Behavioural Brain Research 291 (2015) 189–194
Contents lists available at ScienceDirect
Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr
Short Communication
Absence of rapid eye movements during sleep in adult zebrafish B.B. Árnason a , H. Þorsteinsson a,b , K.Æ. Karlsson a,b,∗ a b
School of Science and Engineering, Reykjavik University, Reykjavik, Iceland 3Z Pharmaceuticals, Reykjavik, Iceland
h i g h l i g h t s • • • • •
Zebrafish are novel yet important model in sleep research. Sleep in zebrafish is measured using behavioral assays that cannot differentiate between REM and NREM sleep. In current study we use image analysis to measure two key REM components: rapid eye movements and respiratory rate. We show that there are no sleep related eye movements but during sleep respiration becomes slower and less regular. We find no evidence for REM–NREM cycles in zebrafish.
a r t i c l e
i n f o
Article history: Received 6 January 2015 Received in revised form 9 May 2015 Accepted 12 May 2015 Available online 20 May 2015 Keywords: REM Sleep Zebrafish Computer vision Motion tracking Activity detection
a b s t r a c t Sleep is not a uniform phenomenon, but is organized in alternating, fundamentally different states, rapid eye movement sleep and non-rapid eye movement sleep. Zebrafish (Danio rerio) have recently emerged as an excellent model for sleep research. Zebrafish are well characterized in terms of development, neurobiology and genetics. Moreover, there are many experimental tools not easily applied in mammalian models that can be readily applied to zebrafish, making them a valuable additional animal model for sleep research. Sleep in zebrafish is defined behaviorally and exhibits the hallmarks of mammalian sleep (e.g. sleep homeostasis and pressure). To our knowledge no attempts have been made to discern if sleep in zebrafish entails alternations of REM–NREM sleep cycles which are critical for further development of the model. In the current experiment we quantify two key REM sleep components, rapid eye movements and respiratory rates, across sleep–wake cycles. We find no sleep-related rapid eye movements. During sleep respiratory rates, however, are reduced and become less regular, further establishing that the behavioral definition used truly captures a change in the fish’s physiology. We thus fail to find evidence for REM–NREM sleep cycles in zebrafish but demonstrate a physiological change that occurs concomitantly with the previously defined behavioral state of sleep. We do not rule out that other phasic REM components (e.g. atonia, cardiac arrhythmias, myoclonic twitches or desynchronized EEG) are coherently expressed during sleep but we conclude that adult zebrafish do not have REM-sleep-related rapid eye movements. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The necessity of using highly prolific genetically tractable animals in sleep research, many of whom do not have differentiated EEG, has resulted in the re-adoption of the older behavioral definition of sleep [1]. According to a behavioral definition a state can be considered sleep if: (i) it is characterized by an absence of voluntary movement; (ii) it is spontaneous, occurring with a circadian rhythm; (iii) it is reversible; (iv) it is characterized by
∗ Corresponding author at: School of Science and Engineering, Reykjavik University, Reykjavik, Iceland. Tel.: +354 8256467. E-mail addresses: [email protected], [email protected] (K.Æ. Karlsson). http://dx.doi.org/10.1016/j.bbr.2015.05.017 0166-4328/© 2015 Elsevier B.V. All rights reserved.
a species-specific posture; (v) it includes an increased arousal threshold; (vi) and is under homeostatic control [2]. Behavioral definition of sleep has been successfully applied to recent, but important, animal models in sleep research, such as fruit flies (Drosophila melanogaster) [3,4] and Zebrafish (Danio rerio) [5,6]. Both adult and larval zebrafish represent an important model organism with established applications in sleep research [7–11]. Since differentiated EEG has not been established in zebrafish, their use as a model in sleep research is limited to particular behavioral definitions of sleep for larvae [6,7,11] and adults [5]. Zebrafish sleep is similar to mammalian sleep as shown by a comparison of behavioral, physiological, and pharmacological characteristics [5,7,11] and their neurobiology is comparable to mammals [12]. Zebrafish behavior has been shown to follow a circadian rhythm
190
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
and to be regulated by a homeostatic mechanism similar to that in mammals [5,9–11]. Studies using both electrical stimulation and light to interrupt sleep show that following 6 s of immobility, adult zebrafish do exhibit homeostasis and sleep pressure, albeit weaker than in mammals [5,9,13]. Furthermore, drugs which affect mammalian sleep have been shown to similarly affect sleep in larval zebrafish [7,14]. Importantly, the current behavioral definition of sleep in zebrafish does not allow a differentiation between REM and NREM sleep states. The fact that REM or REM-like sleep has been shown to exist in both ectothermic and endothermic vertebrates [15,16] warrants a thorough study of the presence of REM or REM-like sleep in fish. Establishing the existence of REM in either adult or larval zebrafish would be highly valuable for sleep research. First, it would open various realms of research on the neural substrates of sleep in general, where full advantage of the unique research tools applicable to zebrafish could be taken. Second, it would allow for novel applications in drug screening where one could screen for drugs with specific effects on either REM or NREM sleep. In the current study we therefore used image analysis in order to determine whether or not adult zebrafish express two of the phasic REM-sleep components, rapid eye movements and/or irregular respiration. 