1. Introduction

2. Methods to study animals sleep and dreams

The standard method used to study sleep is called polysomnography. It uses electrode patches placed on specific parts of the animal's head and body to record electrical activity on a polygraph (resulting in readable data that looks like scribbled lines).

It records simultaneously the electrical activity of the brain, the eye movements, the muscles tone, the heart activity and the respiration of the animal. The electroencephalogram (EEG) is a measurement of bioelectric brain activity, by the mean of electrodes placed on the scalp. When looking at EEG levels, sleep researchers can clearly see distinctions between the different stages of sleep, as well as when the initial sleep onset actually occurs. The electro-oculography (or EOG) is the standard method used to measure eye movements during sleep. The front of the eye (the cornea) is electrically positive compared to the back (the retina). Therefore, when the eye moves the change in voltage can be recorded by electrodes placed on the sleeper’s face and recorded on the polygraph. The muscle activity is measured by a method called the electromyogram (or EMG). Muscles emit electrical potentials when they move, and that electricity is also detected by electrodes and recorded on the polygraph. EMG electrodes are usually placed over the animal’s chin muscles. Similarly, the electrocardiogram (ECG) records from the body surface and registers the differences in electrical potential generated by the heart.

These systems involve either tethered systems that are restrictive and heavy on the animal or wireless systems using transponders that are large relative to the animal and surgically invasive for implantation, thus natural behavior/activity might be altered. But a study (Zielinski, Mark R., et al, 2013) using a novel telemetric system to measure polysomnography biopotentials in freely moving mice showed good results in analysis of sleep architecture.

Fig1

The mouse head cap containing EEG/EMG electrodes is connected to a telemetry transponder (4) encapsulated in a protective covering (2) by short cables contained in a lightweight protective sheath (1). This system is counterbalanced (3), rotates on an O-ring that swivels 360 degrees and slides horizontally between two contained ends, and possesses an additional swivel allowing the maximal range of movement within the mouse housing (5).

3. When do dreams happen during sleep?

3.1 Two states of sleep in mammals

This type of studies - made since 1953 on humans (Aserinski, Kleitman and Dement, USA) and since 1959 on cats (Jouvet, France) – have proved that there is two distinct stages of sleep: the slow sleep or nREM sleep (non Rapid Eyes Movement) and the paradoxical sleep or REM sleep, defined by their electrophysiological characteristics.

Table 1: Physiological characteristics of the awake, nREM and REM sleep states

These phases alternate cyclically in a highly structured pattern throughout the night. Those cycles last about 90 min in human and 28 min in cats. Jouvet (1994) has proved the existence of a duty ratio, calculated by dividing the average duration of sleep cycles by the average duration of REM sleep episodes. It is close to 4 in the vast majority of species (except rabbits and monkeys).

3.2 Dreams and REM sleep

At the end of the nineteenth century, a scientist first made the assumption that rapid eye phenomena were due to feelings and representations of dreams (Manaseina, 1899). Then, studies found that by waking his sleeping human patients during REM sleep and interviewing them, they remember their dreams much clearly. So scientists made the hypothesis that most dreams occur during REM sleep (Dement, William, and Nathaniel Kleitman, 1957).

However, the forced awakening of patients during nREM sleep gives equivocal results. Memorizing dreams is not demonstrated but in 8 to 30% of the cases, elementary mental activity defined as "dream" was found. (Foulkes, 1990)

In animals, it has been experimented on cats by abolishing muscle atony in REM sleep and then observing the behavior of the animal when sleeping. Under these conditions, a cat exhibits during REM sleep a behavior typical of game or hunting and behaviors of aggression or defence that can be related to dreams.

Fig2: Localization of the Locus Coeruleus alpha, responsible of the muscle atonia

Jouvet’ experiences has proved that The destruction of the locus coeruleus alpha prevents muscle atonia observed during REM sleep. In these animals, the cortical synchronization, rapid eye movements and the activity of the pons attest to a REM sleep. However animals simultaneously present motor behavior instead of the normal atony. The behaviors observed are characteristic of the species: predatory attack (like round back, erection of hairs), fear, grooming,exploration. All of these behaviors are never referred to an element of the environment and the animal is largely insensitive to any external stimulus. The animals"live" their dream.

3.3 Pontic origin of the REM sleep

Michel Jouvet (1961), French researcher, found out the origin of the REM sleep by experimenting on “pontic cats”,it mean cats whose brain stem (cerebral trunk) are connected to electrodes.

