(b) Mechanisms underlying the appearance of P.S.
We shall discuss here the question of reflex production of P.S., the effect of deprivation of P.S. in pontile animals and humoral factors.
(i) Reflex production of P.S.
Inhibition of nuchal tonus during P.S. is not dependent on the gamma system alone, since it persists after section of the posterior cervical roots of the intact animal. Since, furthermore, the cerebellum is intact in pontile animals, it would appear that production of P.S. as a reflex depends on the integrity of the pontine reticular formation (posterior part of the nucleus reticularis pontis oralis and anterior part of the N.R.P. caudalis). There is thus no direct inhibitory effect emanating from the bulbar reticular formation (situated at the level of the nucleus giganto-cellularis and nucleus reticularis ventralis (Magoun and Rhines, 1946; brodal, 1957), since P.S. can no longer be triggered by mechanisms distal to the mediopontine or retropontine areas. Participation of adrenalin secretion can also be ruled out, as P.S. can appear immediately after nociceptive stimulation and, furthermore, injection of adrenalin does not produce P.S. It also appears that the integrity of the mesencephalic tegmentum prevents reflex P.S. in mesencephalic animals, although P.S. still occurs spontaneously. Activation of the descending facilitatory system (Magoun, 1950) would thus preclude descending inhibition in mesencephalic animals. We did not analyse the nervous effects of the natural stimuli used in our experimentation, so that it is difficult to isolate the particular type of afferents responsible for triggering P.S. But we do know that the majority of afferents of group III project principally on to the mesencephalic tegmentum, while those of groups I and II project for the most part on to the pontine reticular formation (Pompeiano and Swett, 1963), so that the latter may be responsible for producing reflex P.S.
Thus, under certain conditions, P.S. can be produced by a reflex mechanism. A distinction must however be drawn between generalized atony, a veritable reflex cataplexy, without pontine spikes or 'spindles' and eye movements, and true P.S. The former state presents no refractory phase and can follow any proprioceptive, or nociceptive, stimulus when the pons has been transected at the nucleus reticularis pontis oralis level. Such cataplexy resembles the 'sudden postural collapse' described by Bard and Macht (1958). It has never been possible, on the other hand, to obtain an iterative form of P.S. Thus, the data invite the conclusion that there exist two different systems at the level of the pons. The first, which is not endowed with refractory period, is responsible for inhibition of muscular tonus (probably via the inhibitory bulbar reticular formation), and the second, which controls the first and presents a refractory period, is responsible for pontine spikes and 'spindles' (and probably the EEG phenomena in intact animals) and for the eye movements. This second system can be identified with the 'centre déclenchant' of P.S. (M. Jouvet, 1962a) for this is the level at which the refractory phase producing the periodicity of P.S. appears. Such a system may explain the different aspects of certain reflex cataplectic attacks in humans. The one type - cataplexy - narcolepsy - presents EEG, polygraphic and electromyographic tracings identical with those in s]eep with eye movements (personal observation), and the other cataplexy - is characterized by intact consciousness.
If, therefore, P.S. can be induced as a reflex under certain conditions, this does not appear to be the sole mechanism. Nevertheless, this hypothesis has been maintained by Lissak et al. (1962). These authors contend that inhibition of muscular tonus is the cause, and not the effect, of P.S. But several facts make it possible to preclude the exclusive activation of P.S. by peripheral muscular afferents.
(1) Cortical activation of P.S. often precedes the disappearance of muscular tonus in intact animals.
(2) In pontile animals (without hypothalamic island) long periods of total atony are not infrequently observed. No EMG activity in the neck or other muscles is then present. ]n spite of this complete atony, P.S. can regularly appear, recognizable by pontine EEG criteria, rapid eye movements and autonomic changes. Total muscular atony is thus not per se an inhibitory or facilitatory factor of P.S.
(3) Lastly, total section of the spinal cord at C6 with section of the brachial plexus, whereby a large number of afferents of muscular origin are suppressed, does not affect the appearance of P.S.
Thus, the result of nervous deafferentation experiments is negative. Furthermore, the periodic and regular appearance of P.S. 4 5 days after ablation of the hypothalamus and pituitary rules out a direct and indispensable participation of ACTH and the post-pituitary hormones and the different hypothalamic neurohormones. The decrease in P.S. seen from about the fourth day and before death must be ascribed to delayed ionic and metabolic disturbances resulting from the prolonged absence of pituitary (especially antidiuretic) hormones.
The periodic appearance of P.S. thus cannot be explained by peripheral nervous or hypothalamo-pituitary factors, so that investigation can be restricted to pontine phenomena (neuronal or glial), although peripheral humoral factors cannot be excluded a priori.
