mice

   Nat Med 23:768–774.

   https://doi.org/10.1038/nm.4329

   • PubMed
   • Google Scholar
2. 1. Andersen ML
   2. Tufik S

   (2003) Sleep patterns over 21-day period in rats with chronic constriction of sciatic nerve

   Brain Res 984:84–92.

   https://doi.org/10.1016/s0006-8993(03)03095-6

   • PubMed
   • Google Scholar
3. 1. Azarbarzin A
   2. Ostrowski M
   3. Hanly P
   4. Younes M

   (2014) Relationship between arousal intensity and heart rate response to arousal

   Sleep 37:645–653.

   https://doi.org/10.5665/sleep.3560

   • PubMed
   • Google Scholar
4. 1. Babiloni AH
   2. Beetz G
   3. Tang NKY
   4. Heinzer R
   5. Nijs J
   6. Martel MO
   7. Lavigne GJ

   (2021) Towards the endotyping of the sleep-pain interaction: a topical review on multitarget strategies based on phenotypic vulnerabilities and putative pathways

   Pain 162:1281–1288.

   https://doi.org/10.1097/j.pain.0000000000002124

   • PubMed
   • Google Scholar
5. 1. Berens P

   (2009) CircStat: A MATLAB toolbox for circular statistics

   J Stat Software 31:1–21.

   https://doi.org/10.18637/jss.v031.i10

   • Google Scholar
6. 1. Bergmann BM
   2. Winter JB
   3. Rosenberg RS
   4. Rechtschaffen A

   (1987) NREM sleep with low-voltage EEG in the rat

   Sleep 10:1–11.

   https://doi.org/10.1093/sleep/10.1.1

   • Google Scholar
7. Book

   1. Berry R
   2. Albertario CL
   3. Harding SM
   4. LLoyd RM
   5. Plante DT
   6. Quan SF
   7. Troester MM
   8. Vaughn B

   (2018)

   The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications

   Darien, IL: American Academy of Sleep Medicine.

   • Google Scholar
8. 1. Bjurstrom MF
   2. Irwin MR

   (2016) Polysomnographic characteristics in nonmalignant chronic pain populations: A review of controlled studies

   Sleep Med Rev 26:74–86.

   https://doi.org/10.1016/j.smrv.2015.03.004

   • PubMed
   • Google Scholar
9. 1. Bourquin AF
   2. Süveges M
   3. Pertin M
   4. Gilliard N
   5. Sardy S
   6. Davison AC
   7. Spahn DR
   8. Décosterd I

   (2006) Assessment and analysis of mechanical allodynia-like behavior induced by spared nerve injury (SNI) in the mouse

   Pain 122:e11–e14.

   https://doi.org/10.1016/j.pain.2005.10.036

   • Google Scholar
10. 1. Cardoso-Cruz H
    2. Sameshima K
    3. Lima D
    4. Galhardo V

    (2011) Dynamics of circadian thalamocortical flow of information during a peripheral neuropathic pain condition

    Frontiers in Integrative Neuroscience 5:43.

    https://doi.org/10.3389/fnint.2011.00043

    • PubMed
    • Google Scholar
11. 1. Castelnovo A
    2. Lopez R
    3. Proserpio P
    4. Nobili L
    5. Dauvilliers Y

    (2018) NREM sleep parasomnias as disorders of sleep-state dissociation

    Nat Rev Neurol 14:470–481.

    https://doi.org/10.1038/s41582-018-0030-y

    • PubMed
    • Google Scholar
12. 1. Challa SR

    (2015) Surgical animal models of neuropathic pain: Pros and Cons

    Int J Neurosci 125:170–174.

    https://doi.org/10.3109/00207454.2014.922559

    • PubMed
    • Google Scholar
13. 1. Claude L
    2. Chouchou F
    3. Prados G
    4. Castro M
    5. De Blay B
    6. Perchet C
    7. Garcia-Larrea L
    8. Mazza S
    9. Bastuji H

    (2015) Sleep spindles and human cortical nociception: a surface and intracerebral electrophysiological study

    J Physiol 593:4995–5008.

    https://doi.org/10.1113/JP270941

    • PubMed
    • Google Scholar
14. 1. Craig AD

    (2003) A new view of pain as a homeostatic emotion

    Trends Neurosci 26:303–307.

    https://doi.org/10.1016/s0166-2236(03)00123-1

    • PubMed
    • Google Scholar
15. 1. Decosterd I
    2. Woolf CJ

    (2000) Spared nerve injury: an animal model of persistent peripheral neuropathic pain

    Pain 87:149–158.

    https://doi.org/10.1016/s0304-3959(00)00276-1

    • PubMed
    • Google Scholar
16. 1. Devor M

    (2009) Ectopic discharge in Aβ afferents as a source of neuropathic pain

    Exp Brain Res 196:115–128.

    https://doi.org/10.1007/s00221-009-1724-6

    • PubMed
    • Google Scholar
17. 1. Drew PJ
    2. Duyn JH
    3. Golanov E
    4. Kleinfeld D

    (2008) Finding coherence in spontaneous oscillations

    Nat Neurosci 11:991–993.

    https://doi.org/10.1038/nn0908-991

    • PubMed
    • Google Scholar
18. 1. Dvir H
    2. Elbaz I
    3. Havlin S
    4. Appelbaum L
    5. Ivanov PC
    6. Bartsch RP

    (2018) Neuronal noise as an origin of sleep arousals and its role in sudden infant death syndrome

    Sci Adv 4:eaar6277.

    https://doi.org/10.1126/sciadv.aar6277

    • PubMed
    • Google Scholar
19. 1. Feige B
    2. Baglioni C
    3. Spiegelhalder K
    4. Hirscher V
    5. Nissen C
    6. Riemann D

    (2013) The microstructure of sleep in primary insomnia: an overview and extension

    Int J Psychophysiol 89:171–180.

    https://doi.org/10.1016/j.ijpsycho.2013.04.002

    • PubMed
    • Google Scholar
20. 1. Fernandez LMJ
    2. Lecci S
    3. Cardis R
    4. Vantomme G
    5. Béard E
    6. Lüthi A

    (2017) Quantifying infra-slow dynamics of spectral power and heart rate in sleeping mice

    J Vis Exp 126:e55863.

    https://doi.org/10.3791/55863

    • PubMed
    • Google Scholar
21. 1. Fernandez LM
    2. Vantomme G
    3. Osorio-Forero A
    4. Cardis R
    5. Béard E
    6. Lüthi A

    (2018) Thalamic reticular control of local sleep in sensory cortex

    eLife 7:e39111.

    https://doi.org/10.7554/eLife.39111

    • PubMed
    • Google Scholar
22. 1. Fernandez LMJ
    2. Lüthi A

    (2020) Sleep spindles: mechanisms and functions

    Physiological Reviews 100:805–868.

    https://doi.org/10.1152/physrev.00042.2018

    • PubMed
    • Google Scholar
23. 1. Finnerup NB
    2. Kuner R
    3. Jensen TS

    (2021) Neuropathic pain: from mechanisms to treatment

    Physiol Rev 101:259–301.

    https://doi.org/10.1152/physrev.00045.2019

    • PubMed
    • Google Scholar
24. 1. Franken P
    2. Malafosse A
    3. Tafti M

    (1999)

    Genetic determinants of sleep regulation in inbred mice

    Sleep 22:155–169.