2. Methods 2.1. Fish A total of 19 adult zebrafish of the AB strain, 4–6 months of age, were used in the study. Fish were fed twice a day on a variable diet of TetraMin flakes (Tetra Holding GmbH, Melle, Germany) and Adult Zebrafish Complete Diet (Zeigler Bros, USA). The fish were kept in a 14:10 light:dark cycle in a 3 or 10 L multi tank constant flow system (Aquatic Habitats, Apopka, FL, USA). Water temperature was held at a constant 28.5 ◦ C and replaced at a rate of 10% per day. 2.2. Data acquisition All tracking trials and recordings were done in a 75 L aquarium containing two identical small chambers made from opaque black Plexiglas, isolating the subject within an environment of 14 cm (wide) × 4.5 cm (deep) × 14 cm (high), but with the same light–dark cycle and temperature controls as described above. The aquarium was illuminated with two infrared lights of 0 lx (darkphase), and a fluorescent light of 255 lx (light-phase) positioned above the chamber. Water was circulated from a 50 L reservoir tank at the rate of 2.3 L/min. Thermosensor, filter, air diffuser, and water pump were placed in the aquarium. Behavior was tracked at 5 and 30 Hz (baseline and bout analysis, respectively) with a Basler ace acA2000-50gmNIR infrared camera (Basler, Germany), using a 50 mm CCTV Pentax lens (Pentax, Germany). Data were acquired with custom written programs in MATLAB R2013a (The MathWorks Inc., Natick, MA, 2013). All test subjects were placed in the acclimation chamber for 24 h. Fish were placed in the acclimation chamber between 11:00 and 12:00; the 24 h recording commenced at 12:00 the following day. Recordings were performed during day and night, with lights-on and lights-off (lights-on at 08:00; lights-off at 22:00). All procedures in this study were carried out in strict compliance with the regulations of, and approved by, the National Bioethics Committee of Iceland (regulation 279/2002); permit issued to KÆK 19th May 2008. 2.3. Data processing First, velocity was plotted across 24 h to verify expression of normal sleep behavior. Second, we parsed the data into bouts of
Table 1 Overview of experimental conditions. Subject/trial 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Acclimation √ √ √ √ √ √ √ √ √ √ √ √ √ √
Baseline √ √ √ √ √ √ √ √
Day/light
√ √ √ √ √
Night/dark √ √ √ √ √ √ √ √ √
Night/light
√ √ √ √ √
sleep and wakefulness. Based on previous studies of adult zebrafish [5] fish were considered immobile if their velocity did not exceed 1.5 cm/s and periods of immobility exceeding 6 s were categorized as sleep [5,9,10]; i.e. the 7th second and above were counted as sleep. Thus, a single sleep bout is the time from the 6th second of immobility until movement exceeds 1.5 cm/s. Using custom written MATLAB routines periods of immobility were obtained from each fish representing all recording categories (day/night and lights-off/on). Third, using videos from immobile fish only, two region-of-interest videos were produced: one from around the eye (to produce eye-movement data) and one around the head (to produce respiratory-movement data). Next, activity detection module was run on the cropped videos, followed by low pass (0.1 Hz) filtering on the respiratory-movement data (to remove high frequency noise) and high pass (5.0 Hz) filtering on the eye-movement data (to remove artifacts caused by respiratory movements). Fig. 1A–C depicts the protocol for detecting eye-movements and Fig. 1D–F depicts the protocol for detecting respiratory movements. 2.4. Data analysis Data for analysis were randomly selected from throughout the night from each set of conditions: night with lights-off: 20 random samples from each subject with a duration of at least 7 s (to include at least one second of sleep), a total of 100 bouts. Day with lights-on: a total of 18 bouts with duration of at least 7 s. Night with lights-on: a total of 48 bouts with duration of at least 7 s. Eye and respiratory movements were then compared between bouts of immobility or between bouts of sleep and wakefulness. Table 1 shows the number of fish in each condition. 2.5. Statistics Mean frequencies were compared using Kolmogorov–Smirnoff test of equal means, changes in peak-to-peak distance were compared using Levene’s test of equal variances and comparison of mean rates within sleep-bout quadrants were performed using the Friedman test. All data are presented with their standard error of measurement. 3. Results 3.1. Eye movements Fig. 2A depicts representative data for eye movements. Peaks above a threshold value of 8 were found to accurately represent eye movements (verified by three independent reviewers from video data). No eye movements were registered in any sleep bout at night with lights-off or during day with lights-off. Eye movements were
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
191
Fig. 1. Overview of activity detection. Activity detection was performed on stationary videos only and based on frame-differencing where a video frame (M) at tx is subtracted from a frame at a later point in time t(x+1) and the absolute value is taken of the result. Mt(x +1) = |Mt(x +1) − Mtx |. This difference frame (M) is the absolute difference in pixel values between t(x+1) and tx . (A)–(C) schematic overview of eye movement detection. (D)–(F) schematic overview of respiratory movement detection.
192
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
Fig. 2. Representative eye- and respiratory movements. (A) Green line represents sleep at night during lights-off. Red and blue lines represent sleep and day during lights-on and night during lights-on, respectively. No eye movements were detected during sleep with lights-off during day or night. (B) Representative recording of respiratory movement during night at lights-off. All tracks are derived from the same fish.
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
193
registered in every sleep bout while lights were on, both during day and night. Solid lines show the activity detection data and registered eye movements are represented with dots (for the following analysis n = 14). During day with lights-on eye movement frequency during the first 6 s was 0.6481 ± 0.098 Hz and 0.5025 ± 0.093 during the remainder of the bouts. A Kolmogorov–Smirnov test revealed no significant difference in eye movement frequency (p = 0.7088). During night with lights-on eye movement frequency during the first 6 s was 0.4271 ± 0.051 Hz compared to 0.6132 ± 0.086 during the remainder of the bouts. A Kolmogorov–Smirnov test revealed a significant difference in eye movement frequency (p = 0.0070), the frequency being lower during the first 6 s. When the whole bout of immobility was combined (i.e. combining sleep and wake bouts) eye movement frequency did not differ between night and day. The frequency of eye movements during day with lights-on was 0.6258 ± 0.075 Hz and during night with lights-on 0.5499 ± 0.055 Hz. A Kolmogorov–Smirnov test revealed no significant difference in eye movement frequency (p = 0.7854). Also, a Kolmogorov–Smirnov test, comparing eye movement frequency during sleep bouts, (immobility beyond the first 6 s) for day and night with lights on, showed no difference (p = 0.5685). Furthermore, visual observation and spectrogram analysis of the eye movement signals revealed that the registered eye movements did not come in bursts of rapid saccades but only as singular rotations of the eyes. Average eye movement frequency (during lights-on only) was 0.7022 ± 0.0497 Hz.