(ii) It is difficult with the methods of study employed to specify the mechanisms of P.S. precisely. But several indirect findings all point to this state of sleep being endowed with a certain 'autorhythmicity' and situated in the pons.
The presence of P.S. under conditions of hypothermia at 29° indicates the resistance to cold of its underlying mechanisms. The findings are similar to those in anesthesia: P.S. in fact continues to appear periodically in decorticated animals after doses of 20-30 mg/kg pentobarbitone (M. Jouvet, 1962a). Furthermore, the greater duration of P.S. (and of the intervals between the phases) during hypothermia is reminiscent of enzymatic processes and suggests a metabolic phenomenon.
The action of y-butyrolactone (G.B.L.) is difficult to interpret, in spite of the fact that its structure is simple. Bessman and Skolnik (1964) see G.B.L. as a normal constituent of the brain, an increase in the intracerebral concentration coinciding with anesthesia induced by G.B.L. or gamma-hydroxybutyrate (G.H.B.). Giarman and Roth (1964), on the other hand, found no G.B.L. in the blood or brain and attribute the anesthetic role to y-hydroxybutyrate. G.B.L. does not appear to be a precursor of cerebral GABA, for the level of the latter does not increase after injection of G.B.L ( Giarman and Schmidt, 1963). It has also been shown that injection of G.B.L. can raise the level of acetylcholine in the brain (Giarman and Schmidt, 1963), and specifically in the area of the corpora quadrigemina. Whilst the mechanism behind this increase in cerebral acetylcholine is not known the ata suggest a cholinergic-type link within the P.S. process. Certain peripheral manifestations of P.S. in intact animals (myosis, bradycardia, increased and irregular respiration, arterial hypotension (Candia et al., 1962) inhibitory role of atropine (M. Jouvet, 1962a), support this view. But while such observations suggest the intervention of cholinergic neurons during P.S., they cannot explain the periodicity.
(iii) Deprivation of P.S.
The progressive diminution of the intervals between each incipient phase of P.S. in pontile cats caused by the increasingly frequent repetition of the disturbing shock, suggests a biochemical process whereby the accumulation during wakefulness of an unknown factor among the metabolites of neuronal activity provokes the increasingly rapid induction of a recuperative process. The inescapable nature of this reappearance of P.S. meant that with our technique prolonged deprivation was impossible. It nevertheless appears that a minimal increase in the recuperative process is adequate to restore the earlier conditions and normal rhythm of P.S.
In the intact animal, on the other hand, in which prolonged periods of deprivation are possible, P.S. increases considerably but does not exceed a certain ceiling (60%) during the first 6 h of recuperative sleep. This plateau thus appears to represent the upper limit of the cyclic metabolic processes of P.S. In this connection, it is interesting to note that recuperation does not occur in one long phase of P.S., but the periodicity remains, as though it were impossible for P.S. to last for more than 20-25 min at a stretch. Thus, the P.S. deficit which is accumulated over a long period of deprivation can only be partially and slowly compensated by a phenomenon which remains periodical and is of limited duration. This suggests the existence of a self-regulating cyclic metabolic process which requires several days before it can 'neutralize' the unknown factor accumulated during deprivation.