    • PubMed
    • Google Scholar
25. 1. Franken P

    (2002) Long-term vs. short-term processes regulating REM sleep

    J Sleep Res 11:17–28.

    https://doi.org/10.1046/j.1365-2869.2002.00275.x

    • PubMed
    • Google Scholar
26. 1. Fukunaga M
    2. Horovitz SG
    3. van Gelderen P
    4. de Zwart JA
    5. Jansma JM
    6. Ikonomidou VN
    7. Chu R
    8. Deckers RH
    9. Leopold DA
    10. Duyn JH

    (2006) Large-amplitude, spatially correlated fluctuations in BOLD fMRI signals during extended rest and early sleep stages

    Magn Reson Imaging 24:979–992.

    https://doi.org/10.1016/j.mri.2006.04.018

    • PubMed
    • Google Scholar
27. 1. Fulda S

    (2011) Idiopathic REM sleep behavior disorder as a long-term predictor of neurodegenerative disorders

    EPMA J 2:451–458.

    https://doi.org/10.1007/s13167-011-0096-8

    • PubMed
    • Google Scholar
28. 1. Grandner MA

    (2020) Sleep, Health, and Society

    Sleep Med Clin 15:319–340.

    https://doi.org/10.1016/j.jsmc.2020.02.017

    • PubMed
    • Google Scholar
29. 1. Guida F
    2. De Gregorio D
    3. Palazzo E
    4. Ricciardi F
    5. Boccella S
    6. Belardo C
    7. Iannotta M
    8. Infantino R
    9. Formato F
    10. Marabese I
    11. Luongo L
    12. Novellis de
    13. Maione S

    (2020) Behavioral, biochemical and electrophysiological changes in spared nerve injury model of neuropathic pain

    Int J Mol Sci 21:3396.

    https://doi.org/10.3390/ijms21093396

    • PubMed
    • Google Scholar
30. 1. Halàsz P
    2. Terzano M
    3. Parrino L
    4. Bódizs R

    (2004) The nature of arousal in sleep

    J Sleep Res 13:1–23.

    https://doi.org/10.1111/j.1365-2869.2004.00388.x

    • PubMed
    • Google Scholar
31. 1. Hiltunen T
    2. Kantola J
    3. Abou Elseoud A
    4. Lepola P
    5. Suominen K
    6. Starck T
    7. Nikkinen J
    8. Remes J
    9. Tervonen O
    10. Palva S
    11. Kiviniemi V
    12. Palva JM

    (2014) Infra-slow EEG fluctuations are correlated with resting-state network dynamics in fMRI

    J Neurosci 34:356–362.

    https://doi.org/10.1523/JNEUROSCI.0276-13.2014

    • PubMed
    • Google Scholar
32. 1. Huang ZL
    2. Mochizuki T
    3. Qu WM
    4. Hong ZY
    5. Watanabe T
    6. Urade Y
    7. Hayaishi O

    (2006) Altered sleep-wake characteristics and lack of arousal response to H3 receptor antagonist in histamine H1 receptor knockout mice

    PNAS 103:4687–4692.

    https://doi.org/10.1073/pnas.0600451103

    • PubMed
    • Google Scholar
33. 1. Jaggi AS
    2. Jain V
    3. Singh N

    (2011) Animal models of neuropathic pain

    Fundam Clin Pharmacol 25:1–28.

    https://doi.org/10.1111/j.1472-8206.2009.00801.x

    • Google Scholar
34. 1. Kontinen VK
    2. Ahnaou A
    3. Drinkenburg WH
    4. Meert TF

    (2003) Sleep and EEG patterns in the chronic constriction injury model of neuropathic pain

    Physiol Behav 78:241–246.

    https://doi.org/10.1016/s0031-9384(02)00966-6

    • PubMed
    • Google Scholar
35. 1. Kopp C
    2. Longordo F
    3. Nicholson JR
    4. Lüthi A

    (2006) Insufficient sleep reversibly alters bidirectional synaptic plasticity and NMDA receptor function

    J Neurosci 26:12456–12465.

    https://doi.org/10.1523/JNEUROSCI.2702-06.2006

    • PubMed
    • Google Scholar
36. 1. Kuner R
    2. Kuner T

    (2020) Cellular circuits in the brain and their modulation in acute and chronic pain

    Physiol Rev 101:213–258.

    https://doi.org/10.1152/physrev.00040.2019

    • PubMed
    • Google Scholar
37. 1. Landis CA
    2. Levine JD
    3. Robinson CR

    (1989) Decreased slow-wave and paradoxical sleep in a rat chronic pain model

    Sleep 12:167–177.

    https://doi.org/10.1093/sleep/12.2.167

    • PubMed
    • Google Scholar
38. 1. Lázár Z
    2. Dijk DJ
    3. Lázár AS

    (2019) Infraslow oscillations in human sleep spindle activity

    J Neurosci Methods 316:22–34.

    https://doi.org/10.1016/j.jneumeth.2018.12.002

    • PubMed
    • Google Scholar
39. 1. LeBlanc BW
    2. Lii TR
    3. Silverman AE
    4. Alleyne RT
    5. Saab CY

    (2014) Cortical theta is increased while thalamocortical coherence is decreased in rat models of acute and chronic pain

    Pain 155:773–782.

    https://doi.org/10.1016/j.pain.2014.01.013

    • PubMed
    • Google Scholar
40. 1. LeBlanc BW
    2. Bowary PM
    3. Chao YC
    4. Lii TR
    5. Saab CY

    (2016) Electroencephalographic signatures of pain and analgesia in rats

    Pain 157:2330–2340.

    https://doi.org/10.1097/j.pain.0000000000000652

    • PubMed
    • Google Scholar
41. 1. LeBlanc BW
    2. Cross B
    3. Smith KA
    4. Roach C
    5. Xia J
    6. Chao YC
    7. Levitt J
    8. Koyama S
    9. Moore C
    10. Saab CY

    (2017) Thalamic bursts down-regulate cortical theta and nociceptive behavior

    Sci Rep 7:2482.

    https://doi.org/10.1038/s41598-017-02753-6

    • PubMed
    • Google Scholar
42. 1. Lecci S
    2. Fernandez LM
    3. Weber FD
    4. Cardis R
    5. Chatton JY
    6. Born J
    7. Lüthi A

    (2017) Coordinated infraslow neural and cardiac oscillations mark fragility and offline periods in mammalian sleep

    Sci Adv 3:e1602026.

    https://doi.org/10.1126/sciadv.1602026

    • PubMed
    • Google Scholar
43. 1. Lecci S
    2. Cataldi J
    3. Betta M
    4. Bernardi G
    5. Heinzer R
    6. Siclari F

    (2020) EEG changes associated with subjective under- and overestimation of sleep duration

    Sleep 43:zsaa094.

    https://doi.org/10.1093/sleep/zsaa094

    • PubMed
    • Google Scholar
44. 1. Lefaucheur JP
    2. Aleman A
    3. Baeken C
    4. Benninger DH
    5. Brunelin J
    6. Lazzaro D
    7. Filipovic SR
    8. Grefkes C
    9. Hasan A
    10. Hummel FC
    11. Jaaskelainen SK
    12. Langguth B
    13. Leocani L
    14. Londero A
    15. Nardone R
    16. Nguyen JP
    17. Nyffeler T
    18. Oliveira-Maia AJ
    19. Oliviero A
    20. Padberg F
    21. Palm U
    22. Paulus W
    23. Poulet E
    24. Quartarone A
    25. Rachid F
    26. Rektorova I
    27. Rossi S
    28. Sahlsten H
    29. Schecklmann M
    30. Szekely D
    31. Ziemann U