Fig 3: Location of sections used for a "Pontic cat" (section A)

If you remove the part of the brain located forward of the pons in a cat, we still observe periodic symptoms of REM sleep. Such animals can survive for several months if the following conditions are ensured artificially: thermal balance, nutritional and hydromineral supply. The animal is then subject, with clockwork regularity, to periods of postural weakness. These periods of atony are accompanied by rapid eye movements. These periods of atony are accompanied by rapid eye movements, in relationship with an activity ponto geniculo occipital (PGO). We can conclude that the REM activity arises from the pons. The REM basic mechanisms are responsible for two functions which are complementary to each other. First, they involve an endogenous system of excitations of the brain through the PGO activity. This stimulation causes the excitation of the sensory systems (mostly visual) and motor systems (pyramidal neurons of the motor area). Thus descending signals will respond to these stimuli and go to the spinal cord to trigger actions and behaviors. This is to prevent the motor activity that a second mechanism must arise: a powerful descending inhibition, motor neurons of the spinal cord. Thus, the dreamer is paralyzed and cannot move. Also, the topography of neurons (cholinergic likely) that constitute the endogenous generator of the dream of PGO activity was well defined. It is located in the pontine reticular formation. We also know the paths that lead the PGO activity in the nuclei eye motors (where it triggers the rapid eye movements.

4. What animals dream?

If we admit that REM sleep is accompanied by dreams, all animals that have REM sleep, should also have dreams. Similarly, the absence of REM sleep in the other animals would preclude the existence of dreams (Tafti, 1996). However, dreams use stored elements, both in the short-term and long term memory. Even if the brain structures necessary for memory, exist in many of the animals we know very little about their memory capacity. Nevertheless, we must admit that the vast majority of animals (except maybe insects and fish) probably has traces stored, necessary for the realization of dreams.

Sleep of a hundred of species of mammals (they represent only a small part of living species, they are the most studied), but also birds, reptiles, fishes, insects, has already been studied. A lot of studies have been carried out on over a hundred of species of mammals, the most studied ones, but also on birds, reptiles, fishes and even insects.

4.1 Problem in the definition of sleep states

The problem for defining sleep in general or the phases of nREM and REM sleep, is that standardized physiological criteria applicable for all species don't exist. The sleep state can be defined by the absence or reduction of activity and response to external stimuli, a specific body posture and an easily reversible state. The latter is crucial in order not to confuse that state with coma, hibernation or hypothermia. The behavioural signs of REM sleep are a partial or total atonia (especially in the neck), rapid eyes movements and sudden movements of the extremities.

4.2 Sleep states in warm blooded vertebrates

4.2.1 Mammals

A crucial point in evolution is the divergence between the phyla of live-bearing mammals (marsupials and placental) and egg-laying mammals (monotremes). Oviparous mammals, echidna and platypus, were first thought not to have REM. But, Siegel had them re-tested with newer technology and it appeared that they presented a kind of REM as well (Siegel, 1996). The echidna presents a less pronounced brain wave pattern, but had these tiny REM-like bursts more often. It seemed that the animal was having micro-dreaming throughout the night. The platypus showed many REM indicators too, except that there wasn't the low voltage REM characteristic of most mammals.

In all land viviparous mammals studied, REM sleep exists. It is very easy to recognize by its classic criteria: activation of the EEG, suppression of muscle tone, rapid eye movements.

Modifications are implemented in marine mammals (Moukhametov, 1990), the largest of these being the need to rise to the surface to breathe. In dolphins and in some other cetaceans, sleep is unilateral. While one cerebral hemisphere shows typical signs of REM sleep, the other hemisphere is awake. As breathing is voluntary in these animals, this feature allows them to simultaneously perform two vital functions: sleep and breathe. The dolphin, which never stops to swim even while sleeping, do not seem to have REM sleep. This is an exception because other marine mammals studied, such as seals and sea lions, have a REM sleep comparable to terrestrial mammals. The possible lack of REM sleep in the cetacean is one of the biggest puzzles of the phylogeny of sleep.

4.2.2 Birds

A dozen bird species was studied with polygraph methods (Klein, Michel, Jouvet 1964). There are notable EEG differences between the awake state (rapid activity and low voltage) and sleep (slow activity high voltage). During the behavioral sleep, there is indisputable evidence of the periodic appearance of REM sleep. It is characterized, as in mammals, by activation of the EEG by rapid eye movements and a decrease, sometimes not total, of the muscle tone. Complete atony of the neck, one of the mammals own characteristics, is rarely seen in birds, except the goose. These episodes are very brief (10 to 20 seconds) but are repeated often during the night. There are significant differences in the organization of sleep in different species. During migration, the albatross or swifts can fly nonstop for long periods. It is possible that these birds have a unihemispheric sleep, like dolphins.

4.3 Sleep like state in invertebrates

Sleep research in invertebrates started with bees, cockroaches and scorpions (Kaiser, 1983; Tobler, 1983). More recently, Drosophila (Hendricks, 2000; Shaw, 2000), and the roundworm C. elegans were discovered as promising models for sleep research (Raizen, 2008). In the beginning, the field was rather preoccupied with proving the existence of sleep in invertebrate species. The research strongly contributed to the implementation of general criteria for the definition of sleep. As no typical mammalian EEG signal is present in these species, the identification of sleep-like states in invertebrates relies primarily on behavioral signs like inactivity and the presence of a specific body posture, an increased threshold to arousing stimulation, as well as the demonstration of a rebound in the sleep-like state that occurs as a consequence of experimental sleep deprivation (Vorster, Born, 2015). In fact, invertebrates studied so far like honey bees and flies, with the exception of crayfish, seem to lack synchronous neuronal activity during sleep (Mendoza-Angeles et al., 2007, 2010). Moreover, sleep intensity seems to vary as seen in cock-roaches, bees and flies suggesting the existence of different sleep stages possibly serving different functions (van Alphen et al., 2013). But no proper REM sleep as in mammals has been observed.