3.2. Respiration Fig. 2B shows representative data for respiration rates. Data are represented as average pixel value change per video frame. Peaks in the data signals were automatically registered and peaks above a threshold value of 4 were found to accurately represent respiratory movements (verified by three independent reviewers from video data) (for following analysis n = 14). Fig. 3A shows a comparison of respiratory rates during night with lights-off for the first 6 s of an immobility bout versus the remainder of the bout (sleep). The mean respiratory rate during the first 6 s was 0.4883 ± 0.027 Hz and for the remainder 0.4127 ± 0.034 Hz. A Kolmogorov–Smirnov test showed a significant difference (p = 0.0018) where the respiratory rate during the first 6 s is higher than during the remainder of the bout. The same test was used on respiratory rate data from sleep bouts during day with lights-on and night with lights-on, both cases showed no significant difference (p = 0.4255 and p = 0.6521). Fig. 3B shows a comparison of the peak-to-peak intervals of breaths during night with lights-off between the first 6 s of an immobility bout versus the remainder of the bout (sleep). A Levene’s test of equal variances showed a significant difference between periods (F(1,495) = 6.5702, p < 0.05) where the variance during the first 6 s is lower than during the remainder of the bout. Fig. 3C shows a comparison of respiratory rates during night with lights-off between sleep bout quadrants. Mean respiratory rates for quadrants 1–4 were, respectively, 0.5428 ± 0.042, 0.4489 ± 0.034, 4480 ± 0.034 and 0.4185 ± 0.034 Hz. A Friedman test revealed a significant difference between groups (2 (3,297) = 17.0579, p < 0.05). A multiple comparison test, showed that the third and fourth quadrants were significantly different from the first with respiratory rates lower during the last two quadrants. Average respiratory frequency across all conditions was 0.6396 ± 0.0368 Hz.
Fig. 3. Respiratory parameters at night during lights-off. (A) Comparison of respiratory frequency between first 6 s (wake) and remainder (sleep) of a bout of immobility. (B) Comparison of peak-to-peak variability between first 6 s (wake) and remainder (sleep) of a bout of immobility. (C) Comparison of respiratory frequency between the quadrants of sleeps bouts.
4. Discussion The objectives of this study were to determine whether adult zebrafish in a state of quiescence, or sleep, show rapid eye movements and irregular respiration, two key components of REM sleep. In order to measure this, we used custom written software, recording zebrafish under different light conditions, tracking the behavior, and measuring both eye and respiratory movements.
194
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
Our data showed a clear difference in activity levels between night and day, confirming the normal circadian and sleep behavior of the subjects as described in previous studies [9,10]. To our surprise, we show that during sleep at night with lightsoff, adult zebrafish exhibit no eye movements of any kind. During bouts of immobility while lights are on, day and night, there are plentiful visible and measurable eye movements. No difference in frequency was found between the first 6 s of a bout and the remainder of the bout during day, however, at night the frequencies are lower during the first 6 s. That is, during behaviorally defined wake (first 6 s of immobility bout), there is no difference in frequency of eye movements between day or night whereas during sleep (7th second and above of immobility bout), at night with lights on the eye movement frequency is increased. A comparison of eye movement frequencies, regardless of behavioral state, between night and day showed no difference. These eye movements do not appear to be of the same kind as the rapid eye movements normally exhibited during REM sleep. We speculate that eye movements may be linked to the ambient light levels, that when in complete darkness, there is no need to visually assess the environment and so there will be no eye movements. Possibly, this constant visual input explains the very low level of sleep during lights-on. Importantly, at least larval zebrafish drastically reduce their photosensitivity at night [17]. To our knowledge the only other report of eye movements during immobility in fish was performed in parrot fish were they were reported to have slow eye movements [18]; that study was done under visible light only. We show that respiratory rates drop significantly after 6 s of immobility and are further lowered in the last two quadrants of sleep bouts under normal day–night cycle. Also, under normal day–night cycle peak-to-peak variance of breathing is significantly higher after the 6 s mark. The breathing is therefore slower but more irregular during sleep than wake. The lower respiratory rates in the last two quadrants could simply be caused by prolonged immobility and the ensuing drop in metabolism. These findings suggest that immobility with lights-on is a physiologically different state from immobility with lights-off. REM sleep components may still exists in zebrafish even though sleep-related rapid eye movements, which can be seen as a proxy for REM, do not. Importantly, all neurotransmitters