    (2020) Evidence-based guidelines on the therapeutic use of repetitive transcranial magnetic stimulation (rTMS): An update (2014-2018

    Clin Neurophysiol 131:474–528.

    https://doi.org/10.1016/j.clinph.2019.11.002

    • PubMed
    • Google Scholar
45. 1. Léna C
    2. Popa D
    3. Grailhe R
    4. Escourrou P
    5. Changeux JP
    6. Adrien J

    (2004) β2-containing nicotinic receptors contribute to the organization of sleep and regulate putative micro-arousals in mice

    J Neurosci 24:5711–5718.

    https://doi.org/10.1523/JNEUROSCI.3882-03.2004

    • PubMed
    • Google Scholar
46. 1. Leys LJ
    2. Chu KL
    3. Xu J
    4. Pai M
    5. Yang HS
    6. Robb HM
    7. Jarvis MF
    8. Radek RJ
    9. McGaraughty S

    (2013) Disturbances in slow-wave sleep are induced by models of bilateral inflammation, neuropathic, and postoperative pain, but not osteoarthritic pain in rats

    Pain 154:1092–1102.

    https://doi.org/10.1016/j.pain.2013.03.019

    • PubMed
    • Google Scholar
47. 1. Lo CC
    2. Chou T
    3. Penzel T
    4. Scammell TE
    5. Strecker RE
    6. Stanley HE
    7. Ivanov P

    (2004) Common scale-invariant patterns of sleep-wake transitions across mammalian species

    PNAS 101:17545–17548.

    https://doi.org/10.1073/pnas.0408242101

    • PubMed
    • Google Scholar
48. 1. Malhotra A
    2. Jordan A

    (2016) The importance of arousal in obstructive sleep apnea-updates from the American thoracic society 2016

    Journal of Thoracic Disease 8:S542–S544.

    https://doi.org/10.21037/jtd.2016.06.81

    • PubMed
    • Google Scholar
49. 1. Mang GM
    2. Franken P

    (2012) Sleep and EEG phenotyping in mice

    Curr Protoc Mouse Biol 2:55–74.

    https://doi.org/10.1002/9780470942390.mo110126

    • PubMed
    • Google Scholar
50. 1. Mathias JL
    2. Cant ML
    3. Burke ALJ

    (2018) Sleep disturbances and sleep disorders in adults living with chronic pain: a meta-analysis

    Sleep Med 52:198–210.

    https://doi.org/10.1016/j.sleep.2018.05.023

    • PubMed
    • Google Scholar
51. 1. Mátyás F
    2. Komlósi G
    3. Babiczky Á
    4. Kocsis K
    5. Barthó P
    6. Barsy B
    7. Dávid C
    8. Kanti V
    9. Porrero C
    10. Magyar A
    11. Szücs I
    12. Clasca F
    13. Acsády L

    (2018) A highly collateralized thalamic cell type with arousal-predicting activity serves as a key hub for graded state transitions in the forebrain

    Nat Neurosci 21:1551–1562.

    https://doi.org/10.1038/s41593-018-0251-9

    • PubMed
    • Google Scholar
52. 1. Moisset X
    2. Bouhassira D

    (2007) Brain imaging of neuropathic pain

    NeuroImage 37:S80–S88.

    https://doi.org/10.1016/j.neuroimage.2007.03.054

    • Google Scholar
53. 1. Nobili L
    2. Ferrara M
    3. Moroni F
    4. De Gennaro L
    5. Russo GL
    6. Campus C
    7. Cardinale F
    8. De Carli F

    (2011) Dissociated wake-like and sleep-like electro-cortical activity during sleep

    Neuroimage 58:612–619.

    https://doi.org/10.1016/j.neuroimage.2011.06.032

    • PubMed
    • Google Scholar
54. 1. Nollet M
    2. Hicks H
    3. McCarthy AP
    4. Wu H
    5. Moller-Levet CS
    6. Laing EE
    7. Malki K
    8. Lawless N
    9. Wafford KA
    10. Dijk DJ
    11. Winsky-Sommerer R

    (2019) REM sleep’s unique associations with corticosterone regulation, apoptotic pathways, and behavior in chronic stress in mice

    PNAS 116:2733–2742.

    https://doi.org/10.1073/pnas.1816456116

    • PubMed
    • Google Scholar
55. 1. Parrino L
    2. Milioli G
    3. De Paolis F
    4. Grassi A
    5. Terzano MG

    (2009) Paradoxical insomnia: the role of CAP and arousals in sleep misperception

    Sleep Med 10:1139–1145.

    https://doi.org/10.1016/j.sleep.2008.12.014

    • PubMed
    • Google Scholar
56. 1. Parrino L
    2. Ferri R
    3. Bruni O
    4. Terzano MG

    (2012) Cyclic alternating pattern (CAP): the marker of sleep instability

    Sleep Med Rev 16:27–45.

    https://doi.org/10.1016/j.smrv.2011.02.003

    • PubMed
    • Google Scholar
57. 1. Perlis ML
    2. Merica H
    3. Smith MT
    4. Giles DE

    (2001) Beta EEG activity and insomnia

    Sleep Med Rev 5:363–374.

    https://doi.org/10.1053/smrv.2001.0151

    • Google Scholar
58. 1. Ploner M
    2. Sorg C
    3. Gross J

    (2017) Brain rhythms of pain

    Trends Cogn Sci 21:100–110.

    https://doi.org/10.1016/j.tics.2016.12.001

    • PubMed
    • Google Scholar
59. 1. Price DD

    (2000) Psychological and neural mechanisms of the affective dimension of pain

    Science 288:1769–1772.

    https://doi.org/10.1126/science.288.5472.1769

    • PubMed
    • Google Scholar
60. 1. Sarnthein J
    2. Stern J
    3. Aufenberg C
    4. Rousson V
    5. Jeanmonod D

    (2006) Increased EEG power and slowed dominant frequency in patients with neurogenic pain

    Brain 129:55–64.

    https://doi.org/10.1093/brain/awh631

    • PubMed
    • Google Scholar
61. 1. Sforza E
    2. Jouny C
    3. Ibanez V

    (2000) Cardiac activation during arousal in humans: further evidence for hierarchy in the arousal response

    Clin Neurophysiol 111:1611–1619.

    https://doi.org/10.1016/s1388-2457(00)00363-1

    • PubMed
    • Google Scholar
62. 1. Silva A
    2. Andersen ML
    3. Tufik S

    (2008) Sleep pattern in an experimental model of osteoarthritis

    Pain 140:446–455.

    https://doi.org/10.1016/j.pain.2008.09.025

    • PubMed
    • Google Scholar
63. 1. Silvani A

    (2019) Sleep disorders, nocturnal blood pressure, and cardiovascular risk: A translational perspective

    Auton Neurosci 218:31–42.

    https://doi.org/10.1016/j.autneu.2019.02.006

    • PubMed
    • Google Scholar
64. 1. Stepanski E
    2. Lamphere J
    3. Badia P
    4. Zorick F
    5. Roth T

    (1984) Sleep fragmentation and daytime sleepiness

    Sleep 7:18–26.

    https://doi.org/10.1093/sleep/7.1.18

    • PubMed
    • Google Scholar
65. 1. Stern J
    2. Jeanmonod D
    3. Sarnthein J

