Physiology & Behavior, Vol. 28, pp. 323-333. Pergamon Press and Brain Research Publ., 1982. Printed in the U.S.A.
Subcortical Waking and Sleep During Lateral Hypothalamic "Somnolence" in Rats SHAI SHOHAM
AND PHILIP TEITELBAUM
D e p a r t m e n t o f Psychology, University o f lllinois, Champaign, IL 61820 R e c e i v e d 2 J u n e 1981 SHOHAM, S. AND P. TEITELBAUM. Subcortical waking and sleep during lateral hypothalamic "somnolence" in rats. PHYSIOL. BEHAV. 28(2) 323--333, 1982.--Following extensive bilateral lateral hypothalamic damage, rats appear "somnolent." Cortical EEG shows persistent high voltage delta, reinforcing the impression of sleep. Preoperatively and postoperatively, we simultaneously measured cortical and subcortical (hippocampal and pontine) EEG, muscular events (neck muscle EMG and eye movement EOG), and behavior, which, as aggregates, differentially define quiet sleep, active sleep, and waking. Postoperatively, though cortical activity was persistently slow, subcortical EEG, muscular events, and behavior, as aggregates, revealed quiet sleep, active sleep, and waking, organized subcortically, intact and alternating, but disconnected from the persistent slow cortical activity. For example, preoperatively, active sleep included cortical low voltage fast activity, hippocampal theta, episodic pontine spike bursts, flat EMG, and rapid eye movements, without any organized behavior. Postoperatively, the same aggregate of subcortical and muscular events indicated the presence of active sleep. Similarly so, for subcortically organized quiet sleep and spontaneous waking. Such waking, termed "drowsy-wakefulness," is a low-arousal form, perhaps related to drowsiness in other species, and to human hypersomnia. Drowsiness Pathology of sleep and waking Aphagia and adipsia Recovery of function Methodology of sleep-recording
L A T E R A L hypothalamic damage has been of interest primarily because of its effects on feeding and drinking. Anand and Brobeck  showed that small localized lateral hypothalamic lesions produce aphagia and adipsia in rats and cats, leading to death from inanition. However, if kept alive by tube-feeding, such animals eventually recover , in a sequence of stages that constitute the lateral hypothalamic syndrome  (see  and  for more recent reviews). More extensive damage produces additional symptoms, including sensory neglect [23,24] catalepsy, akinesia, and somnolence [19, 21, 25, 31, 35, 39]. In some of our previous studies [55,56], such somnolence was considered as an abnormality reflecting a loss in endogenous arousal, and thus a factor that could account for deficits in feeding and orienting. In this paper, we analyze lateral hypothalamic somnolence as a phenomenon in its own right, to learn about its relation to normal states of sleep and waking. Therefore, we make a distinction, in this paper, between the clinical label "somnolence" and what we find to underlie such a behavioral state in terms of physiological and behavioral measures. The study of the E E G has been fruitful in understanding the neurophysiological basis of the states of arousal underlying sleep and waking . In recent years, cortical and hip-
pocampal E E G has also proven valuable in the analysis of various kinds of movement , and has provided the lawful correlations described below that served as the framework for our laboratory's initial work in the E E G analysis of " s o m n o l e n c e " . In the waking normal rat, hippocampal rhythmic sinusoidal activity (RSA or " t h e t a " ) always occurs during movements such as locomotion, swimming, orienting, etc., behaviors categorized by Vanderwolfet al.  as Type I (voluntary). This type of theta cannot be abolished by atropine sulfate . Immobility, and stereotyped automatisms such as grooming, tooth-chattering, etc., are categorized by Vanderwolfet al.  as Type II. These can be accompanied by hippocampal irregular activity of large amplitude. Sometimes, during immobility, there is an atropine-sensitive form of hippocampal theta, of a relatively lower frequency and amplitude. This happens, for example, when a rat freezes in reaction to a danger-signalling stimulus . The cortical E E G in normal rats is low voltage fast activity during movement of Type I, and remains so during Type II behaviors, unless the animal becomes drowsy and falls asleep, at which time there is large amplitude slow activity. The cortical low voltage fast activity cannot be abolished by atropine sulfate
1This work was supported by National Institutes of Health Grant #R01 NS11671, a University of Illinois Research Board Award, and a University of Illinois Biomedical Research award to Philip Teitelbaum, to whom reprint requests should be sent.