and brainstem neural circuits generating REM in mammals are conserved in zebrafish [12]. Further research should be done on other REM components. For example, measuring the EKG in freely moving fish, or doing electrophysiological measurements in REM active neurons on head-fixed larvae, using fictive movements as an indication of behavioral state [19,20]. And by assessing sleep-related atonia, previously reported in catfish [21]. Possibly, REM components could be more easily detected in young fish or larvae. Young mammals (and even drosophila) have a greater ratio of REM to NREM sleep (and more sleep in total) than adults and therefore it might be possible to more easily identify REM components in young fish [10,16,22–24]. Previous studies have established a definition of zebrafish sleep as a period after the fish has been immobile longer than 6 s where the velocity of the fish at no point exceeds 1.5 cm/s [7–10,13]. This study focused on two of the phasic REM components, rapid eye movements and irregular breathing, other aspects of REM sleep or REM-like stages have yet to be examined. None of the results of this study indicate any additional stages of sleep in zebrafish beyond what has been previously defined whereas the findings
further solidify the notion that there is a physiological change in adult zebrafish after approximately 6 s of immobility. Acknowledgements Supported by the Icelandic Centre for Research grant no. 214002. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.bbr.2015.05.017 References [1] Hendricks JC, Finn SM, Panckeri KA, Chavkin J, Williams JA, Sehgal A, et al. Rest in Drosophylia is a sleep like state. Neuron 2000;25:129–38. [2] Campbell SS, Tobler I. Animal sleep: a review of sleep duration across phylogeny. Neurosci Biobehav Rev 1984;8:269–300. [3] Hendricks JC, Sehgal A, Pack AI. The need for a simple model to understand sleep. Prog Neurobiol 2000;61(4):339–51. [4] Shaw PJ, Cirelli C, Greenspan RJ, Tononi G. Correlates of sleep and waking in Drosophila melanogaster. Science 2000;287(5459):1834–7. [5] Yokogawa T, Marin W, Faraco J, Pezeron G, Appelbaum L, Zhang J, et al. Characterization of sleep in zebrafish and insomnia in hypocretin receptor mutants. PLoS Biol 2007;5(10):2379–97. [6] Zhdanova IV, Wang SY, Leclair OU, Danilova NP. Melatonin promotes sleep-like state in zebrafish. Brain Res 2001;903(1–2):263–8. [7] Rihel J, Prober DA, Arvanites A, Lam K, Zimmerman S, Jang S, et al. Zebrafish behavioral profiling links drugs to biological targets and rest/wake regulation. Science 2010;327(5963):348–51. [8] Rihel J, Prober DA, Schier AF. Monitoring sleep and arousal in zebrafish. Methods Cell Biol 2010;100:281–94. [9] Sigurgeirsson B, Thorsteinsson H, Sigmundsdottir S, Lieder R, Sveinsdottir HS, Sigurjonsson OE, et al. Sleep–wake dynamics under extended light and extended dark conditions in adult zebrafish. Behav Brain Res 2013;256:377–90, http://dx.doi.org/10.1016/j.bbr.2013.08.032. [10] Sorribes A, Thorsteinsson H, Arnardottir H, Johannesdottir I, Sigurgeirsson B, de Polavieja GG, et al. The ontogeny of sleep–wake cycles in zebrafish: a comparison to humans. Front Neural Circuits 2013;7:178, http://dx.doi.org/10.3389/fncir.2013.00178. [11] Zhdanova IV. Sleep in zebrafish. Zebrafish 2006;3(2):225. [12] Panula P, Chen YC, Priyadarshini M, Kudo H, Semenova S, Sundvik M, et al. The comparative neuroanatomy and neurochemistry of zebrafish CNS systems of relevance to human neuropsychiatric diseases. Neurobiol Dis 2010;40(1):46–57. [13] Prober DA, Rihel J, Onah AA, Sung RJ, Schier AF. Hypocretin/orexin overexpression induces an insomnia-like phenotype in zebrafish. J Neurosci 2006;26(51):13400–10. [14] Sigurgeirsson B, Thorsteinsson H, Arnardottir H, Johannesdottir IT, Karlsson KA. Effects of modafinil on sleep–wake cycles in larval zebrafish. Zebrafish 2011;8(3):133–40, http://dx.doi.org/10.1089/zeb.2011.0708. [15] Siegel J. REM sleep. In: Kryger MH, Roth T, Dement WC, editors. Principles and Practice of Sleep Medicine. Philadelphia: Saunders; 2005. p. 120–35. [16] Siegel JM. Phylogeny and the function of REM sleep. Behav Brain Res 1995;69:29–34. [17] Emran F, Rihel J, Adolph AR, Dowling JE. Zebrafish larvae lose vision at night. Proc Natl Acad Sci U S A 2010;107(13):6034–9, http://dx.doi.org/10.1073/pnas.0914718107. [18] Tauber E, Weitzman E. Eye movements during behavioral inactivity in certain Bermuda feed fish. Commun Behav Biol A 1969;3:131–5. [19] Fetcho JR. The utility of zebrafish for studies of the comparative biology of motor systems. J Exp Zool B Mol Dev Evol 2007;308(5):550–62. [20] Masino MA, Fetcho JR. Fictive swimming motor patterns in wild type and mutant larval zebrafish. J Neurophysiol 2005;93(6):3177–88. [21] Titkov ES. Diurnal periodicity of wakefulness and sleep in the catfish Ictalurus nebulosus. Zh Evol Biokhim Fiziol 1976;12(4):335–40. [22] Jouvet-Mounier D, Astic L, Lacote D. Ontogenesis of the states of sleep in rat, cat, and guinea pig during the first postnatal month. Dev Psychobiol 1970;2:216–39. [23] Roffwarg HP, Muzio JN, Dement WC. Ontogenetic development of the human sleep-dream cycle. Science 1966;152:604–19. [24] Shaw P, Ocorr K, Bodmer R, Oldham S. Drosophila aging 2006/2007. Exp Gerontol 2008;43(1):5–10, http://dx.doi.org/10.1016/j.exger.2007.10.008.