    (2006) Persistent EEG overactivation in the cortical pain matrix of neurogenic pain patients

    Neuroimage 31:721–731.

    https://doi.org/10.1016/j.neuroimage.2005.12.042

    • PubMed
    • Google Scholar
66. 1. Tan LL
    2. Oswald MJ
    3. Heinl C
    4. Retana Romero OA
    5. Kaushalya SK
    6. Monyer H
    7. Kuner R

    (2019) Gamma oscillations in somatosensory cortex recruit prefrontal and descending serotonergic pathways in aversion and nociception

    Nat Commun 10:983.

    https://doi.org/10.1038/s41467-019-08873-z

    • PubMed
    • Google Scholar
67. 1. Tokunaga S
    2. Takeda Y
    3. Shinomiya K
    4. Yamamoto W
    5. Utsu Y
    6. Toide K
    7. Kamei C

    (2007) Changes of sleep patterns in rats with chronic constriction injury under aversive conditions

    Biol Pharm Bull 30:2088–2090.

    https://doi.org/10.1248/bpb.30.2088

    • PubMed
    • Google Scholar
68. 1. Van Someren EJ
    2. Cirelli C
    3. Dijk DJ
    4. Van Cauter E
    5. Schwartz S
    6. Chee MW

    (2015) Disrupted sleep: from molecules to cognition

    J Neurosci 35:13889–13895.

    https://doi.org/10.1523/JNEUROSCI.2592-15.2015

    • PubMed
    • Google Scholar
69. 1. van Someren EJW

    (2020) Brain mechanisms of insomnia: new perspectives on causes and consequences

    Physiol Rev 101:995–1046.

    https://doi.org/10.1152/physrev.00046.2019

    • Google Scholar
70. 1. Vanhatalo S
    2. Palva JM
    3. Holmes MD
    4. Miller JW
    5. Voipio J
    6. Kaila K

    (2004) Infraslow oscillations modulate excitability and interictal epileptic activity in the human cortex during sleep

    PNAS 101:5053–5057.

    https://doi.org/10.1073/pnas.0305375101

    • PubMed
    • Google Scholar
71. 1. Vassalli A
    2. Franken P

    (2017) Hypocretin (orexin) is critical in sustaining theta/gamma-rich waking behaviors that drive sleep need

    PNAS 114:E5464–E5473.

    https://doi.org/10.1073/pnas.1700983114

    • PubMed
    • Google Scholar
72. 1. Wall PD
    2. Devor M

    (1983) Sensory afferent impulses originate from dorsal root ganglia as well as from the periphery in normal and nerve injured rats

    Pain 17:321–339.

    https://doi.org/10.1016/0304-3959(83)90164-1

    • PubMed
    • Google Scholar
73. 1. Watson BO

    (2018) Cognitive and physiologic impacts of the infraslow oscillation

    Front Syst Neurosci 12:44.

    https://doi.org/10.3389/fnsys.2018.00044

    • PubMed
    • Google Scholar
74. 1. Yüzgeç O
    2. Prsa M
    3. Zimmermann R
    4. Huber D

    (2018) Pupil size coupling to cortical states protects the stability of deep sleep via parasympathetic modulation

    Curr Biol 28:392–400.

    https://doi.org/10.1016/j.cub.2017.12.049

    • PubMed
    • Google Scholar
75. 1. Zimmermann M

    (1983) Ethical guidelines for investigations of experimental pain in conscious animals

    Pain 16:109–110.

    https://doi.org/10.1016/0304-3959(83)90201-4

    • PubMed
    • Google Scholar

Acceptance summary:

The manuscript addresses the relationship between chronic pain and the patterns and quality of sleep in mice. The study employs high quality EEG analyses to show that although mice with neuropathic pain retain general sleep architecture, they are more easily aroused in sleep and show alterations in infraslow oscillatory activity in the brain in conjunction with behavioral abnormalities. The findings provide insights into heightened arousability during sleep in pain and bad sleep quality as a comorbidity of chronic pain.

Decision letter after peer review:

Thank you for submitting your article "Cortico-autonomic local arousals and heightened somatosensory arousability during NREM sleep of mice in neuropathic pain" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Christian Büchel as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Carl Y Saab (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) Although the study reports interesting data on alterations in sleep in chronic pain, the functional relationship between pain and sleep is not addressed. In order to enable a significant scientific advance, behavioural data on pain in conjunction with altered sleep are necessary.

2) There are a number of concerns about the analysis of the EEG data. A significant reanalysis of EEG data is necessary and it is recommended to address all points related to analysis and interpretation of data on oscillatory activity.

Reviewer #1:

In the present manuscript, the authors study the relationship of chronic pain and sleep in mice. Specifically, they first established a mouse model of neuropathic pain and show that the general sleep architecture is fully preserved in mice. Then using mechanicovibrational stimulation, they show that these mice are easier arousable in sleep. Critically, they link this phenomenon to an infraslow (approx. 0.02 Hz, every 50s) rhythm. The findings are being interpreted as heightened arousability in pain during sleep, which might constitute a promising target for future interventions.

The paper is very well written, has great illustration and nicely guides the reader from the initial hypothesis to the main findings. It provides an interesting link between sleep physiology and pain perception. While overall very convincingly written, several conceptual and methodological questions remain:

1. Why is the increased arousability only linked to mechanical stimulation? Wouldn't one expect that neuropathic pain patients suffer more constantly and are thus, more likely to experience sub-optimal night sleep with more microarousals?

2. How can the authors claim specificity for the infraslow oscillation when no alternative frequency bands or spectral phenomena were probed? The general link to the 0.02 Hz oscillations is interesting, but for the naïve reader the step from Figure 1 to Figure 2 is hard to understand: Besides that they authors first described the 0.02 Hz oscillation in the Lecci 2017 paper, there is no direct obvious link to Figure 1.

In particular, since Figure 1 indicates no differences between groups. Hence, any other spectral metric could have been investigated. Maybe the authors could elaborate on the reasoning here.

3. How do the authors interpret the fact that the observed effects (e.g. Figure 3) are not frequency specific, but seem to be present in almost every frequency band? Overall, the presentation of the results would benefit from clearly separating physiology vs. pathophysiology-based assumptions. Right now, the manuscript is difficult to follow because both a presented in an interleaved manner and one wonders which effects are truly specific to the pain model.

4. The visualization of the coupling in Figure 5 appears very cluttered. Could the authors maybe include circular representations of the data in addition to the histogram-based ones (with phase on the x-axis)? Specifically to avoid the binary contrasts between two large (270-90{degree sign} and 90-270{degree sign}) bins that were used for statistical comparisons? Why not utilize appropriate statistics that do not require rendering the data into linear format, such as Watson-Williams or Harrison-Kanji tests when testing for coupling interactions.

This seems critical, since the authors essentially present a cross-frequency coupling analysis (0.02 Hz to power or arousal coupling), which might benefit from the guidelines formulatedby Aru et al. (2015) Current Opinion in Neurobiology

5. Interpretation of oscillatory signatures is sometimes a bit unspecific. E.g. in the introduction they authors discuss 'enriched oscillations', i.e. indexing increased spectral power – but how would they disentangle that from e.g. movement artifacts, which cover multiple frequency bands after spectral transformations, which is the defining element of microarousals?

6. Overall, the manuscript is difficult to read, because lots of different analyses are being introduced and detailed, but it is difficult to extract the main gist why the specific analysis was carried out.