C o p y r i g h t © 1982 B r a i n R e s e a r c h P u b l i c a t i o n s Inc.--0031-9384/82/020323-11503.00/1
324 during movements of Type I, but it is abolished during immobility and generally during movements of Type II, when it is replaced by cortical high voltage slow activity. Thus, two pharmacologically differentiated systems reflect the organization of movement in the EEG. The atropine-sensitive one correlates with automatisms and immobility [34,47], while the atropine-resistant system correlates with movements perhaps subject to a more complex level of control. Since the lateral hypothalamic-damaged rat shows only immobility or automatisms such as grooming, toothchattering, etc. in stage I of recovery (spontaneous locomotion, orienting, scanning and ingestion are absent, see [41,43]), De Ryck and Teitelbaum  expected that the EEG of the cortex and hippocampus would correspondingly reflect such a simplification of behavior. They studied the cortical and hippocampal E E G in relation to movement in stage I and in stage II. Indeed, after extensive bilateral lateral hypothalamic damage, hippocampal theta was simplified-there was now only cholinergic theta, which could be blocked completely by atropine. It could appear now during movements when it never before was present, i.e., during movements of Type II, and sometimes, in trains, during immobility. This implies that an inhibitory control over cholinergic theta, which acts in the normal rat to limit cholinergic theta to some forms of immobility, is now missing. Cortical activity was also simplified. Fast low voltage activity was persistently absent, leaving only slow high amplitude. Fast activity did not appear, even when the animal initiated movement. However, as recovery progressed, the cortex desynchronized in correlation with theta whenever the animal moved. (These results verified similar findings by Kolb and Whishaw [t9] and Robinson and Whishaw [31[.) Therefore, early in recovery the cortex was functionally disconnected from subcortical (hippocampal) correlation with movement, and was functionally re-connected as recovery progressed. The behavioral lethargy [21, 27, 31] and the persistent slow cortical E E G [6, 13, 19, 25, 39] that follow sufficient lateral or posterior hypothalamic damage led to the interpretation that the " s o m n o l e n c e " is an abnormally deep and persistent form of sleep. An alternative view, expressed recently , is that this involves a complete "disorganization" of sleep and waking states, including a loss of active sleep. However, on the basis of De Ryck and Teitelbaum's study  showing cortical disconnection from the subcortical organization of movement, we thought that during lateral hypothalamic " s o m n o l e n c e , " there might be a similar disconnection and subcortical organization of states of sleep and waking. These subcortically organized states might very well be different from normal. To test this, we measured the various EEG and muscular activities which, as aggregates, define waking, quiet (slow wave) sleep, and active (REM) sleep. In our rats, simultaneously with cortical EEG, we studied subcortical E E G activity in the hippocampus and pons. We also recorded neck muscle EMG, eye movements (EOG), and behavioral movements. In accordance with earlier studies of sleep in the rat [30, 32, 45, 54], three well established patterns of cortical and subcortical E E G and peripheral activities were found that define the three normal states of arousal: waking, quiet sleep, and active sleep. Postoperatively, as the data shown below demonstrate, the same three sleep and waking states appeared and alternated, but they were organized only at a subcortical level. Therefore, lateral hypothalamic rats are not really " s o m n o l e n t , " in the sense
SHOHAM A N D T ! ITI!;IBAt:M of being in a tmitary persistent, unchanging stat,~: of deep slow wave sleep, nor is their sleep "'disorganized," lhe,. alternate through waking and both forms of sleep. But lhu cortical EEG is disconnected from this subcorlicat organiza tion of sleep and waking, being persistently slow and of high amplitude. The subcortical wakefulness, which wc term "'drowsy wakefulness," appears to be a low-arousal state, transitional in the normal between waking and sleep, perhaps related to that seen normally in other species  and in human hypersomnia [4,10]. Indeed, from earlier studies of decerebrate animals, where components of sleep were demonstrated, one might have expected such a subcortic:d organization after lateral hypothalamic partial transection METHOD
Twenty male Long-Evans hooded rats weighing 350-550 grams were implanted with chronic lesion electrodes in the lateral hypothalamus, and with electrographic recording electrodes in the cortex, hippocampus, and pons+ Neck muscle EMG and eye movement recording electrodes were also implanted at that time. Ten of the animals were implanted with cortical, hippocampal, pontine, and neck muscle EMG electrodes. The other ten animals were implanted with cortical, hippocampat, eye movement EOG, and neck muscle EMG electrodes. Under equithesin anesthesia (3.0 cc per 100 grams body weight), with the animal's head fixed in the stereotaxic instrument, level between bregma and lambda, seven stainless steel miniature screws were screwed into the dorsal bony cranium to support cement attachment of electrodes and head-connector. One screw, over the cerebellum, served as reference for EEG recording. The lesion electrodes were 00 stainless steel insect pins insulated with Formvar, except at the 0.5 mm conical tip. They were implanted using bregma as reference: P2.5, 1,2.0, and 8.0 mm ventral from the dural surface. EEG recording electrodes consisted of stainless steel insulated pins, 250 micrometers in diameter, insulated except at their tip. They were implanted in pairs, aligned adjacent to each other along their length, with one millimeter separating their lower tips in the vertical dimension (after De Ryck and Teitelbaum ). For placing the lower tip of each pair of electrodes in the dorsal hippocampus and frontal neocortex, the bregma was also used as reference; hippocampus: P4.0, L2.5, and 3.0 mm ventral from dura; neocortex: A4.0, L2.0, and 2.0 mm ventral from dura. For the dorsal pons, the interaural line was used as reference (bregma as midline): P2.0, LI.5, and 6.5 mm ventral from dura (based on Farber et al. ). EMG electrodes were made of Biomed 5-stranded stainless steel wire ( Medwire Corp., Mt. Vernon, NY) exposed by flame at the tip and tied to the superficial neck muscles, one wire on each side of the midline, about 10 mm apart. The same type of wire. slipped through the connective tissue over the eye balls, was used for EOG recording. All electrodes were connected to a male 9-pin ITT Cannon miniature connector (MDI-9PLI), which was firmly attached to the skull with dental cement. EEG Re('ording Techniques and Experimental Pro('('durc.+ Recording techniques were basically as described by Schallert, De Ryck and Teitelbaum . However, because sleep may be disrupted in an unfamiliar environment, all recordings were made while the rat was in its home cage, in the room in which it was housed. The cage was made of Plexiglas to allow observation of behavior and had a removable top to allow entry of a recording cable. When recording,
SLEEP AND WAKING DURING "SOMNOLENCE"
FIG. 1. A representative section of the brain of a rat with bilateral lateral hypothalamic damage. HI=Hippocampus, TH=Thalamus, RTI=Inferior Thalamic Radiations, CAI=Internal Capsule, HMV=Ventro-Medial Hypothalamus, TO=Optic Tract. (Abbreviations from K6nig and Klippel, .)