Contents lists available at ScienceDirect
Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr
Short Communication
Absence of rapid eye movements during sleep in adult zebrafish B.B. Árnason a , H. Þorsteinsson a,b , K.Æ. Karlsson a,b,∗ a b
School of Science and Engineering, Reykjavik University, Reykjavik, Iceland 3Z Pharmaceuticals, Reykjavik, Iceland
h i g h l i g h t s • • • • •
Zebrafish are novel yet important model in sleep research. Sleep in zebrafish is measured using behavioral assays that cannot differentiate between REM and NREM sleep. In current study we use image analysis to measure two key REM components: rapid eye movements and respiratory rate. We show that there are no sleep related eye movements but during sleep respiration becomes slower and less regular. We find no evidence for REM–NREM cycles in zebrafish.
a r t i c l e
i n f o
Article history: Received 6 January 2015 Received in revised form 9 May 2015 Accepted 12 May 2015 Available online 20 May 2015 Keywords: REM Sleep Zebrafish Computer vision Motion tracking Activity detection
a b s t r a c t Sleep is not a uniform phenomenon, but is organized in alternating, fundamentally different states, rapid eye movement sleep and non-rapid eye movement sleep. Zebrafish (Danio rerio) have recently emerged as an excellent model for sleep research. Zebrafish are well characterized in terms of development, neurobiology and genetics. Moreover, there are many experimental tools not easily applied in mammalian models that can be readily applied to zebrafish, making them a valuable additional animal model for sleep research. Sleep in zebrafish is defined behaviorally and exhibits the hallmarks of mammalian sleep (e.g. sleep homeostasis and pressure). To our knowledge no attempts have been made to discern if sleep in zebrafish entails alternations of REM–NREM sleep cycles which are critical for further development of the model. In the current experiment we quantify two key REM sleep components, rapid eye movements and respiratory rates, across sleep–wake cycles. We find no sleep-related rapid eye movements. During sleep respiratory rates, however, are reduced and become less regular, further establishing that the behavioral definition used truly captures a change in the fish’s physiology. We thus fail to find evidence for REM–NREM sleep cycles in zebrafish but demonstrate a physiological change that occurs concomitantly with the previously defined behavioral state of sleep. We do not rule out that other phasic REM components (e.g. atonia, cardiac arrhythmias, myoclonic twitches or desynchronized EEG) are coherently expressed during sleep but we conclude that adult zebrafish do not have REM-sleep-related rapid eye movements. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The necessity of using highly prolific genetically tractable animals in sleep research, many of whom do not have differentiated EEG, has resulted in the re-adoption of the older behavioral definition of sleep [1]. According to a behavioral definition a state can be considered sleep if: (i) it is characterized by an absence of voluntary movement; (ii) it is spontaneous, occurring with a circadian rhythm; (iii) it is reversible; (iv) it is characterized by
∗ Corresponding author at: School of Science and Engineering, Reykjavik University, Reykjavik, Iceland. Tel.: +354 8256467. E-mail addresses: [email protected], [email protected] (K.Æ. Karlsson). http://dx.doi.org/10.1016/j.bbr.2015.05.017 0166-4328/© 2015 Elsevier B.V. All rights reserved.
a species-specific posture; (v) it includes an increased arousal threshold; (vi) and is under homeostatic control [2]. Behavioral definition of sleep has been successfully applied to recent, but important, animal models in sleep research, such as fruit flies (Drosophila melanogaster) [3,4] and Zebrafish (Danio rerio) [5,6]. Both adult and larval zebrafish represent an important model organism with established applications in sleep research [7–11]. Since differentiated EEG has not been established in zebrafish, their use as a model in sleep research is limited to particular behavioral definitions of sleep for larvae [6,7,11] and adults [5]. Zebrafish sleep is similar to mammalian sleep as shown by a comparison of behavioral, physiological, and pharmacological characteristics [5,7,11] and their neurobiology is comparable to mammals [12]. Zebrafish behavior has been shown to follow a circadian rhythm
190
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
and to be regulated by a homeostatic mechanism similar to that in mammals [5,9–11]. Studies using both electrical stimulation and light to interrupt sleep show that following 6 s of immobility, adult zebrafish do exhibit homeostasis and sleep pressure, albeit weaker than in mammals [5,9,13]. Furthermore, drugs which affect mammalian sleep have been shown to similarly affect sleep in larval zebrafish [7,14]. Importantly, the current behavioral definition of sleep in zebrafish does not allow a differentiation between REM and NREM sleep states. The fact that REM or REM-like sleep has been shown to exist in both ectothermic and endothermic vertebrates [15,16] warrants a thorough study of the presence of REM or REM-like sleep in fish. Establishing the existence of REM in either adult or larval zebrafish would be highly valuable for sleep research. First, it would open various realms of research on the neural substrates of sleep in general, where full advantage of the unique research tools applicable to zebrafish could be taken. Second, it would allow for novel applications in drug screening where one could screen for drugs with specific effects on either REM or NREM sleep. In the current study we therefore used image analysis in order to determine whether or not adult zebrafish express two of the phasic REM-sleep components, rapid eye movements and/or irregular respiration. 