7. Queries per specific Figures:

Figure 1A – Why is more time spent awake in the 'dark phase'? Shouldn't this be the other way round?

Figure 1E – How were the spectra normalized?

Figure 4F-H implies a non-frequency specific effect, why not quantify and display the true comodulation to assess the spectral extent?

Figure 6 – Isn't the presented analysis circular in nature? The authors predict labels using machine learning based on already classified stages using mutually-dependent criteria.

Figure 7 – A Visualization of when stimulation is most effective in controls would help.

8. Please report t values and effect sizes and not just p-values, specifically in the first part of the manuscript.

9. When discussing their finding 1 in the main discussing: What is the early stage of the disease and to what data are the authors referring to?

Taken together, the authors present an interesting study with multitude of results using a variety of analyses. Overall, this manuscript provides a better understanding for sleep physiology by further adding evidence for the importance of the 0.02 Hz oscillation. In addition, it provides an interesting perspective on pain and sleep. However, in its present form it is difficult to extract the main findings.

Overall, this is an interesting paper well within the scope of eLife, which requires several clarifications, but is a strong candidate for publication.

10. The authors should try to clearly separate hypotheses, predictions and analyses concerning physiology vs. pathophysiology. In its present form both appear interleaved, making it very difficult to follow the manuscript and reasoning.

Reviewer #2:

Rigorous investigation of potential correlation between chronic pain and sleep disorder in a mouse model of spared sciatic nerve injury (SNI). Methodology includes conventional sleep analysis using EEG, EMG and a novel arousal assay within epochs referred to as 0.02 Hz fragility-continuity during light phase NREM. No difference was reported in global sleep architecture. However, mice with SNI manifest increased cortical fragility in somatosensory hindlimb (S1) contralateral to SNI, but not in ipsilateral S1 or prelimbic cortex, compared to sham. These heart changes also correlated with increased rate. Moreover, mice with SNI were hypersensitive to spontaneous arousal, as well as arousal evoked by mechano-vibration timed to fragility periods in a closed-loop manner.

Strengths:

The topic of chronic pain, brain abnormality, EEG and sleep is timely.

Very well written manuscript, logical flow, clear diagrams.

Methodology is robust, uses conventional techniques and develops a new assay to investigate fragility-continuity that enabled novel findings. This novel assay is potentially useful to sleep researchers.

Conclusions are conservative and warranted by the data and the statistics.

Literature is comprehensively reviewed.

Weaknesses:

1. Study reads as if it is designed to explore sleep differences in SNI mice, rather than to test a hypothesis based on a clearly defined neurophysiological mechanism. There is no mechanistic evidence (other than a speculation about sustained peripheral drive from injury site, which is more stochastic than regular).

Significant findings have considerable limitations. They apply only to NREMS, during light phase, in absence of EMG, and in S1. This raises concerns about relevance and validity, especially that one animal model was used, no evidence for pain presented (and if so what pain category or correlation with pain severity), one time point at D20+, changes are minor, and other cited studies have noted local cortical arousal in humans (Nobili 2011).

2. Link between pain and sleep is bi-directional, as stated by the authors, but pain is causal to sleep disorder, so recommendations to better understand sleep to treat pain should be reworded more carefully. Also recommendations for treatment options (e.g. TMS localized to S1) should be re-considered, especially when mechanisms are still unknown.

3. Analysis of EEG data requires more clarification. Normalization of EEG is common practice, though standards differ, along with data interpretation. Due to the inherent variability in EEG, the most accurate analysis is within-subject comparison. In this study, EEG was normalized dividing by the average mean power from 0.75-47 Hz. This could lead to bias. For e.g., if absolute EEG power in the 1-45 Hz is modulated in one group, normalization will dilute this modulation and artificially enhance power in bands outside this range.

4. Movement artifact is a common source of noise in EEG. How was this dealt with? It is recommended that an automated algorithm be used to limit subjectivity and bias.

5. Authors conclude no change in theta power in awake mice with SNI. However, cited paper by Leblanc 2014 clearly showed increased theta in awake rats after CCI (another sciatic pain model). The same group has since published 7 papers replicating these findings in multiple pain models, including one study in humans. Thalamic mechanisms for theta was presented by Leblanc, as correctly discussed by the authors (however that was Leblanc 2017, not 2014, please correct). The only explanation for the discrepancy here is the fact that Leblanc recorded theta during brief (~15 min) awake resting state with no apparent movement once per day. Analysis of EEG during the entire wake period when an animal is engaged in day-time activity (potentially distracted from pain) might dilute this effect on theta power.

6. Following the same logic, it should also be noted that the increased gamma reported by Tan 2019 refers to the CFA model of acute inflammatory pain, which is different from chronic neuropathic pain.

7. The study is largely based on the premise that mice are in chronic pain at D20+ after SNI. Typically, studies using animal models include behavioral data to verify the pain state (for e.g. sensory hypersensitivity, etc.) Whereas SNI is a widely used model, verification of model reproducibility is essential. Moreover, it appears from the literature that the predominant evidence for pain in this model is sensory hypersensitivity evoked by mechanical and thermal stimuli. Do we know if rodents (particularly mice) with SNI manifest spontaneous pain? This is important because only one model is used in this study and at only one time-point. This is also relevant when making translational claims about human pain, which is multi-dimensional (reference to model mimicking human pain should be avoided).

8. Authors cite several studies showing abnormal findings in alpha, beta and low-gamma during sleep (p4). These findings seem inconsistent with this paper. Please discuss (also make sure you're making apples-to-apples comparisons, including epoch lengths, normalization, model, etc.)

9. Figure legends too crowded with statistics.

10. Some icons in legends hard to interpret (e.g. Figure 1G).

11. What are units in Figure 1F?

Essential revisions:

1) Although the study reports interesting data on alterations in sleep in chronic pain, the functional relationship between pain and sleep is not addressed. In order to enable a significant scientific advance, behavioural data on pain in conjunction with altered sleep are necessary.

This is an important remark that is clearly essential for this study. We provide here two novel datasets for this outstanding issue that helps resolve it in two complementary aspects.

i. We present behavioral responses to mechanical and thermal (heat) stimulation after SNI and sham surgeries in mice. The data are from another ongoing experiment in our laboratory, and they include measurements at times points similar to the present study. In our hands, and in the hands of many other labs, all nerve-injured animals with the well performed SNI technique develop tactile and thermal allodynia/hyperalgesia-like behavior (for review, see Guida et al., 2020; doi:10.3390/ijms21093396).These data are included in the rebuttal to the concerns of Reviewer 2.

ii. We re-analyze the data from the mice included in this study regarding the presence of elevated theta and high-frequency power. Such elevations have been repeatedly linked to pain hypersensitivity in experimental models and described as an “electroencephalographic signatures of pain” by LeBlanc et al. (2017). They are found in acute and chronic experimental models and respond to analgesics. When suppressed by optogenetic manipulation of thalamic activity in mice, or via thalamic lesions in humans, pain relief has been observed (Sarnthein et al., 2006) (LeBlanc et al., 2016) (LeBlanc et al., 2017). We do find changes in the power spectrum of the SNI groups during quiet wakefulness (a state determined polysomnographically), which strongly supports the presence of neuropathic pain-like hypersensitivity in the animals used in our study. These data are included in a new Figure 2 in the main manuscript.

In conclusion, we now provide the most sensitive physiological measures of experimental chronic pain currently available. In particular, we were able to correlate it with the sleep data, as requested by the Editor.