the cage was placed in a 60x60 cm Plexiglas open field. A multiconductor shielded cable linked the head connector to a 9-channel mercury commutator, which in turn was connected to AC preamplifiers (P5 and 7P511) of a Model 7 Grass polygraph. The neocortical, hippocampal, pontine, neck muscle, and ocular electrographic activities were recorded bipolarly and were differentially amplified. The bandwidth (3 dB cutoff points) for neocortical E E G was 1-35 Hz, for hippocampal 1-30 Hz, for EMG 10-90 Hz, and for EOG 3-15 Hz. Movement-related neocortical desynchronization appeared to be devoid of RSA (theta). Movement-related hippocampal RSA had amplitudes of 300 /xV-1.5 mV. Thus, criteria for optimal electrode placement were fulfilled . A simple movement sensor was used. Six flexible rubber elements supported the bottom of the open field, which touched the needle of a phonograph pick-up cartridge (M44E Shure) placed underneath. Movements of the rat in its cage produced mechanical displacement of the cartridge needle which was AC amplified (Grass 7P51 l, 0.3-30 Hz bandwidth). This yielded an analog voltage output, whose frequency and amplitude varied with the extent and rhythmicity of movement. With the cage on the floor of the open field, and the rat connected to the polygraph via the multiconductor shielded cable, the rat's movements could be recorded simultaneously with electrographic activity. This helped us distinguish sleep from waking, and quiet from active sleep. Two weeks after implantation surgery, recordings were begun for a three-hour period each day (3-6 p.m.) for a week,
to obtain sleep baseline records. Room temperature was 23 degrees centigrade. Lights were on from 9 a.m. to 9 p.m. After baseline recordings were obtained, bilateral lesions were made. The awake animal was immobilized by wrapping it in an elastic bandage, except for its head. F o r each lesion, a direct current of 1 mA for 30 seconds was passed, with the cathodal electrode in the rectum; each chronic lesion electrode served in turn as the anode. Lesions were made while the animals were awake and unanesthetized in order to observe the full range of behavioral and E E G consequences of the damage, and to avoid confounding the effects of anesthesia with the effects of lesions. Postoperatively, recordings were made at intervals throughout the light part of the dark-light cycle. Initially, a continuous period of three hours was sampled at intervals of about one to two hours. As the EEG and behavioral phenomena stabilized, intervals between each three-hour recording session were increased to as long as six hours. All data reported in the present paper were obtained from animals in stage I of recovery, as defined by lack of any eating or drinking, even when offered palatable wet foods in a dish on the floor of the home cage .
Histology After one to two weeks (the duration of stage I), in which sleep and waking were recorded, the animals were anesthetized with Nembutal and perfused through the heart, first with saline, then with 10 percent Formalin. Forty-
SHOHAM AND TI..II'I,II BAt.! M
326 micrometer sections of frozen tissue were mounted on slides and stained for cell bodies, using cresyl violet. RESULTS
Histology The lesions produced extensive damage in the region of the lateral hypothalamus (Fig. 1). Laterally, they destroyed the medial part of the internal capsule (CAI). Medially, they encroached upon the lateral aspect of the ventromedial nuclei (HMV). Posteriorly, the lesions typically extended to include the H1 and H2 fields of Forel, and in a few animals beyond them to the level of the red nucleus and the substantia nigra. The anterior-posterior range was estimated to extend from A5340 to A2970 in the Krnig and Klippel atlas . Ventrally, the lesions bordered on the optic tract (TO). Dorsally, they encroached upon the zona incerta, ventral thalamic nuclei, medial lemniscus, and the inferior thalamic radiations (RTI). These large lesions represent partial transections at the level of the lateral hypothalamus. They destroy, in varying degree, the neural systems that pass through this area, including the medial forebrain bundle, the nigro-striatal bundle, and the internal capsule (CAI). Such disruption of rostrocaudal systems may contribute to the various disconnection phenomena observed in the lateral hypothalamic syndrome [22, 41, 42, 53]. The damage to the inferior thalamic radiations may also be involved , particularly in some aspects of EEG disconnection (see Discussion section).
Sleep and Waking in Normal Rats Preoperatively, three states were identified: waking, quiet (slow wave) sleep, and active (REM) sleep. Behaviorally, waking was characterized by organized spontaneous acts such as walking, rearing, sniffing, face grooming, scratching, etc. The animals startled, froze, or oriented in response to loud noise, or touch on the flank. Electrographically, such acts were accompanied by cortical low voltage fast activity, hippocampal theta during Type I acts , occasional eye movements, and a high voltage and variable neck muscle tone. The sensor under the cage also reflected the behavioral movements of the waking rat (Fig. 2; W A K I N G , PRE-OP). During sleep, the rats typically assumed a curved sleep posture, with closed eyes, and generally did not react to the environment. Electrographically, during quiet sleep, there was cortical high voltage slow activity with intermittent spindling, hippocampal irregular activity, no eye movements, and intermediate voltage neck muscle tone. No behavioral movements were detected by the movement sensor, except for occasional postural adjustments (Fig. 2; Q U I E T SLEEP, PRE-OP). During active sleep, there was cortical low voltage fast activity, hippocampal theta, rapid eye movements, and neck muscle atonia (fiat EMG). Except for phasic muscle twitches, there was no organized behavior (Fig. 2; ACTIVE SLEEP, PRE-OP).
Sleep and Waking in Rats after Bilateral Lateral Hypothalamic Damage In the first 24-48 hours postoperatively the rats were hyperactive. Hyperactivity consisted, during the first 8 to 12 hours of either forward locomotion or climbing-like movements along the walls of the cage or box in which they were placed. These two types of hyperactivity could appear in the same subject alternating with each other or in combination.