2. Methods 2.1. Fish A total of 19 adult zebrafish of the AB strain, 4–6 months of age, were used in the study. Fish were fed twice a day on a variable diet of TetraMin flakes (Tetra Holding GmbH, Melle, Germany) and Adult Zebrafish Complete Diet (Zeigler Bros, USA). The fish were kept in a 14:10 light:dark cycle in a 3 or 10 L multi tank constant flow system (Aquatic Habitats, Apopka, FL, USA). Water temperature was held at a constant 28.5 ◦ C and replaced at a rate of 10% per day. 2.2. Data acquisition All tracking trials and recordings were done in a 75 L aquarium containing two identical small chambers made from opaque black Plexiglas, isolating the subject within an environment of 14 cm (wide) × 4.5 cm (deep) × 14 cm (high), but with the same light–dark cycle and temperature controls as described above. The aquarium was illuminated with two infrared lights of 0 lx (darkphase), and a fluorescent light of 255 lx (light-phase) positioned above the chamber. Water was circulated from a 50 L reservoir tank at the rate of 2.3 L/min. Thermosensor, filter, air diffuser, and water pump were placed in the aquarium. Behavior was tracked at 5 and 30 Hz (baseline and bout analysis, respectively) with a Basler ace acA2000-50gmNIR infrared camera (Basler, Germany), using a 50 mm CCTV Pentax lens (Pentax, Germany). Data were acquired with custom written programs in MATLAB R2013a (The MathWorks Inc., Natick, MA, 2013). All test subjects were placed in the acclimation chamber for 24 h. Fish were placed in the acclimation chamber between 11:00 and 12:00; the 24 h recording commenced at 12:00 the following day. Recordings were performed during day and night, with lights-on and lights-off (lights-on at 08:00; lights-off at 22:00). All procedures in this study were carried out in strict compliance with the regulations of, and approved by, the National Bioethics Committee of Iceland (regulation 279/2002); permit issued to KÆK 19th May 2008. 2.3. Data processing First, velocity was plotted across 24 h to verify expression of normal sleep behavior. Second, we parsed the data into bouts of
Table 1 Overview of experimental conditions. Subject/trial 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Acclimation √ √ √ √ √ √ √ √ √ √ √ √ √ √
Baseline √ √ √ √ √ √ √ √
Day/light
√ √ √ √ √
Night/dark √ √ √ √ √ √ √ √ √
Night/light
√ √ √ √ √
sleep and wakefulness. Based on previous studies of adult zebrafish [5] fish were considered immobile if their velocity did not exceed 1.5 cm/s and periods of immobility exceeding 6 s were categorized as sleep [5,9,10]; i.e. the 7th second and above were counted as sleep. Thus, a single sleep bout is the time from the 6th second of immobility until movement exceeds 1.5 cm/s. Using custom written MATLAB routines periods of immobility were obtained from each fish representing all recording categories (day/night and lights-off/on). Third, using videos from immobile fish only, two region-of-interest videos were produced: one from around the eye (to produce eye-movement data) and one around the head (to produce respiratory-movement data). Next, activity detection module was run on the cropped videos, followed by low pass (0.1 Hz) filtering on the respiratory-movement data (to remove high frequency noise) and high pass (5.0 Hz) filtering on the eye-movement data (to remove artifacts caused by respiratory movements). Fig. 1A–C depicts the protocol for detecting eye-movements and Fig. 1D–F depicts the protocol for detecting respiratory movements. 2.4. Data analysis Data for analysis were randomly selected from throughout the night from each set of conditions: night with lights-off: 20 random samples from each subject with a duration of at least 7 s (to include at least one second of sleep), a total of 100 bouts. Day with lights-on: a total of 18 bouts with duration of at least 7 s. Night with lights-on: a total of 48 bouts with duration of at least 7 s. Eye and respiratory movements were then compared between bouts of immobility or between bouts of sleep and wakefulness. Table 1 shows the number of fish in each condition. 2.5. Statistics Mean frequencies were compared using Kolmogorov–Smirnoff test of equal means, changes in peak-to-peak distance were compared using Levene’s test of equal variances and comparison of mean rates within sleep-bout quadrants were performed using the Friedman test. All data are presented with their standard error of measurement. 3. Results 3.1. Eye movements Fig. 2A depicts representative data for eye movements. Peaks above a threshold value of 8 were found to accurately represent eye movements (verified by three independent reviewers from video data). No eye movements were registered in any sleep bout at night with lights-off or during day with lights-off. Eye movements were
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
191
Fig. 1. Overview of activity detection. Activity detection was performed on stationary videos only and based on frame-differencing where a video frame (M) at tx is subtracted from a frame at a later point in time t(x+1) and the absolute value is taken of the result. Mt(x +1) = |Mt(x +1) − Mtx |. This difference frame (M) is the absolute difference in pixel values between t(x+1) and tx . (A)–(C) schematic overview of eye movement detection. (D)–(F) schematic overview of respiratory movement detection.
192
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
Fig. 2. Representative eye- and respiratory movements. (A) Green line represents sleep at night during lights-off. Red and blue lines represent sleep and day during lights-on and night during lights-on, respectively. No eye movements were detected during sleep with lights-off during day or night. (B) Representative recording of respiratory movement during night at lights-off. All tracks are derived from the same fish.
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
193
registered in every sleep bout while lights were on, both during day and night. Solid lines show the activity detection data and registered eye movements are represented with dots (for the following analysis n = 14). During day with lights-on eye movement frequency during the first 6 s was 0.6481 ± 0.098 Hz and 0.5025 ± 0.093 during the remainder of the bouts. A Kolmogorov–Smirnov test revealed no significant difference in eye movement frequency (p = 0.7088). During night with lights-on eye movement frequency during the first 6 s was 0.4271 ± 0.051 Hz compared to 0.6132 ± 0.086 during the remainder of the bouts. A Kolmogorov–Smirnov test revealed a significant difference in eye movement frequency (p = 0.0070), the frequency being lower during the first 6 s. When the whole bout of immobility was combined (i.e. combining sleep and wake bouts) eye movement frequency did not differ between night and day. The frequency of eye movements during day with lights-on was 0.6258 ± 0.075 Hz and during night with lights-on 0.5499 ± 0.055 Hz. A Kolmogorov–Smirnov test revealed no significant difference in eye movement frequency (p = 0.7854). Also, a Kolmogorov–Smirnov test, comparing eye movement frequency during sleep bouts, (immobility beyond the first 6 s) for day and night with lights on, showed no difference (p = 0.5685). Furthermore, visual observation and spectrogram analysis of the eye movement signals revealed that the registered eye movements did not come in bursts of rapid saccades but only as singular rotations of the eyes. Average eye movement frequency (during lights-on only) was 0.7022 ± 0.0497 Hz.