4.4 Activity and inactivity in cold blooded vertebrates

4.4.1 Fishes

Many studies (Bauchot 1984, Karmanova 1982, Tobler 1992), mainly behavioral, has proved the existence of sleep-like in many fish species. Some sink into the sand to spend the night. Other changes color at night. Some parrotfish can secrete a mucous envelope in which they remain motionless all night. During these states rest, the fish react more slowly to sensory stimuli. In tench (Tinca Tinca), were recorded simultaneously the electrical activity of the brain, the activity of gills and different muscles, respiratory rate, heart rate and finally reactivity to sensorial stimulation (Hartse, 1994). There are no significant differences between the rapid and low voltage electrical activity of the brain during awake state (at night) or during rest (day). (Zhdanova, 2006, for similar results in the zebra fish) Moreover, no periods of eye movement accompanied by autonomic changes has been observed during resting period. There is therefore no evidence to suggest that periodic state, similar to REM sleep, is present in fish.

4.4.2 Amphibians

Amphibians are probably the most poorly studied taxa among tetrapods with sleep–wake data existing for only a handful of species (0.14% of all known species). Despite conflicting results from some studies, the amphibian species studied to date display behavioural characteristics of sleep (Libourel, Herrel, 2015). The majority of studies did not report complete atonia or eye movements during the resting state suggesting the absence of an active sleep state in amphibians. However, one did report an ‘activation phase’ with an EEG similar to that observed during the awake state, motor automatisms, and a phasic transitory heart rate increase in R. temporaria. (Karmanova & Lazarev, 1979)

4.4.3 Reptiles

Three studies on turtles reported the presence of a mammalian active sleep-like state, and one the presence of mammalian quiet sleep-like state. Turtle is among the few reptiles believed to have REM sleep with rapid eye movement and suppression of muscle tone of the neck. In crocodiles, in contrast to other reptilian species, the EEG amplitude and frequency during sleep-like states appears to change, similar to what is observed during mammalian quiet sleep. Only one study reported eye movements and motor automatisms during behavioural sleep not associated with muscle atonia. All studies agree on the presence of sleep-like states in squamates, but the diversity of findings prevents clear conclusions regarding the electrophysiological nature of sleep in these animals. None of the studies on squamates performed to date have examined all of the electrophysiological and physiological traits that allow the characterization of active sleep in mammals. As a conclusion, if reptiles do have REM, it must be quite different from mammalian REM. Reptiles don’t have the brain development of mammals and don't show the extreme EEG differences that mammals do between wake and sleep. Further, reptiles don't seem to have atonia during sleep. This means that the three measurements, brain activity, input/output gating and neuromodulation are all going to be quite different, if they exist at all (Libourel, Herrel, 2015).

5. Factors influencing the amount and rhythmicity of REM sleep

Studies have highlighted some correlations regarding the amount of REM sleep in mammals and birds, but some exceptions remain:

• The average duration of REM sleep episodes increases with the increase in body and brain weight (Parmeggiani, 2011).

• The more an animal is immature at birth, the more he is likely to have big amount of REM sleep. The amount of REM sleep then decreases with ontogenesis (Jouvet-Mounier, 1968).

• The safety of the refuge and the availability of food positively influences the amount of REM sleep. Ruminants, which need considerable amounts of grass to meet their food needs, sleep and have very little REM sleep when grazing, but sleep more and have a greater amount of REM sleep if they are to the barn with unlimited access to food. (Allison, 1976).

• There is a conjecture that REM changes with the predatory/prey conditions of an environment. Predatory animals show more REM than herd and prey animals (Allison, 1976).

• Low temperatures usually decrease the amount of REM sleep (Haskell, Edwin, 1981).

• Sleeping postures that need a minimal muscle tone are less compatible with REM sleep.

REM sleep is not continuous during sleep: it appears periodically and its rhythmic recurrence structures sleep cycles. Periodicity is unique for each animal species: it occurs roughly every 4 minutes of sleep in mice, every 12 minutes in squirrels, every 27 minutes in cats, every 60 minutes in the horse, every 90 minutes in humans and every 100 minutes in the elephant.

As other physiological parameters such as the duration of the cardiac, respiratory or sexual cycles, the periodicity of the REM sleep is inversely proportional the basal metabolic rate of the animals (exceptions are ruminants). Thus the mouse, which has a basal metabolic rate 25 times greater than the elephant (that is to say, it consumes twenty five times more oxygen per gram of body weight per hour), has REM sleep 25 times more often the elephant. If the dream mouse more frequently, she dreams shorter than the elephant.

6. Phylogeny of sleep states

7. Dreams content in animals

8. Roles of dreaming for animals

9. References