Hyperactivity was subsequently replaced gradually t~3 "somnolence." There was less locomotion and antigravi~3 support in general, but even when not walking, the animal would engage in head bobbing, or occasional alternating limb movements. As "'somnolence" became dominant, locomo tion appeared to be inhibited, but would occasionally ~eern to burst forth in brief whole body jerks forward. This restlessness was replaced by "somnolence," usually in about 48 hours postoperatively. During the first 24-48 hours post lesion a variety of seizure activities often appeared in the EEG tracings, consisting of spike and wave patterns, series of spikes, or spikes on a background of slow waves. There did not seem to be a lawful relation between the EEG and behavior during EEG seizures. After about 24 hours post lesion, persistent cortical high voltage slow wave activity was established. Thus, it could appear during the hyperac-tive phase, sometimes being replaced by seizure patterns. In summary, the first 24-hour phase of post-lesion EEG and hyperactive behavior can be seen as reflecting phenomena associated with the initial damage. During this phase the rats do not seem to sleep at all. and their movement appears to be driven, as though by electrical or chemical stimulation of the brain. These observations replicate previous studies of lateral hypothalamic damaged rats [5,191. As hyperactivity subsided it was replaced by' ~ somnolence." Such "'somnolence" was manifested in a variety of symptoms: righting in the air or on the ground (contact righting) was often sluggish. The rats' movements were slow in general, their eyes were often closed or semi-closed, they were immobile most of the time, with little or no anti-gravity support (Fig. 5A). They were not comatose, however. When the tail was pinched, for instance, such an animal could show a startle reaction and might orient slowly with its head toward its tail, or move forward away from the painful stimulus. The cortical EEG was persistently in high voltage, slow activity (Fig. 2: COR. POST-OP). reinforcing the impression that the animal was asleep. Cortical activity remained slow and of high voltage even when the rat interrupted its immobility, for example, to groom (Fig. 5B). Thus, as judged only from the cortical EEG and the behavior, the rats appeared to be in a form of sleep, abnormal in its persistence and depth. This verifies earlier studies of the syndrome in rats [6, 19, 21, 25,311. However, when the subcortical and peripheral electrographic activities that also define normal sleep and waking were observed, three aggregates appeared that corresponded to the three subcortical and peripheral activity patterns that had been present preoperatively. These continued to alternate, and were still sufficient to define the three states of waking, quiet sleep, and active sleep. Because the cortical EEG remained in high voltage slow activity and no longer varied with these subcortical, muscular, and behavioral events, we considered it functionally disconnected. One aggregate, similar to active sleep (Fig. 2; ACTIVE SLEEP, POST-OP), consisted of hippocampal theta of relatively low amplitude and frequency, neck muscle atonia (flat EMG), rapid eye movements, and no behavioral movements except phasic muscular twitches. Except for the fact that it was accompanied by cortical high voltage slow activity, this aggregate indicated the presence of an active sleep episode. Another aggregate, similar to quiet sleep (Fig. 2; QUIET SLEEP, POST-OP), consisted of hippocampal irregular activity, intermediate voltage neck muscle tone, no eye movements, and no behaviors or muscular twitches. Even without the slow wave activity, postoperatively, this aggregate indi-
SLEEP AND WAKING DURING "SOMNOLENCE"
COR~ S L ~
EOG -'EMG '
~~ " : .....
0.5mv 0.15my .~11~_ . IL . . . . d ~ , ~ .
FIG. 2. Identification of the three states of sleep and waking in a typical rat, before and after bilateral lateral hypothalamic damage. Recordings include neocortical and dorsal hippocampal EEG, neck muscle EMG, and eye movements (EOG). Behavioral movements were recorded by a sensor placed under the cage. Three typical aggregates of activity can be discerned preoperatively, and are defined as three states: QUIET SLEEP, A C T I V E SLEEP, and WAKING. Postoperatively (POST-OP), cortical activity (COR) is persistently slow and of high voltage. However, as can be seen in this figure, the remaining measures form three subcortical and peripheral aggregates that correspond to the ones observed preoperatively (PRE-OP). The POST-OP records are all from day three following the damage. Note: Polygraph pen sensitivity is higher for the EMG recording.
S H O H A M ANI)
l J l t IJ~,AUM
~ ~ ' ~
' r "'',,','-~''"
:--.~ .,-, ......
•~,. . . . ,-'
,'vT,~,l,,,,1-!,',= O . 0 5 m v
I O.05mv 1 SEC
FIG. 3. QUIET SLEEP and ACTIVE SLEEP in a rat, before and after bilateral lateral hypothalamic damage. Both preoperatively (PRE-OP) and postoperatively (POST-OP), episodic pontine (PON) spike bursts appear during ACTIVE SLEEP, but generally not during QUIET SLEEP. Thus, pontine activity is a component of the EEG and muscular activity aggregates that define the states of sleep and waking (as explained in the text and in Figs. 2 and 4). This reinforces the conclusion that postoperatively, there are subcortically organized sleep states, from which cortical activity is disconnected. The POST-OP records are all from day 3 following the damage. COR=Neocortex, HIP=Dorsal Flippocampus, PON=Pons, EMG=Neck muscle electromyogram, MOV=Movement sensor. Note: Polygraph pen sensitivity is higher for neck muscle EMG and for pontine (PON) recordings.
cated the presence of quiet sleep. Note that the slow cortical activity was devoid of spindling, in contrast to normal slow waves of quiet sleep (see Discussion), and that it appeared all the time. Thus, there appeared to be two subcortically organized states of sleep, in which cortical E E G did not participate. If, indeed, there are two distinct forms of subcortically organized sleep, they should differ in pontine activity as well. Jouvet and Michel  and Bizzi and Brooks  have shown, in the cat, that the pons is the origin of the PGO (pontine-geniculo-occipital) spikes that are characteristic of active (REM) sleep but are absent in quiet sleep. Recently, in the rat, pontine spikes (but not geniculate or occipital) have been shown to accompany active sleep, not quiet sleep . Therefore, in the present experiment, in 10 out of the 20 rats, pontine activity was recorded in addition to all the other electrographic cortical, subcortical and peripheral events described above, except the EOG (there were only five recording channels available on the equipment, forcing the omission of E O G whenever pontine activity was added). Preoperatively, as shown in Fig. 3, during active sleep, episodic spike potentials appeared in the pons (PON), along with the other cortical, subcortical, and peripheral events that define this state of sleep. Thus, the cortex was in low voltage fast activity, there was hippocampal them, complete absence of neck muscle tone (fiat EMG), and only phasic muscle twitches. Preoperatively, during quiet sleep, pontine spikes were generally not observed (Fig. 3; Q U I E T S L E E P , PRE-OP). When they did occur during quiet sleep, it was generally at the transition to active sleep. The other cortical, subcortical, and peripheral events were as described for quiet sleep in Fig. 2 (except EOG). Thus, we confirm Farber e t a l . . Postoperatively, as also is shown in Fig. 3, pontine spike
PON EMG MOV
. . . . . . . . . .
FIG. 4. Preoperative and postoperative pontine spike trains, characteristic of waking. Along with the behavioral acts and the other aggregate activities that define waking, characteristic pontine spike trains appear. Compare with Fig. 3, which shows that such trains do not appear during quiet sleep (no spikes) or during active sleep (episodic spike bursts). COR=Neocortex, HIP=Dorsal Hippocampus, PON=Dorsal Pons, MOV=Movement sensor. POST-OP is taken 3 days following the damage. Note: EMG and PON activity are recorded with higher sensitivity. potentials were present during active sleep, but generally not during quiet sleep, except occasionally, during transitions to active sleep. This supports our hypothesis that in lateral hypothalamic " s o m n o l e n c e " in stage I of recovery, there are two distinct forms of sleep, corresponding to quiet and active sleep, but subcortically organized, and functionally disconnected from cortical activity.