(iv) It is a difficult task to interpret the relationships between osmolarity of the blood and P.S., for the amounts of ions normally present in the cerebrospinal fluid and the brain are not known. The post-pituitary hormone could be cited as an indirect factor in the increase of P.S. during withdrawal of liquids and following injection of hypertonic saline solution. But injection of ADH does not promote P.S., which can occur in transitory diabetes insipidus (i.e. in the absence of post-pituitary hormones). There must therefore be another mechanism at work, which cou]d be none other than a direct influence of the osmolarity of the blood. In the case of pontile animals with an intact hypothalamic island, the relationship between the blood and the brain must always be considered in particular at the level of the barrier between the capillaries and the neuroglial cells, to which an important and active role is accredited in what is termed the blood-brain barrier (Edström, 1964). According to recent findings (De Robertis and Gerschenfeld, 1961), the neuroglia appears to be the tissue that plays the greatest part in the mechanisms of ion and water exchange between the internal milieu and the neurons. It constitutes a water ion pool (Gerschenfeld et al., 1959) between the blood and the neurons, playing a part in the transportation of metabolites and the storage or elimination of K and Na at the neuronal level. An active barrier, which has been likened to the renal glomeruli (Tschirgt, 1958; Edström, 1964), would thus keep the whole neuronal complex relatively independent of variations in the iOI1 content of the extracellular fluid. Furthermore, the osmolarity of the brain is closely related to that of the blood and quickly adjusts itself to variations in the latter (Stern and Coxon, 1964). There are thus grounds for supposing that variations in the osmolarity of the blood above all affect the glial cells. lt is therefore possible that an electro]yte concentration exceeding that of these cells facilitates the enzymatic mechanisms responsible for P.S., whereas a lower concentration would inhibit them. Lastly, the appearance of regular rhythmic activity during rehydration (after fluid deprivation) only during P.S. would appear to signify a close relationship between the electrical activity of the pons in P.S. and the osmolarity of the blood. One might therefore presume that at this moment, and only at this moment, the glial structures participate in the recurrence of a certain type of cerebral homeostasis. This hypothetical periodic glial mechanism, with its purpose either of maintaining cerebral homeostasis (cerebrostasis) or as an active process for the elimination of certain metabolites, would protect the brain so that P.S. would appear as a periodic phenomenon interrupting waking activity in the pontile animal and slow sleep in the intact animal, to restore equilibrium at the neuronal level. It must be admitted that if such a mechanism exists its location in the pons would be ideal. For, at that level, the nervous and glial cells could receive biochemical information on the activity of the neurons from the whole nervous system, since the afferents from the rostral and caudal regions of the nervous system converge at that point (Brodal, 1957), whilst efferent ascending and descending pathways lead out from the pons to the cortex and spinal cow (Scheibel and Scheibel, 1957).
(v) The nature of P.S.
Finally, we must gather the facts together and attempt to draw up some hypotheses on the nature of P.S.
It would appear to be established that P.S. is dependent on a periodic mechanism of an unknown nature situated in the pons. This mechanism triggers an ascending tonic neuronal activity similar to (or perhaps identical with) that of the waking state and a phasic activity which is specific to it. However intense the cortical, reticular or pyramidal neuronal activity during P.S., it is prevented from expressing itself (except for the rapid eye movements) in appropriate tonic motor phenomena by the inhibitory reticular formation set in action by the pons. The findings we have discussed also show P.S. to be a state differing qualitatively from slow sleep. The relation between the two is nonetheless a close one, however, for slow sleop is normally a precondition for P.S. Finally, numerous observations in humans show that dreaming is the subjective equivalent of P.S. (Dement and Kleitman, 1957; M. Jouvet and D. Jouvet, 1964).
Passing from the facts to hypothetical considerations, a number of points can be made. Recollection of recent or past events during P.S. suggests that certain memory processes occur during this phase. The relatively high incidence of P.S. in the neonatal period, when the 'plastic' processes of learning are at their most active, is also indicative of a probable relationship between P.S. and the process of memory. A further sign is the presence of a particular rhythmic 6-activity in the limbic system. The relationship of such activity during wakefulness to storage of information in the CNS has been discussed by Adey (1964). Finally the parallelism between variations in P.S. and the osmolarity of the blood suggests the intervention of the neuroglia, perhaps as regulator of certain processes of protein synthesis essential to storage in the neurons; this is in keeping with the findings of Hydén and Pigon (1960).
Is it possible that the biological clock in the pons which activates P.S. may be the mechanism responsible for the complex biological processes by which we retain (or lose) the memory of past events during dreaming ? It must certainly be admitted that if molecular changes are to occur at the level of the sensory and motor neurons it would be logical for them to occur during sleep and at a moment when a safety mechanism prevents peripheral motor expression of the discharges due to protein synthesis within the neurons, for otherwise the dreamer would run the risk of behaviourally reacting to dangerous hallucinations.
Paradoxical sleep, the physiological substratum of dreaming, would thus appear as the expression of a periodic function of storing information at the molecular level. Why this mechanism expresses a need, as objectified by deprivation experiments, remains to be explained. Are we to imagine a threshold in the functional information storage processes occurring in the waking state which initiates slow sleep? A large quantity of information (external stimuli, prolonged low-frequency stimuli) does in fact quickly produce slow sleep in the intact animal (supraliminal inhibition of the Pavlovian type, hypnogenic stimulation). It is possible that, in the pons, where the majority of neurons of the central nervous system converge, certain cells are subject to a biochemical mechanism which represents a certain threshold of functional storage of information. Once this threshold has been passed, a pontine mechanism would initiate the storage process at the molecular level. It is not unreasonable to presume that such processes are cyclic and autoregulatory and that they cannot exceed a certain limit of activity of 60%, as is the case during recuperation following deprivation of paradoxical sleep.