3.2. Respiration Fig. 2B shows representative data for respiration rates. Data are represented as average pixel value change per video frame. Peaks in the data signals were automatically registered and peaks above a threshold value of 4 were found to accurately represent respiratory movements (verified by three independent reviewers from video data) (for following analysis n = 14). Fig. 3A shows a comparison of respiratory rates during night with lights-off for the first 6 s of an immobility bout versus the remainder of the bout (sleep). The mean respiratory rate during the first 6 s was 0.4883 ± 0.027 Hz and for the remainder 0.4127 ± 0.034 Hz. A Kolmogorov–Smirnov test showed a significant difference (p = 0.0018) where the respiratory rate during the first 6 s is higher than during the remainder of the bout. The same test was used on respiratory rate data from sleep bouts during day with lights-on and night with lights-on, both cases showed no significant difference (p = 0.4255 and p = 0.6521). Fig. 3B shows a comparison of the peak-to-peak intervals of breaths during night with lights-off between the first 6 s of an immobility bout versus the remainder of the bout (sleep). A Levene’s test of equal variances showed a significant difference between periods (F(1,495) = 6.5702, p < 0.05) where the variance during the first 6 s is lower than during the remainder of the bout. Fig. 3C shows a comparison of respiratory rates during night with lights-off between sleep bout quadrants. Mean respiratory rates for quadrants 1–4 were, respectively, 0.5428 ± 0.042, 0.4489 ± 0.034, 4480 ± 0.034 and 0.4185 ± 0.034 Hz. A Friedman test revealed a significant difference between groups (2 (3,297) = 17.0579, p < 0.05). A multiple comparison test, showed that the third and fourth quadrants were significantly different from the first with respiratory rates lower during the last two quadrants. Average respiratory frequency across all conditions was 0.6396 ± 0.0368 Hz.
Fig. 3. Respiratory parameters at night during lights-off. (A) Comparison of respiratory frequency between first 6 s (wake) and remainder (sleep) of a bout of immobility. (B) Comparison of peak-to-peak variability between first 6 s (wake) and remainder (sleep) of a bout of immobility. (C) Comparison of respiratory frequency between the quadrants of sleeps bouts.
4. Discussion The objectives of this study were to determine whether adult zebrafish in a state of quiescence, or sleep, show rapid eye movements and irregular respiration, two key components of REM sleep. In order to measure this, we used custom written software, recording zebrafish under different light conditions, tracking the behavior, and measuring both eye and respiratory movements.
194
B.B. Árnason et al. / Behavioural Brain Research 291 (2015) 189–194
Our data showed a clear difference in activity levels between night and day, confirming the normal circadian and sleep behavior of the subjects as described in previous studies [9,10]. To our surprise, we show that during sleep at night with lightsoff, adult zebrafish exhibit no eye movements of any kind. During bouts of immobility while lights are on, day and night, there are plentiful visible and measurable eye movements. No difference in frequency was found between the first 6 s of a bout and the remainder of the bout during day, however, at night the frequencies are lower during the first 6 s. That is, during behaviorally defined wake (first 6 s of immobility bout), there is no difference in frequency of eye movements between day or night whereas during sleep (7th second and above of immobility bout), at night with lights on the eye movement frequency is increased. A comparison of eye movement frequencies, regardless of behavioral state, between night and day showed no difference. These eye movements do not appear to be of the same kind as the rapid eye movements normally exhibited during REM sleep. We speculate that eye movements may be linked to the ambient light levels, that when in complete darkness, there is no need to visually assess the environment and so there will be no eye movements. Possibly, this constant visual input explains the very low level of sleep during lights-on. Importantly, at least larval zebrafish drastically reduce their photosensitivity at night [17]. To our knowledge the only other report of eye movements during immobility in fish was performed in parrot fish were they were reported to have slow eye movements [18]; that study was done under visible light only. We show that respiratory rates drop significantly after 6 s of immobility and are further lowered in the last two quadrants of sleep bouts under normal day–night cycle. Also, under normal day–night cycle peak-to-peak variance of breathing is significantly higher after the 6 s mark. The breathing is therefore slower but more irregular during sleep than wake. The lower respiratory rates in the last two quadrants could simply be caused by prolonged immobility and the ensuing drop in metabolism. These findings suggest that immobility with lights-on is a physiologically different state from immobility with lights-off. REM sleep components may still exists in zebrafish even though sleep-related rapid eye movements, which can be seen as a proxy for REM, do not. Importantly, all