SLEEP AND WAKING DURING "SOMNOLENCE"
FIG. 5. Groom-arrest during "drowsy wakefulness." Figure 5A shows a rat in stage I of recovery. It does not eat or drink and spends most of its time immobile, asleep, or in "drowsy wakefulness" as shown here (note the open eyes). At such times, even though its cortex remains in slow high voltage activity (Figs. 2 and 4), it may engage in spontaneous behaviors which reveal a subcortically organized low-arousal state of drowsy-wakefulness (see text). In B-D, grooming begins, but this is suddenly interrupted, and the rat gradually sinks into sleep.
Waking Postoperatively Besides quiet and active sleep, there was also a subcortical and peripheral aggregate of electrographic activities (Figs. 2 and 4; W A K I N G ) which was accompanied by organized behavioral movements, and thus indicated the presence of a subcortical form of waking. We call this " D r o w s y Wakefuln e s s " (see Discussion). The movements were either spontaneous (face grooming, chewing, tooth-chattering, scratching, yawning, and stretching) or were elicited (tail-pinchinduced startle). The electrographic aggregate (see Fig. 2; W A K I N G ) consisted of hippocampal theta, high voltage and variable neck muscle EMG, behavioral movements, occasional eye movements, and occasional trains of pontine spikes uniquely characteristic of waking (see Fig. 4, W A K ING, PRE- and POST-OP). It should be noted that both preoperatively as well as postoperatively, in each of the three states, pontine activity consisted, for the most part, o f uniform low voltage fast
activity, which did not differentiate any of the states. However, during active sleep and waking, spiking activity did appear which was sufficient to differentiate these two states from slow wave sleep. During active sleep, such spiking consisted of brief episodic bursts. During waking, such spiking consisted of longer trains. Behaviorally, " d r o w s y wakefulness" was very different from normal. One aspect was the animal's apparent inability to maintain spontaneous wakeful activities for a normal length of time. F o r example, there was the " g r o o m arrest" phenomenon, first described by Levitt and Teitelbaum,  (Fig. 5B-D). Often, as the rat bends to groom its genital or abdominal area, it suddenly arrests, and gradually subsides into a head-down posture in which it falls asleep. Also, if the rat was roused to action by brief external stimulation such as a tail pinch, it quickly relapsed into sleep. Thus, the low level of endogenous arousal seemed insufficient to sustain prolonged waking.
SHOHAM AND TIIIqI:-:.|+BAtJM
We identified all three states--waking, quiet sleep and active sleep once stage I "somnolence" was established, following the hyperactivity and seizure phase. Once hyperactivity subsided and was replaced by "somnolence" these three subcortically organized states cycled regularly. Out of 20 implanted and bilaterally damaged rats, 17 were in stage I of recovery . Of these 17, 15 showed the complete cortical disconnection described above. In the other two, cortical low voltage fast activity appeared during behavioral acts or during the phasic twitches of active sleep. One rat of the remaining three died 24 hours post lesion, and the others showed only slight behavioral effects of the damage, quickly recovering normal feeding patterns. In summary, there appears to be a subcortical organization of states of sleep and waking, as reflected in electrographic aggregates of subcortical and muscular activities, and in behavior. This includes a subcortical form of waking, termed "drowsy wakefulness," much reduced in arousal. DISCUSSION
Our results confirm that sufficient bilateral lateral hypothalamic damage, in addition to aphagia and adipsia , catalepsy , and sensory neglect [23,24] produces what looks like a persistent unvarying deep sleep or "somnolence"  accompanied by cortical high voltage slow activity [6,19]. However, these earlier studies did not simultaneously measure the behavioral acts along with the cortical and subcortical E E G and muscular (EMG and EOG) electrographic activities associated with sleep and waking. On doing so, we found a subcortical organization of the states of sleep and waking. This is in contrast to Danguir and Nicolaidis  who observed behavior and recorded only cortical EEG, and concluded from its persistent high voltage slow activity that there was a breakdown of the organization of sleep and waking. In particular, they suggested that active sleep was completely abolished. On the contrary, the present paper shows that subcortical and peripheral components of sleep and waking remain intact, still organized as aggregates, but functionally disconnected from the persistent, slow cortical activity. These subcortical states alternate regularly and appear to correspond to the three states of waking, quiet sleep, and active sleep. However, Danguir and Nicolaidis  were correct in emphasizing that the cortical slow waves are abnormal, for two reasons: (1) there are no sleep spindles, and (2) they remain unchanging, uncorrelated with organized behavioral acts (also see [6, 19, 31, 36]). We verify these findings (see Figs. 2, 3, and 4). However, based on our simultaneous recording of subcortical and muscular events, we believe, not that there is a total disruption of sleep, but that the cortex is functionally disconnected from the states of sleep and waking, now organized only subcortically. The loss of sleep spindles may be related to damage in the inferior thalarnic radiations . The disconnection may reflect a deactivation of cortical function, directly, due to disruption of ascending activating systems [4, 12, 17, 21, 31, 36, 55, 57], and/or indirectly, via a disruption of descending activating inputs to the reticular formation , as well as the thalamocortical systems [36, 50, 51, 57]. Alternatively, there may be a disinhibition of inhibitory systems, such as the anterior hypothalamic descending inhibition of the reticular system [3,37] or the inhibitory reticular systems themselves . In a wider context, our results may be valuable in understanding normal and abnormal states of sleep and wakeful-
ness. In particular, we distinguish a subcortically organized form of waking. On a background of low arousal, and within the constraints of a drastically limited behavioral repertoire. a residue of organized acts still occurs in association with the same subcortical EEG and muscular events that characterized waking preoperatively. As shown in recent studies. early in stage I of recovery, even when not asleep, lateral hypothalamic rats do not explore the environment and do not eat. They cannot spontaneously self-activate the movement subsystems involved in forward locomotion, head-scanning, head-orienting and mouthing of food or fluid [6. 9, 43]+ In other words, they are akinetic, display sensory neglect, and are aphagic and adipsic. However, they retain postural sup+ port, and engage in organized body-oriented acts. such a~ grooming, scratching, head and body shake, stretching, mouthing in air, and yawning, ([6, 9, 21, 31] and the present paper). When these appear, the eyes are usually open, and in the EEG there is hippocampal theta and pontine low voltage activity with occasional spike trains. Eye-movements appear in the EOG accompanying high voltage neck muscle E M G Thus, except for the functionally disconnected persistent cortical slow activity, the entire aggregate of EEG and muscular events appears, that characterized waking preoperatively, distinct from that of quiet or active sleep. We call this subcortical low-arousal state +'drowsy wakefulness." It seems analogous to a similar state, identified as "drowsiness" in normal cattle by Ruckebusch 133]. Cows show cortical high voltage slow activity while awake