2) There are a number of concerns about the analysis of the EEG data. A significant reanalysis of EEG data is necessary and it is recommended to address all points related to analysis and interpretation of data on oscillatory activity.

In response to this essential revision request, we have carried out all the analyses requested by the reviewers. We provide detailed answers to these in a point-to-point manner.

Reviewer #1:

[…] The paper is very well written, has great illustration and nicely guides the reader from the initial hypothesis to the main findings. It provides an interesting link between sleep physiology and pain perception. While overall very convincingly written, several conceptual and methodological questions remain:

1. Why is the increased arousability only linked to mechanical stimulation? Wouldn't one expect that neuropathic pain patients suffer more constantly and are thus, more likely to experience sub-optimal night sleep with more microarousals?

The increased arousability is linked to both – mechanical stimulation on the skull, and a higher frequency of brief arousals that we characterize as “cortico-autonomic arousals” in the title. These are not the usual microarousals commonly described in mice. Instead, these are cortical arousals occurring locally in the cortex but that are nevertheless accompanied by heart rate increases. They take place in the hindlimb primary somatosensory cortex that is linked to the discriminative sensory component of pain processing (localization, intensity and modality).

We conclude that nerve injury, and thus neuropathic pain-like alterations, augments both, arousability to external mechanical stimulation and spontaneous arousability in precisely those brain areas that are concerned with pain processing.

2. How can the authors claim specificity for the infraslow oscillation when no alternative frequency bands or spectral phenomena were probed? The general link to the 0.02 Hz oscillations is interesting, but for the naïve reader the step from Figure 1 to Figure 2 is hard to understand: Besides that they authors first described the 0.02 Hz oscillation in the Lecci 2017 paper, there is no direct obvious link to Figure 1.

In particular, since Figure 1 indicates no differences between groups. Hence, any other spectral metric could have been investigated. Maybe the authors could elaborate on the reasoning here.

We are grateful for this comment that made us realize that we have to describe more in details the reasoning for the use of the infraslow oscillation. We do this now explicitly in the rewritten Introduction (3^rd paragraph), in which we review evidences that the infraslow time scale stands out in terms of identifying moments of sleep vulnerability.

Regarding the link from Figure 1 to Figure 2 that the reviewer describes as hard to understand: The figure sequence has now been changed such that the physiology of the infraslow oscillation comes first, and is then followed by studies in the experimental model of neuropathic pain (SNI).

Together, with these re-arrangements, we believe we have solved the major concerns of the reviewer: the issue of specificity, the clarity of the figure arrangement, and the reasoning for the choice of the infraslow oscillation.

3. How do the authors interpret the fact that the observed effects (e.g. Figure 3) are not frequency specific, but seem to be present in almost every frequency band? Overall, the presentation of the results would benefit from clearly separating physiology vs. pathophysiology-based assumptions. Right now, the manuscript is difficult to follow because both a presented in an interleaved manner and one wonders which effects are truly specific to the pain model.

We are not sure which observed effects in Figure 3 the reviewer refers to. The data in this figure do not concern frequency specificity. In case the reviewer thought about Figure 4, then we remark the following: the observed effects cover precisely the frequency bands that need to change to give rise to an arousal: low-frequency bands decrease in power, while high-frequency bands increase.

We have taken up the reviewer’s excellent suggestion to better separate physiology versus pathophysiology. To do so, we position the Figure describing the phase-coupling of microarousals in healthy mice to the infraslow oscillation as the new Figure 1 (original Figure 2). We additionally modified the figure to make the infraslow oscillation more clear. Then, we introduce a new Figure 2 describing the physiological correlates of pain signature in SNI operated mice. The remaining figures are then focused on sleep pathophysiology. In this way, the separation between physiology and pathophysiology should be clear.

4. The visualization of the coupling in Figure 5 appears very cluttered. Could the authors maybe include circular representations of the data in addition to the histogram-based ones (with phase on the x-axis)? Specifically to avoid the binary contrasts between two large (270-90{degree sign} and 90-270{degree sign}) bins that were used for statistical comparisons? Why not utilize appropriate statistics that do not require rendering the data into linear format, such as Watson-Williams or Harrison-Kanji tests when testing for coupling interactions.

This seems critical, since the authors essentially present a cross-frequency coupling analysis (0.02 Hz to power or arousal coupling), which might benefit from the guidelines formulatedby Aru et al. (2015) Current Opinion in Neurobiology.

We have simplified the presentation of Figure 5 (now the new Figure 6) to amend the reviewers’ concern. The changes involve:

– The data of the original panels 5 H and I are now moved to the supplemental figure 2, panels A-C;

– The data from the original panel J are now shown more expanded in the new panel 6 H;

– The data from the original panels K, L are now shown expanded in the new panels 6 I, J.

Regarding the use of circular representations, we agree that this could be a worthwhile alternate way of presenting data that involve phase relationships. We are actually using this kind of representation in the lab (as recently done for the tuning of head direction units in the 360 ° space, see Vantomme et al., 2020, doi: 10.1016/j.celrep.2020.107747). Regarding the infraslow oscillation, we would prefer to keep a linear presentation that we have already started using some years ago (Lecci et al., 2017, doi: 10.1126/sciadv.1602026) when we first reported on the infraslow oscillation. Furthermore, we find the representation of the microarousal clustering around the trough of the oscillation (see new Figure 1) clearer as it brings the importance of the trough of the oscillation for sleep’s vulnerability to the forefront.

This representation also makes it clear that we are indeed interested in the two large bins described by the reviewer. These bins are defined by the natural clustering of microarousals. We currently have no incentive to use a finer graduation of angles as the microarousal distribution does not give us a physiological reason to do so.

5. Interpretation of oscillatory signatures is sometimes a bit unspecific. E.g. in the introduction they authors discuss 'enriched oscillations', i.e. indexing increased spectral power – but how would they disentangle that from e.g. movement artifacts, which cover multiple frequency bands after spectral transformations, which is the defining element of microarousals?

The term “enriched oscillations” was used in the context of the description of insomnia patients. This part has been removed and the term “enriched oscillations” is no longer present.

Regarding movement artifacts.

The Intan recording system we used to obtain the data presented in this study renders movement artefacts very small. The reason is that the signals are digitized at the level of the pre-amplifier that is present on the animals head. In this way, movements that primarily arise from the movement of the cables are eliminated. Please also realized that when we study the novel cortico-autonomic arousals the EMG is completely flat (panel C of the new Figure 6). Therefore, in the analysis of these novel events, movement artefacts are not a potential source of confound. Similarly, when we analyze the spectral composition of quiet wakefulness, the EMG levels are very low (see the new Figure 2, panel C).

More generally, movement artefacts become a minor concern once recordings are obtained intracranially, as is the case for our LFP recordings. The problem is much greater in the case of surface EEG recordings, as is frequently done in humans. In our experiments, EEG signals were only used for scoring of microarousals and for the extraction of the infraslow dynamics in Figure 1. For both these analyses, movement artefacts played no role.

6. Overall, the manuscript is difficult to read, because lots of different analyses are being introduced and detailed, but it is difficult to extract the main gist why the specific analysis was carried out.

We appreciate this comment because it helped us realize the need of improving the clarity and structure of the manuscript. As already described, the work is now separated into sleep physiology- and abnormal sleep related to an experimental model of neuropathic pain. The transition between these parts is more explicitly described and so are the analyses.

7. Queries per specific Figures:

Figure 1A – Why is more time spent awake in the 'dark phase'? Shouldn't this be the other way round?

