Discussion

(a) Duality of the states of sleep

The first section will be limited to positive evidence of the duality of the states of sleep and thus of the specificity and autonomy of paradoxical sleep in relation to slow sleep.

(i) Structural duality of the two states of sleep

It is difficult to explain the totality of the results we have just described on the basis of a 'unitary' theory of behavioural sleep, according to which similar mechanisms and identical structures are responsible for the two states of sleep. To explain the desynchronization of P.S. following synchronization of slow sleep in accordance with this theory, it has been necessary to presume the existence of a single progressively ascending inhibitory process, which attacks first the ascending activating reticular system and secondarily the thalamic synchronizing structures (Hernandez Peon, 1963). In point of fact, the structural details we have described necessarily imply different nervous structures responsible for the two states of sleep. The results of experiments with section and coagulation all make it possible to localize the structures triggering off P.S. in the pons (M. Jouvet, 1961; Cadilhac and Passouant-Fontaine, 1962; Rossi et al. 1963). We have discussed elsewhere (M. Jouvet, 1962a) the different arguments in favour of a descending origin of slow sleep from the telencephalon. But the initiation of this slow activity has not been finally cleared up. Some results indicate that synchronizing structures which could be responsible for slow sleep are to be found in the medulla oblongata (Moruzzi, 1960; Magnes et al., 1961). It has been suggested ( Rossi et al., 1963) that these structures are situated immediately behind those responsible for P.S. (in the caudal part of the pontine reticular formation). This hypothesis does not however take into account the fact that coagulation directly behind the nucleus reticularis pontis caudalis at the level of the anterior part of the nucleus gigantocellularis does not suppress slow sleep. Also there is no intermediary phase of sleep between wakefulness and P.S. in pontile animals as shown by either EEG or behaviour. If the synchronizing structures were situated in the lower portion of the brain stem, at the level of the pons and the medulla oblongata, we would have to assume that they could produce no EEG or behavioural manifestations in pontile animals. Moreover, the ascending synchronizing influences would have to act on structures rostral to the pons to produce behavioural sleep of an intermediary type between wakefulness and P.S. without slow activity, as occurs in decorticate animals (M. Jouvet, 1962a), whilst the cortex would be necessary to produce slow subcortical activity in intact animals.

(ii) Duality of the mechanisms of the states of sleep (Fig. 25)

Like the structural findings, the EEG, phylogenetic and ontogenetic data speak for a duality of mechanisms. The results of cortical and subcortical recordings which show the fundamental difference between P.S. and slow sleep are in fact confirmed by other methods. Thus, the cortical D.C. potential (Dement, 1964; Wurtz, 1964) in cats shows a sudden negative shift at the beginning of P.S., following the positive shift of slow sleep. The variations in cerebral impedance ( Birzis and Tachibana, 1964) and cerebral blood flow ( Kanzow et al., 1962) are similar to those observed during wakefulness and contrary to those observed during slow sleep. Lastly, studies with electrodes have shown a sudden increase in cortical (Evarts, 1962) or reticular activity (Huttenlocher, 1961) compared with that of slow sleep.

Variations in both the autonomic and somatic spheres show that P.S. is indeed a state qualitatively different from slow sleep, for blood pressure drops suddenly at the beginning of the former, while there is little change in the course of the latter (Candia et al., 1962). Finally, the monosynaptic spinal reflexes change little during slow sleep, and disappear entirely in P.S. (Giaquinto et al., 1964). All the results reviewed here cannot be explained on the basis of a 'unitary' theory of sleep.

Such a theory also necessarily implies that slow sleep (considered as a phase of light sleep) precedes P.S. (deep sleep). If the hypothesis is to be accepted, there ought to be a parallelism between the phylogenic and ontogenic evolution of the two states of sleep. But this is not in fact the case. Whereas slow sleep appears to be a characteristic of all vertebrate species studied polygraphically up to the present time (from reptiles to mammals), P.S. does not appear to be related to slow sleep in their phylogenic evolution. What is more, the absence of P.S. in chelonians, and its very rudimentary form in birds, contrast with its constancy and relative importance in all mammals. The number of species so far studied is too small to permit definite conclusions to be drawn, but it does appear that P.S. occurs from birds upwards on the evolutionary scale. It may therefore be presumed that with its appearance a new function came into being which is not necessarily related to sleep, since it does not appear in reptiles.

Ontogenic aspects also make a differentiation between the two states of sleep possible. At birth, in fact, P.S. occurs relatively more often and is less dependent on slow sleep than in later life. Periods of sleep with rapid eye movements have also been observed in newborns (Roffwarg et al., 1963; Delange et al., 1961). These phases of sleep similar to P.S. can also follow immediately on wakefulness. Thus in the newborn mammal, slow sleep is not a prerequisite of P.S. Until now, sleep states have been investigated primarily in animals that are very immature at birth and it is therefore diffficult when analysing the greater amounts and autonomy of P.S. compared with slow sleep to know what to attribute to the immaturity of the nervous system and what to the conditions of early life - milk diet, etc. Nevertheless, one fact stands out: in the cat the average duration of P.S. is almost from the first identical with that in the adult animals. It is only by a reduction in frequency that the proportion of P.S. versus total sleep diminishes. It must therefore be assumed that the P.S. mechanism is already at birth what it will be later in the adult, whereas slow sleep requires further deve]opment. In other words, P.S. depends on an 'innate' mechanism, whereas the mechanism of slow sleep is acquired after birth.

Whilst in adulthood a behavioural distinction can be made polygraphically between the two states of sleep, the fact that slow sleep habitually precedes P.S. would permit the conclusion that the former is a prerequisite of the latter. But our technique of selective deprivation makes it very easy to separate them. Our results confirm those of Dement (1960) in humans and show that a need for P.S. appears as soon as an attempt is made to suppress it. This suggests that a particular function is connected with the appearance of P.S. It is interesting to note in this context that a cat deprived of P.S. for more than 72 h resembles a newborn kitten in its recuperative sleep. In both cases P.S. constitutes the greater part of behavioural sleep (up to 80% of the first hour of recuperative sleep), phasic phenomena and periodicity are very much increased and, especially, P.S. can follow immediately on wakefulness, without being preceded by a phase of slow sleep. P.S. following directly on a state of wakefulness has also been observed in adult man during narcoleptic attacks (Rechtschaffen et al., 1963). Thus under particular conditions - selective deprivation of P.S., narcolepsyÑnot yet fully understood, it is possible to distinguish between the mechanisms of P.S. and slow sleep in adults.

Next page