SLEEP AND WAKING DURING "SOMNOLENCE" chronization, independently of the cycling of states below the transection [13,49]. Furthermore, after complete removal of cortical tissue, the brain below is still capable of organized sleep and waking cycles . However, in studying animals with localized lesions, it is often assumed that the nervous system, even though abnormal, is acting as a whole. Thus, the persistent slow cortical EEG and lethargic appearance of lateral or posterior hypothalamic-damaged animals has in the past been taken as a sign of a unitary, persistent sleep-like state [13, 14, 21, 26, 29, 31, 39]. Alternatively, when such slow EEG did not agree with the presence of organized behavioral acts, some investigators have interpreted this as a complete disorganization of the EEG, and correspondingly, as a total breakdown of sleep . What is new in the present paper is the demonstration in such animals of a f u n c t i o n a l disconnection of cortical activity from subcortically organized, intact and alternating, sleep and waking states. When, for the most part, such an animal is lethargic, close behavioral observation combined with subcortical and peripheral measures reveals periods of waking--our "drowsy wakefulness." On the other hand, when the animals engage in organized acts, and the cortex remains slow, one must realize that these movements are disconnected from the cortical activity . Perhaps it would be more fruitful to view lateral hypothalamic damage, not as localized lesions, which abolish unitary functions, but rather as partial transections, which decompose hierarchically organized functions into simpler components. After such damage, sleep and waking are at first functionally disconnected from cortical activity, but recovery can occur. For instance, shortly after lateral hypothalamic damage, the association of hippocampal theta with movement still occurs, but functionally isolated from the slow cortical activity. Later in recovery, cortical desynchronization reappears, once again linked with hippocampal theta . Correspondingly, in recovery, one should expect the reconnection of cortical activity with sleep and waking. One might argue that "drowsy wakefulness" is active sleep, not waking. After all, they both have, in common, hippocampal theta, eye movements and pontine activity. The cortex, being functionally disconnected, offers us no clue to distinguish them, and even if it were re-connected, it would still be desynchronized, whether the state was one of waking or of active sleep. The major differences are: during "drowsy wakefulness," the neck EMG is not fiat, and organized behavioral acts occur. But this is what happens in active sleep without atonia in cats after pontine damage. Therefore, why not label our state of "drowsy wakefulness" as sleep without atonia? First of all, we found, preoperatively, that rhythmic pontine spike trains (see Fig. 4) occur only during waking, whereas, during active sleep, episodic spike bursts occur (see Fig. 3). Postoperatively, rhythmic pontine spike trains still occurred, strengthening our interpretation that "drowsy wakefulness" was, in fact, a form of waking. Furthermore, active sleep was also present, with its flat EMG and episodic pontine spike bursts. The mechanisms responsible for linking neck muscle atonia to the subcortically organized state of active sleep appear to remain intact after lateral hypothalamic damage, thus making the interpretation
331 of active sleep without atonia even less likely. However, "drowsy wakefulness" and active sleep in our rats should be explored further. Shivering should not occur, if drowsy wakefulness is, in fact, active sleep without atonia [11,28]. Furthermore, in recovery, more organized behavioral acts should occur during drowsy wakefulness as it recovers toward wakefulness; conversely, if it recovers toward active sleep, organized behaviors should diminish, and neck muscle EMG atonia should increase. Comparing the present report and other reports from our laboratory [6, 9, 21, 56] with reports of other investigators , it appears that our lateral hypothalamic damaged rats show, not only akinesia, catalepsy, sensory neglect and aphagia and adipsia, but also "somnolence." This may be due to differences in the size of the lesion, ours presumably being larger. It may also be due to differences in the observation and interpretation of behavior. Other investigators have tested lateral hypothalamic-damaged rats in running wheels and challenged them in water tanks, thus eliciting running and swimming . Such tests demonstrate that these animals can be galvanized into action, perhaps even to normal levels. On the other hand, when one observes the spontaneous behavior of these animals, they are sluggish and akinetic. When there is spontaneous behavior in these animals it is of short duration and tends quickly to be replaced by sleep. Therefore, external sources of activation are necessary to maintain wakeful, highly organized, environmentally-directed behavior in these animals. We see their dependence on external activation as reflecting a low level of endogenous activation or arousal. Throughout this paper, we have regarded the persistent, unchanging high amplitude slow cortical activity of the LHdamaged rat as indicating that the cortex is functionally disconnected from the subcortical and peripheral events that do continue to reflect changes in states of waking and sleep. In other words, we take the cortex to be isolated from subcortical input or output, being unresponsive to, and having no effect on those alternating subcortical events. This may be incorrect. The cortex might be changing its activities in ways not reflected by our recording methods, and might also be exerting active inhibitory actions on subcortical systems. (We thank Allan Rechtschaffen of the University of Chicago for suggesting this.) We do not rule out inhibitory actions, but until we can detect those actions, we prefer the simpler assumption that the cortex is merely passively isolated by the lesions, functionally disconnected from the subcortical events that go on without being modulated by it. In summary, after lateral hypothalamic damage in rats, three subcortically organized alternating states of sleep and waking were identified, functionally disconnected from cortical activity, which remains slow and unchanging during what has previously been considered an unchanging state of persistent deep sleep, labelled as "somnolence." Within the state of "somnolence," not only are there periods of quiet and active sleep, but also periods of organized behavioral acts, accompanied by the same subcortical and muscular signs which, as an aggregate, identify waking. We call this subcortically organized, low-arousal form of waking, "drowsy-wakefulness."