Mice are nocturnal animals. Therefore, they spent most of their time awake during the dark phase (night), and most of their time asleep during the light phase (day).

Figure 1E – How were the spectra normalized?

We clearly describe the normalization procedure in the Methods on p.30-31 of the clean manuscript (lines 638-645). For the spectra of quiet wakefulness, we followed the procedure specified in the papers of C. Saab and colleagues, 2014, doi:10.1016/j.pain.2014.01.013 to enable direct comparisons between models (p.30, lanes 630-637).

Figure 4F-H implies a non-frequency specific effect, why not quantify and display the true comodulation to assess the spectral extent?

We have already explained above that we motivate our choice of analysis based on the natural physiology of how microarousals clusters around the trough of the infraslow oscillation. Our major questions were: what are the spectral changes at these moments of high sleep vulnerability? Do they coincide with power changes in frequency patterns that coincide with an arousal? Our results show that this is indeed the case: at moments of elevated sleep vulnerability, LFP data collected in the S1 hindlimb cortex show a pattern of frequency changes that indicate that an arousal takes place. An arousal is not a frequency-specific change, it consists of a decrease in low- and an increase in the power of higher frequencies. This is exactly what we see.

Figure 6 – Isn't the presented analysis circular in nature? The authors predict labels using machine learning based on already classified stages using mutually-dependent criteria.

For this analysis, we first train the algorithm on datasets that are already scored. It is only then that the online monitoring of continuity and fragility periods is done. Thus, the presented analysis is not circular. The training was only done to enable the machine to predict the continuity and fragility periods online. No information about the MAs themselves was given during the training of the network. Moreover, the training was done on completely separate animals compared to the one used in the paper.

Figure 7 – A Visualization of when stimulation is most effective in controls would help.

We are not sure what the reviewer means by most effective. In this figure (new Figure 8), we clearly indicate two representative traces of wake-up and sleep-through (panel C). Moreover, using color-coded arrows, we present several NREMS bouts and indicate when stimulation resulted in a wake-up or a sleep-through.

8. Please report t values and effect sizes and not just p-values, specifically in the first part of the manuscript.

Please find now a new Supplementary Table for Statistics, arranged according to figure panel, statistical test used, group size, value of test statistics, p values and Cohen’s effect sizes. We have instead removed this info from all figure legends to simplify reading.

9. When discussing their finding 1 in the main discussing: What is the early stage of the disease and to what data are the authors referring to?

We study sleep 22-23 days after SNI surgery. In this model, pain hypersensitivity is initiated during the first 3 days after surgery, reaches a plateau a week after and is stable for weeks, even months at. Anxiety- and depressive-like behavior as well as cognitive impairment are seen during late phases. Three weeks after SNI, peripheral abnormal activity and central sensitization mechanisms responsible for pain hypersensitivity are established. On the other hand, at this time point, there is only low level of psychopathology associated with peripheral nerve injury and cognitive bias (see review of Guida et al., 2020 doi:10.3390/ijms21093396). In addition, the absence of an increased REMS in our animals (seen in Nollet et al., 2019, doi:10.1073/pnas.1816456116) also tends toward the conclusion that nerve-injury related pain is predominant during this time frame.

10. The authors should try to clearly separate hypotheses, predictions and analyses concerning physiology vs. pathophysiology. In its present form both appear interleaved, making it very difficult to follow the manuscript and reasoning.

With the above-mentioned points, we hope to have responded to these recommendations so that the paper is now easy to follow in terms of reading and reasoning.

Reviewer #2:

[…] 1. Study reads as if it is designed to explore sleep differences in SNI mice, rather than to test a hypothesis based on a clearly defined neurophysiological mechanism. There is no mechanistic evidence (other than a speculation about sustained peripheral drive from injury site, which is more stochastic than regular).

Significant findings have considerable limitations. They apply only to NREMS, during light phase, in absence of EMG, and in S1. This raises concerns about relevance and validity, especially that one animal model was used, no evidence for pain presented (and if so what pain category or correlation with pain severity), one time point at D20+, changes are minor, and other cited studies have noted local cortical arousal in humans (Nobili 2011).

We thank the reviewer for this detailed summary that clearly lays out the revisions required to improve the strengths of the paper. We hope that with the revisions added, we have amended the major concerns. Please note in particular that we now do present evidence that mice might be suffering from pain electroencephalography evidences. We have now also expanded the analysis beyond NREMS to active and quiet wakefulness, alleviating the concern of the limited significance.

2. Link between pain and sleep is bi-directional, as stated by the authors, but pain is causal to sleep disorder, so recommendations to better understand sleep to treat pain should be reworded more carefully. Also recommendations for treatment options (e.g. TMS localized to S1) should be re-considered, especially when mechanisms are still unknown.

We agree that we have kept the text on the implications of our findings for treatment options short and over-conclusive on the dimension of sleep intervention as a new approach for pain alleviation. We have now formulated a novel paragraph at the end of the discussion to discuss this topic – we discuss both, the implications for animal models of sleep disorders, and for diagnosis and management strategies in patients. In addition, bad sleep decreases highly the quality of life in patients suffering from neuropathic pain. A large subset of patients have an intractable form of neuropathic pain, and improving quality of life by reducing pain interferences such as sleep and mood disorders is essential for their treatment. In this line, we think that we point out early mechanisms for sleep disturbances that can be further explored to understand better their contribution to the persistence of pain and development of other related alterations.

3. Analysis of EEG data requires more clarification. Normalization of EEG is common practice, though standards differ, along with data interpretation. Due to the inherent variability in EEG, the most accurate analysis is within-subject comparison. In this study, EEG was normalized dividing by the average mean power from 0.75-47 Hz. This could lead to bias. For e.g., if absolute EEG power in the 1-45 Hz is modulated in one group, normalization will dilute this modulation and artificially enhance power in bands outside this range.

We appreciate the reviewers’ concerns regarding normalization. Normalization of EEG power spectra is common practice. Normalization removes the variations in absolute power values between individuals that naturally arise from the different positioning and/or impedance of recording electrodes with respect to the brain and from variations in brain size. Moreover, normalization permits to see directly how much power changes in one frequency band relative to another one. This is exactly the information that is needed to estimate chronic alterations in brain states after in SNI in mice.

What is critical is how normalization is done. In our case, we used the 0.75-47 Hz band to determine relative power changes between vigilance states. When it came to the quantification of gamma power that lies outside this band, we used only power value ratios between baseline and after SNI within individuals. Therefore, there were no artificial enhancements of power values.

For the quantification of changes in theta power in quiet wakefulness, we used normalization procedures described previously for the chronic constrictive model of neuropathic pain (CCI) model and other pain models (LeBlanc et al., 2014, doi:10.1016/j.pain.2014.01.013) in the interest of comparison (see also response to reviewer 1 and below). This consisted of normalizing the spectra by their mean power between 2-30 Hz.

4. Movement artifact is a common source of noise in EEG. How was this dealt with? It is recommended that an automated algorithm be used to limit subjectivity and bias.

We have received a similar comment from Reviewer 1. Therefore, we reply here in the same words.

The Intan recording system we used to obtain the data presented in this study renders movement artefacts very small. The reason is that the signals are digitized at the level of the pre-amplifier that is present on the animals head. In this way, movements that primarily arise from the movement of the cables are eliminated. Please also realized that when we study the novel cortico-autonomic arousals the EMG is completely flat (panel C of the new Figure 6). Therefore, in the analysis of these novel events, movement artefacts are not a potential source of confound. Similarly, when we analyze the spectral composition of quiet wakefulness, the EMG levels are very low (see the new Figure 2, panel C).

