Evidence of the duality of the states of sleep
(a) EEG and behavioural findings
It is believed at the present time that slow sleep is characterized (in the chronic adult cat) by two successive EEG stages. The first is made up of cortical spindles of 15-18 c/s, the second of slow high-voltage waves of 2-4 c/s which invade the cortex and subcortical structures. These two stages are accompanied by tonic nuchal activity in the EMG and almost complete absence of eye movements. The appearance of spindles requires the integrity of the thalamus (Naquet et al. 1964), while that of slow cortical and subcortical waves necessitates the neocortex (M. Jouvet, 1962a). The mechanisms of the appareance of slow sleep have been discussed elsewhere (M. Jouvet, 1962a; Moruzzi, 1960).
Paradoxical sleep differs entirely from slow sleep in EEG, behavioural, tonic and phasic aspects.
(i) Tonic aspects (Fig. 1)
(1) EEG. ln normal adult cats not deprived of sleep, P.S. always occurs after a phase of slow sleep. Its average duration is 6 min, but periods of 15-20 min are often recorded. It represents an average proportion of 20-25% of behavioural sleep (i.e. approximately 15% of the total time). lt is characterized by fast, low-voltage neocortical, diencephalic and mesencephalic activity (20-30 c/s), similar to that of cortical activation which regularly accompanies a state of intense alertness or attention. But certain local electrical peculiarities make it possible to draw a formal distinction between the electrical cerebral activity of P.S. and that of behavioural alertness. The appearance of continuous 6-activity in the dorsal and ventral hippocampus is highly characteristic: it is more regular, more rapid (5-7 c/s) and, above all, of a more extensive topography than that observed during intense alertness (4-4.5 c/s) in the dorsal hippocampus—the presence of a 6-rhythm in the ventral hippocampus occurring only exceptionally in the waking state. A 6-rhythm has also been registered in the peri-aqueductal grey matter, the anterior portion of the pons and the limbic midbrain area (M. Jouvet, 1962a).
(2) Behavioural criteria. In contrast to slow sleep, which is not clearly defined behaviourally, the beginning and end of P.S. can be located to within a few seconds on the basis of behavioural criteria alone. The complete extinction of muscular activity, especially of the neck, is the most striking sign of the inhibition of muscular tonus that characterizes P.S. (M. Jouvet et al., 1959). Before, or some seconds after the cortical activity of P.S. begins, the lack of activity in the EMG is accompanied by a sudden dropping of the animal's head if it has been in an unsupported position during slow sleep. The end of P.S. is marked by a renewal, usually sudden, of considerable activity in the EMG, the awakening of the animal or transition back to the state of slow sleep.
(ii) Phasic aspects
Among the phasic phenomena characteristic of P.S. the eye movements are of such great importance that we shall consider them separately. But they are not in fact an isolated phenomenon and P.S. is punctuated in a strange and disordered manner by sudden movements of the ears, whiskers, limbs (flexion) and tail, and sometimes veritable clonic jerks of the muscles of the back (Fig. 2). These phasic phenomena are particularly developed in the young cat after birth and are characteristically increased after long privation of P.S.
Phasic phenomena of the oculomotor system. One of the most remarkable characteristics of P.S. is the appearance of rapid eye movements, accompanied by phasic ponto-geniculo-visual electrical activity.
Rapid eye movements (Fig. 3). Rapid eye movements occur from the beginning of cortical activation. With a frequency of 60-70 per min, they differ in speed, distribution, and pattern from ocular movements of observation during the waking state (Jeannerod and Mouret, 1963). They may be isolated or occur in short bursts of less than 5 movements (as during observation), but most characteristic are bursts of a greater number of movements, up to 50 without a pause. The ratio between the number of movements during the bursts and the total number of movements remains constant for each animal (50%). During P.S., myosis is at a maximum most of the time while the nictitating membranes are relaxed. Nevertheless, sudden mydriasis with retraction of the nictitating membranes can on occasions accompany the bursts of eye movements (Berlucchi et al., 1964). Analysis of the structures responsible for the appearance of isolated movements and bursts gave the following results (Jeannerod et al., 1965): pontile cats in which the superior colliculi were destroyed showed only isolated lateral and external movements (dependent on N. VI). In the mesencephalic cat with superior colliculi intact the larger bursts persisted. In contrast, coagulation of a zone in the superior colliculus and the mesencephalic tegmentum in intact animals suppressed these bursts. The latter were, on the other hand, very much enhanced in the totally decorticated animal. The role of the cortex is not unequivocal, for ablation of the visual cortex will decrease the number of bursts and isolated movements, sometimes to a considerable degree, whilst frontal decortication increases the number of bursts.