1. Anand, B. K. and J. R. Brobeck. Hypothalamic control of food intake in rats and cats. Yale J. Biol. Med. 24: 123-140, 1951.
2. Bizzi, E. and D. C. Brooks. Functional connections between pontine reticular formation and lateral geniculate nucleus during deep sleep. Archs ital. Biol. 101: 666--680, 1963.
S H O H A M A N I ) t !:I! !-~! B gl.: ~,,~
3~2 3. Bremer, F. Preoptic hypnogenic focus and mesencephalic reticular formation. Brain Reds. 21: 132-134, 1970. 4. Cairns, H., R. C. Oldfield, J. B. Pennybacker and D. Whitteridge. Akinetic mutism with an epidermoid cyst of the 3rd ventricle. Brain 64: 273-290, 1941. 5. Danguir, J. and S. Nicolaidis. Cortical activity and sleep in the rat hypothalamic syndrome. Brain Res. 195: 305-321, 1980. 6. De Ryck, M. and P. Teitelbaum. Neocortical and hippocampal EEG in normal and lateral hypothalamic-damaged rats. Physiol. Behav. 20: 403-409, 1978. 7. Epstein, A. N. The lateral hypothalamic syndrome: Its implications for the physiological psychology of hunger and thirst. In: Progress in Psyehobiology and Physiological Psychology, vol. 4, edited by E. Stellar and J. M. Sprague. New York: Academic Press, 1971, pp. 263-317. 8. Farber, J., G. A. Marks and H. P. Roffwarg. Rapid eye movement sleep PGO-type waves are present in the dorsal pons of the albino rat. Science 209: 615-618, 1980. 9. Golani, I., D. L. Wolgin and P. Teitelbaum. A proposed natural geometry of recovery from akinesia in the lateral hypothalamic rat. Brain Res. 164: 237-267, 1979. 10. Guilleminault, C., R. Philips and W. C. Dement. A syndrome of hypersomnia with automatic behavior. Eleetroeneeph. elin. Neurophysiol. 38: 403-413, 1975. 11. Hendricks, J. C., R. M. Bowker and A. R. Morrison. Functional characteristics of cats with pontine lesions during sleep and wakefulness and their usefulness for sleep research. In: Sleep 1976. edited by W. Koella and P. Levin. Basel: Karger, 1977, pp. 207-210. 12. Henley, K. and A. R. Morrison. A reevaluation of the effects of lesions of the pontine tegmentum and locus coeruleus on phenomena of paradoxical sleep in the cat. Acta neurobiol, exp. 34: 215-232, 1974. 13. Hobson, J. A. The effects of chronic brain-stem lesions on cortical and muscular activity during sleep and waking in the cat. Eleetroeneeph. olin. Neurophysiol. 19: 41-62, 1965. 14. Ingram, W. R., J. R. Knott, M. D. Wheatley and T. D. Summers. Physiological relationships between hypothalamus and cerebral cortex. Eleetroeneeph. olin. Neurophysiol. 3: 37-57, 1951. 15. Jones, B. E., S. T. Harper and A. E. Halaris. Effects of locus coeruleus lesions upon cerebral monoamine content, sleepwakefulness states and the response to amphetamine in the cat. Brain Res. 124: 473-496, 1977. 16. Jouvet, M. and F. Michel. Correlations electromyographiques du sommeil chez le Chat decortique et mesencephalique. C.r. S~;ane. Soe. Biol. 153: 422-425, 1959. 17. Jouvet, M. Recherches sur les structures nerveuses et les mecanismes responsables des differentes phases du sommeil physiologique. Archs ital. Biol. 100: 125-206, 1962. 18. Kawamura, H. and E. F. Domino. Hippocampal slow ("arousal") wave activation in the rostral midbrain transected cat. Eleetroenceph. clin. Neurophysiol. 25: 471-480, 1968. 19. Kolb, B. and I. Q. Whishaw. Effects of brain lesions and atropine on hippocampal and neocortical electroencephalograms in the rat. Expl Neurol. 56: 1-22, 1977. 20. Krnig, F. F. R. and R. A. Klippel. The Rat Brain. Baltimore: Williams and Wilkins, 1963. 21. Levitt, D. R. and P. Teitelbaum. Somnolence, akinesia and sensory activation of motivated behavior in the lateral hypothalamic syndrome. Proe. natn. Acad. Sci. U.S.A. 72: 281%2823, 1975. 22. Marshall, J. F., J. S. Richardson and P. Teitelbaum. Nigrostriatal bundle damage and the lateral hypothalamic syndrome. J. eomp. physiol. Psychol. 87: 808-830, 1974. 23. Marshall, J. F., B. H. Turner and P. Teitelbaum. Sensory neglect produced by lateral hypothalamic damage. Science 174: 523-525, 1971. 24. Marshall, J. F. and P. Teitelbaum. Further analysis of sensory inattention following lateral hypothalamic damage in rats. J. eomp. physiol. Psyehol. 86: 375-395. 1974.
25. McGinty, D. J. Somnolence. recovery and t~3 po~ol~ni~ iollo~ ing ventromedial diencephalic lesion~ in the r:~ 1~/,~.~.,.,,,.m~p; olin, Neurophvsi, d. 26: 70-79. 1969, 26. Moruzzi. G. The nleep waking cycle, l:~Tz,'/,,. //~, ~,~/ 64: t~.t,, 1972.