More generally, movement artefacts become a minor concern once recordings are obtained intracranially, as is the case for our LFP recordings. The problem is much greater in the case of surface EEG recordings, as is frequently done in humans. In our experiments, EEG signals were only used for scoring of microarousals and for the extraction of the infraslow dynamics in Figure 1. For both these analyses, movement artefacts played no role.

5. Authors conclude no change in theta power in awake mice with SNI. However, cited paper by Leblanc 2014 clearly showed increased theta in awake rats after CCI (another sciatic pain model). The same group has since published 7 papers replicating these findings in multiple pain models, including one study in humans. Thalamic mechanisms for theta was presented by Leblanc, as correctly discussed by the authors (however that was Leblanc 2017, not 2014, please correct). The only explanation for the discrepancy here is the fact that Leblanc recorded theta during brief (~15 min) awake resting state with no apparent movement once per day. Analysis of EEG during the entire wake period when an animal is engaged in day-time activity (potentially distracted from pain) might dilute this effect on theta power.

This is an excellent point raised by the reviewer that we have now followed up using an analysis of quiet wakefulness. The results are now included (see figure summarized in a new Figure 2. Briefly, after SNI, mice do show an enhanced power at frequencies above ~5 Hz). When binned in frequency bands for theta (5-10 Hz), σ (10-15 Hz), β (16-25 Hz) and low (26-40 Hz) and high (60-80 Hz), we find significant increases in theta, beta and low gamma power.

We now cite adequately the LeBlanc studies.

6. Following the same logic, it should also be noted that the increased γ reported by Tan 2019 refers to the CFA model of acute inflammatory pain, which is different from chronic neuropathic pain.

We now appropriately cite this reference.

7. The study is largely based on the premise that mice are in chronic pain at D20+ after SNI. Typically, studies using animal models include behavioral data to verify the pain state (for e.g. sensory hypersensitivity, etc.) Whereas SNI is a widely used model, verification of model reproducibility is essential. Moreover, it appears from the literature that the predominant evidence for pain in this model is sensory hypersensitivity evoked by mechanical and thermal stimuli. Do we know if rodents (particularly mice) with SNI manifest spontaneous pain? This is important because only one model is used in this study and at only one time-point. This is also relevant when making translational claims about human pain, which is multi-dimensional (reference to model mimicking human pain should be avoided).

This is an important question and we divide our answer in several parts.

Regarding the need of behavioral data: We have intentionally refrained from testing the animals for recording stimulus-evoked pain hypersensitivity. First, we have large evidence in our lab that all operated mice develop pain-related behavior. Second, we know that behavioral testing of pain sensitivity is stressful for the animals (the tests induce nocifensive behavior only in SNI animals plus testing itself is disrupting the animal). It can likely be a cause of augmented sleep deprivation in the SNI group and would be a confounding factor in our experiments. We aimed at having only one distinct variable that is the sham or the SNI surgery

We add that, all along the years, all animals with SNI in our lab (and in the lab of many colleagues using the SNI model), display nocifensive behavior to non-nociceptive stimuli (Von Frey threshold assays with up and down method) or lower latency to heat-induced pain-like behavior. We are thus entirely confident that the mice used in the present work would have pain-hypersensitivity like behavior, an indirect sign of peripheral neuropathic pain-like behavior.

Regarding the presence of spontaneous pain. We agree that this is an indirect sign of pain, which does not correspond to spontaneous pain, but the discussion on this topic is not in the scope of our work on sleep. We can only mention that other laboratories studied the appearance of signs of spontaneous pain in this model (Langford et al., 2010, doi: 10.1038/nmeth.1455; Tappe-Theodor and Kuner, 2014, doi: 10.1111/ejn.1264). To be very clear, we have now modified the text and use appropriate phrasing to avoid any confusion between an experimental model of neuropathic pain and the presence of spontaneous pain. As stated by the reviewers’ using experimental model of nerve injury and claiming that the animal is suffering from neuropathic pain is over conclusive, and it is the limitation that all pain scientists encounter. Nonetheless, the physiological data provided in the new Figure 2 provide new indirect signs – as said by the author – a signature of pain in the mice.

8. Authors cite several studies showing abnormal findings in α, β and low-γ during sleep (p4). These findings seem inconsistent with this paper. Please discuss (also make sure you're making apples-to-apples comparisons, including epoch lengths, normalization, model, etc.)

These portions of the introduction have been removed and the introduction rewritten to better comply with the specific points address in this study. Moreover, analysis for the frequency bands has been done and added in a new Figure 2, as discussed repeatedly above.

9. Figure legends too crowded with statistics.

We have now added a new Supplementary Table for Statistics and removed all detailed statistics from the figure legends.

10. Some icons in legends hard to interpret (e.g. Figure 1G).

We have now removed the small icons that referred to statistics. There are no longer icons that are hard to interpret on any figure.

11. What are units in Figure 1F?

These units represent power relative to baseline. The indication is present below the plots. The new figure arrangement presents this in a much clearer way.

Author details

1. Romain Cardis

   1. Department of Fundamental Neurosciences, Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland
   2. Pain Center, Department of Anesthesiology, Lausanne University Hospital (CHUV), Lausanne, Switzerland

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing – review and editing

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0002-0595-4175

2. Sandro Lecci

   Department of Fundamental Neurosciences, Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland

   Contribution

   Conceptualization, Data curation, Formal analysis, Methodology

   Competing interests

   none

3. Laura MJ Fernandez

   Department of Fundamental Neurosciences, Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland

   Contribution

   Data curation, Methodology

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0002-7942-3369

4. Alejandro Osorio-Forero

   Department of Fundamental Neurosciences, Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland

   Contribution

   Methodology

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0003-4341-4206

5. Paul Chu Sin Chung

   Pain Center, Department of Anesthesiology, Lausanne University Hospital (CHUV), Lausanne, Switzerland

   Contribution

   Formal analysis, Writing – review and editing

   Competing interests

   None

6. Stephany Fulda

   Sleep Medicine Unit, Neurocenter of Southern Switzerland, Civic Hospital (EOC) of Lugano, Lugano, Switzerland

   Contribution

   Conceptualization, Validation, Writing – review and editing

   Contributed equally with

   Isabelle Decosterd and Anita Lüthi

   For correspondence

   stephany.fulda@gmail.com

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0001-8416-2610

7. Isabelle Decosterd

   Pain Center, Department of Anesthesiology, Lausanne University Hospital (CHUV), Lausanne, Switzerland

   Contribution

   Conceptualization, Data curation, Funding acquisition, Project administration, Validation, Visualization, Writing – original draft, Writing – review and editing

   Contributed equally with

   Stephany Fulda and Anita Lüthi

   For correspondence

   isabelle.decosterd@chuv.ch

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0002-4820-5289

8. Anita Lüthi

   Department of Fundamental Neurosciences, Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland

   Contribution

   Conceptualization, Data curation, Funding acquisition, Project administration, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing

   Contributed equally with

   Stephany Fulda and Isabelle Decosterd

   For correspondence

   anita.luthi@unil.ch

   Competing interests

   none

   "This ORCID iD identifies the author of this article:" 0000-0002-4954-4143

• 1,245

  Page views

• 224

  Downloads

• 6

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.