These observations can be summarized as follows: the rapid eye movements in P.S. are not identical with those of the waking state and persist, for example in decorticated, pontile cats, when eye movements in the waking state are completely impossible. They are also present during P.S. in kittens immediately after birth and still blind (Valatx et al., 1964). The mechanisms responsible for the rapid eye movements of P.S. must thus be different from those regulating the eyes during observation. The results suggest that these eye movements are initiated in the pons and are rendered more complex in the superior colliculi and the midbrain, whilst the process of 'cortical integration' (facilitating visual cortex and inhibiting frontal cortex) would act on this latter region.
Phasic ponto-geniculo-occipItal. activity (Fig. 4). Owing to the difficulties encountered and the slow development of a broad and systematic study of the cortical and subcortical structures under chronic conditions several years were necessary before a common link was found between the 'spontaneous' potentials observed during eye movements in P.S. First observed in the pontine reticular formation, monophasic peaks of 200 300 microV of 100 msec duration, often appearing in groups of five or six (whence their appearance as pseudo-spindles) (M. Jouvet et al., 1959), were then observed in the lateral geniculate nucleus (Mikiten et al., 1961), the occipItal. cortex (Mouret et al., 1963), the superior colliculus and oculomotor nucleus (Brooks and Bizzi, 1963; Michel et al., 1964a), and the pulvinar and parietal cortex (Hobson, 1964). The pontine and geniculate spikes are the earliest signs of incipient P.S. They can in fact precede rapid cortical activity and extinction of the nuchal EMG by several seconds. More rarely, this phasic activity can occur in fleeting bursts during slow sleep without the appearance of P.S.
Lesions of the nucleus reticularis pontis caudalis or in front of the pons, in the dorsal or central part of the brain stem, may suppress the appearance of geniculate and visual spikes during P.S. (Hobson, 1964). On the other hand, monophasic spikes persist in the pontine reticular formation (PRF) in the pontile animal during P.S. lt is thus probable that an ascending ponto-geniculo-occipItal. system, of which the topography has still to be clarified, reacts phasically during the rapid eye movements.
But the relationship between this phasic activity and the eye movements is not a simple one. Neither darkness, retinal coagulation, nor even total ablation of the eyeballs and the extrinsic muscles of the eye (Michel et al., 1964a) suppresses the pontovisual peaks, which therefore cannot be regarded as a possible feedback of retinal (on and off effect) or extrinsic muscular origin (Fig. 4). Moreover, this phasic activity precedes the eye movements by some 30-90 sec at the beginning of P.S., and the movement can occur without demonstrable spike activity, but in the majority of cases there is a relation in time between the monophasic ponto-geniculo-occipItal. spike and the activity of the extrinsic muscles of the eye. This activity appears above all as rapid phasic bursts, whilst in the waking state a tonic element occurs (Michel et al., 1964b).
It would be premature to try to correlate this phasic activity and the rapid eye movements at the present time, and it is sufficient to note the essential difference between these phenomena and those occurring during the eye movements of observation (when such phasic activity is not recorded) and especially during slow sleep, during which phasic phenomena occur neither in the motor effectors nor in the EEG.
These findings thus enable us to draw a clear distinction between P.S. and slow sleep on the basis of their EEG and tonic and phasic behavioural aspects. By these criteria P.S. appears as distinct from slow sleep as the latter is from the waking state.
But we cannot affirm on the basis of EEG methods and polygraphy alone that slow sleep and P.S. are the result of different mechanisms and structures. In order, therefore, to obtain more evidence in favour of the dualist concept for the two states of sleep ( M. Jouvet et al., 1959) we studied the possibility of differentiating them, either in the course of their phylogenic or ontogenic evolution, by selective deprivation, or by central nervous lesions.