27. Nauta, W. J. H. Hypothalamic regulation of ~ieep ~n v:~ls. Al~ experimental study. J. Nettrophysiol. 9:285-316. 1946. 28. Parmeggiani, P. L. Interaction between sleep and thermoregn. lation. Waking Sh'epin~, I : 123-132, 1977. 29. Ranson, S. W. Somnolence caused by hypothalamic lesions m the monkey. Arclt~' Nearol. Psy('hiat.. Chiea~,,o 41: 1-23, 1939. 30. Robinson, T. E., R. C. Kramis and C. H. Vanderwolf. Two types of cerebral activation during active sleep: relations to behavior. Brain Re.~. 124: 544-549, 1977. 31. Robinson, T. E. and I. Q. Whishaw. Effects of posterior hypothalamic lesions on voluntary behavior and hippocampal electroencephalograms in the rat. J. ~,,mp. physiol P~v~.h,d. 86: 768-786, 1974. 32. Rosenberg, R. S., H. Zepelin and A. Rechtschaffen. Sleep in young and old rats..I. Geront. 34: 525-532, 1979. 33. Ruckebusch, Y. The relevance of drowsiness in the circadian cycle of farm animals. Anita. Behav. 20: 637--643. 1972. 34. Schallert, T., M. De Ryck and P. Teitelbaum. Atropine stereotypy as a behavioral trap: A movement subsystem and EEG analysis. ,I. co,tp, phy.~iol. P.~ychol. 94: 1-24, 1980. 35. Schallert, T., 1. Q. Whishaw, M. De Ryck and P. Teitelbaum The postures of catecholamine-depletion catalepsy: Their possible adaptive value in thermoregulation. Phy.siol. Behav. 21: 817-820, 1978. 36. Skinner, J. E. and D. B. Lindsley. Electrophysiological and behavioral effects of blockade of the nonspecific thalamocortical system. Brain Res. 6:95-118. 1967. 37. Sterman, M. B. and C. D. Clemente. Forebrain inhibitory mechanisms: sleep patterns induced by basal forebrain mechanisms underlying deep sleep in the cat. Archs' ira/ Biol. 99: 270--292, 1961. 38. Sterman, M. B., T. Knauss, L. Lehmann and C. D. Clementc. Circadian sleep and waking patterns in the laboratory cat. Ele,troenceph, olin. Neurophysiol. 19: 50%517, 1965. 39. Swett, C. P. and J. A. Hobson. The effects of posterior hypothalamic lesions on behavioral and electrographic manifestations of sleep and waking in cats. Arehs ital. Biol. 106: 283-293, 1968. 40. Teitelbaum, P. Levels of integration of the operant. In: Handboo~ ,)fOperant Behavior. edited by W. K. Honig and J. E. R Staddon. Englewood Cliffs, N J: Prentice-Hall, 1977, pp. 7-27. 41. Teitelbaum, P. Disconnection and antagonistic interaction of movement subsystems in motivated behavior, t~roc¢.eding.~ ,~1" the Fir~t Institute o f Neurological Scien('e.~: ~%'mp,~.sium ,m Neurohiology. October, 1980, in press. 42. Teitelbaum, P. and A. N. Epstein. The lateral hypothalamic syndrome: Recovery of feeding and drinking after lateral hypothalamic lesions. Psy~hol. Rev. 69: 74-90, 1962. 43. Teitelbaum, P., T. Schallert, M. De Ryck, I. Q. Whishaw and 1. Golani. Motor subsystems in motivated behavior, ln: Neural Mechanisms ~1 Goal-Directed Behavi~, and Learning,,. edited by R. F. Thompson, L. H. Hicks and V. B. Shvyrkov. New York: Academic Press, 1980, pp. 127-t43. 44. Teitelbaum, P. and E. Stellar. Recovery from the failure to eat produced by hypothalamic lesions. Scie,ce 120: 894-895, 1954. 45. Timo-Iaria, C.. N. Negrad, W. R. Schmidek, K. Hoshino, C. Lobato de Menezes and T. Leme da Rocha. Phases and states of sleep in the rat. Physiol. Behav. 5: 1057-1062, 1970. 46. Vanderwolf, C. H. Neocortical and hippocampal activation in relation to behavior: effects of atropine, eserine, phenothiazines and amphetamine. J. eomp. physiol. Psychol. 88: 300-328, 1975. 47. Vanderwolf, C. H., R. C. Kramis, L. A. Gillespie and B. H. Bland. Hippocampal rhythmical slow activity and neocortical low voltage fast activity: relations to behavior. In: The Hippoearnpus: A Comprehensive Treatise. edited by R. L. Isaacson and K. H. Pribram. New York: Plenum Press, 1975, pp. 101-128.
SLEEP AND WAKING DURING "SOMNOLENCE"
48. Villablanca, J. Electroencephalograms in the permanently isolated forebrain of the cat. Science 138:44 46, 1962. 49. Villablanca, J. Behavioral and polygraphic study of "sleep" and "wakefulness" in chronic decerebrate cats. Electroenceph. clin. Neurophysiol. 21: 562-577, 1966. 50. Villablanca, J. and M. E. Salinas-Zeballos. Sleep-wakefulness, EEG, and behavioral studies of chronic cats without the thalamus: The 'athalamic' cat. Archs ital. Biol. 110: 383--411, 1972. 51. Villablanca, J. and J. Schlag. Cortical control of thalamic spindle waves. Expl Neurol. 20: 432-444, 1968. 52. Vital-Durand, F. and F. Michel. Effets de la desafferentation peripherique sur le cycle veille-sommeil chez le chat. Archs ital. Biol. 109: 166-186, 1971. 53. Whishaw, I. Q., T. E. Robinson, T. Schallert, M. De Ryck and V. D. Ramirez. The electrical activity of the hippocampus and neocortex in rats depleted of brain dopamine and norepinephrine: relations to behavior and effects of atropine. Expl Neurol. 62: 748-767, 1978.
54. Whishaw, I. Q. and C. H. Vanderwolf. Hippocampal EEG and behavior: changes in amplitude and frequency of RSA (Theta Rhythm) associated with spontaneous and learned movement patterns in rats and cats. Behav. Biol. 8: 461-484, 1973. 55. Wolgin, D. L., J. Cytawa and P. Teitelbaum. The role of activation in the regulation of food intake. In: Hunger: Basic Mechanisms and Clinical Implications, edited by D. Novin, W. Wyrwicka and G. Bray. New York: Raven Press, 1976, pp. 179-191. 56. Wolgin, D. L. and P. Teitelbaum. Role of activation and sensory stimuli in recovery from lateral hypothalamic damage in the cat. J. comp. physiol. Psychol. 92: 474-500, 1978. 57. Wright, J. J. and M. D. Craggs. Changes in cortical activation and the lateral hypothalamic syndrome: A study in the split brain cat. Brain Res. 151: 632-636, 1978.