neurotransmitters and brainstem neural circuits generating REM in mammals are conserved in zebrafish [12]. Further research should be done on other REM components. For example, measuring the EKG in freely moving fish, or doing electrophysiological measurements in REM active neurons on head-fixed larvae, using fictive movements as an indication of behavioral state [19,20]. And by assessing sleep-related atonia, previously reported in catfish [21]. Possibly, REM components could be more easily detected in young fish or larvae. Young mammals (and even drosophila) have a greater ratio of REM to NREM sleep (and more sleep in total) than adults and therefore it might be possible to more easily identify REM components in young fish [10,16,22–24]. Previous studies have established a definition of zebrafish sleep as a period after the fish has been immobile longer than 6 s where the velocity of the fish at no point exceeds 1.5 cm/s [7–10,13]. This study focused on two of the phasic REM components, rapid eye movements and irregular breathing, other aspects of REM sleep or REM-like stages have yet to be examined. None of the results of this study indicate any additional stages of sleep in zebrafish beyond what has been previously defined whereas the findings
further solidify the notion that there is a physiological change in adult zebrafish after approximately 6 s of immobility. Acknowledgements Supported by the Icelandic Centre for Research grant no. 214002. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.bbr.2015.05.017 References [1] Hendricks JC, Finn SM, Panckeri KA, Chavkin J, Williams JA, Sehgal A, et al. Rest in Drosophylia is a sleep like state. Neuron 2000;25:129–38. [2] Campbell SS, Tobler I. Animal sleep: a review of sleep duration across phylogeny. Neurosci Biobehav Rev 1984;8:269–300. [3] Hendricks JC, Sehgal A, Pack AI. The need for a simple model to understand sleep. Prog Neurobiol 2000;61(4):339–51. [4] Shaw PJ, Cirelli C, Greenspan RJ, Tononi G. Correlates of sleep and waking in Drosophila melanogaster. Science 2000;287(5459):1834–7. [5] Yokogawa T, Marin W, Faraco J, Pezeron G, Appelbaum L, Zhang J, et al. Characterization of sleep in zebrafish and insomnia in hypocretin receptor mutants. PLoS Biol 2007;5(10):2379–97. [6] Zhdanova IV, Wang SY, Leclair OU, Danilova NP. Melatonin promotes sleep-like state in zebrafish. Brain Res 2001;903(1–2):263–8. [7] Rihel J, Prober DA, Arvanites A, Lam K, Zimmerman S, Jang S, et al. Zebrafish behavioral profiling links drugs to biological targets and rest/wake regulation. Science 2010;327(5963):348–51. [8] Rihel J, Prober DA, Schier AF. Monitoring sleep and arousal in zebrafish. Methods Cell Biol 2010;100:281–94. [9] Sigurgeirsson B, Thorsteinsson H, Sigmundsdottir S, Lieder R, Sveinsdottir HS, Sigurjonsson OE, et al. Sleep–wake dynamics under extended light and extended dark conditions in adult zebrafish. Behav Brain Res 2013;256:377–90, http://dx.doi.org/10.1016/j.bbr.2013.08.032. [10] Sorribes A, Thorsteinsson H, Arnardottir H, Johannesdottir I, Sigurgeirsson B, de Polavieja GG, et al. The ontogeny of sleep–wake cycles in zebrafish: a comparison to humans. Front Neural Circuits 2013;7:178, http://dx.doi.org/10.3389/fncir.2013.00178. [11] Zhdanova IV. Sleep in zebrafish. Zebrafish 2006;3(2):225. [12] Panula P, Chen YC, Priyadarshini M, Kudo H, Semenova S, Sundvik M, et al. The comparative neuroanatomy and neurochemistry of zebrafish CNS systems of relevance to human neuropsychiatric diseases. Neurobiol Dis 2010;40(1):46–57. [13] Prober DA, Rihel J, Onah AA, Sung RJ, Schier AF. Hypocretin/orexin overexpression induces an insomnia-like phenotype in zebrafish. J Neurosci 2006;26(51):13400–10. [14] Sigurgeirsson B, Thorsteinsson H, Arnardottir H, Johannesdottir IT, Karlsson KA. Effects of modafinil on sleep–wake cycles in larval zebrafish. Zebrafish 2011;8(3):133–40, http://dx.doi.org/10.1089/zeb.2011.0708. [15] Siegel J. REM sleep. In: Kryger MH, Roth T, Dement WC, editors. Principles and Practice of Sleep Medicine. Philadelphia: Saunders; 2005. p. 120–35. [16] Siegel JM. Phylogeny and the function of REM sleep. Behav Brain Res 1995;69:29–34. [17] Emran F, Rihel J, Adolph AR, Dowling JE. Zebrafish larvae lose vision at night. Proc Natl Acad Sci U S A 2010;107(13):6034–9, http://dx.doi.org/10.1073/pnas.0914718107. [18] Tauber E, Weitzman E. Eye movements during behavioral inactivity in certain Bermuda feed fish. Commun Behav Biol A 1969;3:131–5. [19] Fetcho JR. The utility of zebrafish for studies of the comparative biology of motor systems. J Exp Zool B Mol Dev Evol 2007;308(5):550–62. [20] Masino MA, Fetcho JR. Fictive swimming motor patterns in wild type and mutant larval zebrafish. J Neurophysiol 2005;93(6):3177–88. [21] Titkov ES. Diurnal periodicity of wakefulness and sleep in the catfish Ictalurus nebulosus. Zh Evol Biokhim Fiziol 1976;12(4):335–40. [22] Jouvet-Mounier D, Astic L, Lacote D. Ontogenesis of the states of sleep in rat, cat, and guinea pig during the first postnatal month. Dev Psychobiol 1970;2:216–39. [23] Roffwarg HP, Muzio JN, Dement WC. Ontogenetic development of the human sleep-dream cycle. Science 1966;152:604–19. [24] Shaw P, Ocorr K, Bodmer R, Oldham S. Drosophila aging 2006/2007. Exp Gerontol 2008;43(1):5–10, http://dx.doi.org/10.1016/j.exger.2007.10.008.
