Sleep, 2(2):161-173
© 1979 Raven Press, New York
Activity of Human Hippocampal Formation and Amygdala Neurons During Sleep Luigi Ravagnati, Eric Halgren, Thomas L. Babb, and Paul H. Crandall Reed Neurological Research Center, Brain Research Institute and Department of SlIrgery/Nellrological, Unil'ersity of Calij()rnia School of Medicine, Los Angeles, California
Summary: Fine wire microelectrodes were implanted for diagnostic purposes in 17 patients suffering from psychomotor epilepsy, Single- and multiunit activity during waking and natural nocturnal slow wave sleep and REM sleep was recorded in the hippocampus (n = 42), hippocampal gyrus (n = 53), and amygdala (n = 32). The firing rates of hippocampal gyrus units usually decreased during slow wave sleep and then increased to levels equal to or above waking during REM. In contrast, the firing rates of hippocampal neurons generally increased during slow wave sleep and fell to very low levels during REM. The amygdala presented a more mixed pattern. Since the projection from the hippocampal gyrus to hippocampus is excitatory, their opposite patterns during sleep suggest that the tonic firing patterns of He neurons may be mainly the result of subcortical afferents. Key Words: Sleep-HippocampusAmygdala-Unit activity-REM sleep.
The human hippocampal formation is necessary for the recall of recent events (Milner, 1968), an essential component ofthe manifest content of dreams (Calkins, 1893; Freud, 1950). Electrical stimulation or pathological irritation of this region in the waking human occasionally evokes vivid experiences similar to dreams in manifest and latent content and in subjective quality (Ferguson et al., 1969; Jackson and Colman, 1898; Penfield and Perot, 1963). More intense stimulation evokes more bizarre and more dreamlike hallucinations (Halgren et al., 1978b). These facts suggest the hypothesis that the human hippocampal formation becomes highly activated during REM sleep as compared with slow wave sleep (SWS) or with a waking period not involving recent memory or bizarre thought processes. Direct evidence for this hypothesis is sparse. Previous observations in humans Accepted for publication September 1979. Dr. Ravagnati's present address is University of Milan, Milan, Italy. Address reprint requests to Dr. Halgren at Brain Research Institute 73-364 CHS, UCLA, Los Angeles, California 90024.
161 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
162
L. RAVAGNATI ET AL.
is limited to Moiseeva and Aleksanyan's (1976) report that multiunit activity in the hippocampus is greatest during waking, least during SWS, and intermediate during REM. However, this conclusion is apparently based on only a single or a few· observations. More extensive animal studies have failed to yield consistent findings (see Discussion), and, in any case, their applicability to humans is questioned by the strong species differences in the behavioral physiology of the hippocampal formation (Halgren et aI., 1977b, 1978a). The present paper appears to be the first full study of human hippocampal multi- and single-unit activity recorded during different stages of natural sleep and wakefulness. We have found that the neural activity of the hippocampus appears to be greatly reduced during REM, whereas the activity of its primary afferent structure, the hippocampal gyrus, is generally enhanced. METHODS Recording Procedures Patient characteristics, surgical technique, electrode localization, and recording methods have been described in detail in previous papers (Babb and Crandall, 1976; Halgren et aI., 1977a, 1978b). Briefly, 17 psychomotor epileptics (5 female) without severe intellectual or personality handicaps were each chronically implanted with 14 gross electrodes in order to localize their epileptic foci (6 right temporal lobe, 5 left TL, 3 bitemporal, 3 extra temporal).! The stereotactic procedure, based on direct visualization of the lateral ventricle, possessed an overall accuracy of approximately 88% according to histological examination of excised lobes from 13 patients.ofthis series (Halgren et aI., 1978b). Electrodes directed toward the hippocampal gyrus usually terminated in the entorhinal cortex, and amygdala electrodes lay in the basolateral nuclear complex. Hippocampal electrodes were in all fields of Ammon's horn and in the dentate gyrus. Some of the electrodes were hollow cannulae through which were inserted bundles of 40 ILm wires, insulated except for the tip. The activity from 106 of these microelectrodes displaying action potentials with the greatest signal-to-noise ratio was recorded on a wide-band 14-track tape recorder for off-line analysis. Direct confirmation ofthe location of the microelectrodes was not possible because their tracks were not visible histologically. Sleep Recordings Sleep recordings began at approximately 11 p.m. and continued until the patient had completed one or two complete sleep cycles (4-8 hr). Active awake periods were obtained by engaging the patient in mental arithmetic, spelling, or divergent production tasks for 3 - 5 min immediately preceding "lights-out" or immediately on awakening. Eye movements and submental muscle activity were recorded
I The procedures were approved by the Human Subjects Protection Committee of this School of Medicine under guidelines established by the National Institutes of Health. Patients were fully informed of the steps in the research procedures both during interviews and in explicit written form.
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
HUMAN LIMBIC UNITS DURING SLEEP
/63
according to the standard procedures of Rechtschaffen and Kales (1968). In addition to limbic depth activity, cortical electroencephalogram (EEG) was recorded from nails placed during surgery into the outer table of the skull and referenced to each other (e.g., T3-T5), or, in later patients, to the contralateral earlobe (C4- AI). Behavioral state was also monitored by recording the patient's voice and, in later patients, via a closed-circuit low-illumination television system. In all cases, sleep staging according to the guidelines of Rechtschaffen and Kales (1968) was attempted. However, due primarily to the epileptic pathology, but also probably to their varying medications and to the inevitably disruptive, uncomfortable, and stressful hospital routine, it was not always possible to unambiguously distinguish between different stages of NREM sleep. For example, it was not always possible to distinguish slow waves due to sleep from those due to medication or pathology. Therefore, for quantitative analysis we considered only one stage of NREM sleep: "slow wave sleep" with the apparent behavioral and electrographic characteristics of stage 3 and/or stage 4 sleep. Data Analysis Neural activity was recorded simultaneously from 6 to 10 microelectrodes in each patient. This activity was passed through a high-pass filter into a pulse height window discriminator. From 1 to 3 nonoverlapping windows were utilized per electrode. Only about 1110 of the electrodes recorded action potentials whose constant waveform and amplitude, and clear separation from lower amplitude action potentials, permitted triggering action potentials from single neurons. In all cases, the minimum level of the lowest window was greater than twice the amplitude of the background multiunit "hash." The output of each window discriminator therefore represents the action potentials of one to a few neurons and is termed here a "unit." Forty-two hippocampal, 53 hippocampal gyrus, and 32 amygdala units were recorded. The outputs of the window discriminators (1 msec pulses) were played out together with the recorded EEG, electro-oculogram, and electromyogram onto a pen-writing oscillograph running at a slow chart speed (0.3-1.5 mmlsec) for direct visual analysis. These records plus the 15 mmlsec tracings acquired on-line were used to choose periods representative of the various sleep stages for quantitative analysis. The analyzed periods were 3-10 min in duration, lacked movement artifact, and were homogeneous in depth of sleep. The unit activity during each period was analyzed with the aid ofa PDP-12 computer, which provided the mean and standard deviation of the interspike intervals, an interval histogram, and autoand cross-correlograms (Woods et aI., 1975a,b; Wyss and Handwerker, 1971). The means and standard deviations of the interspike intervals from different periods of sleep were compared using Student's I-test (two-tailed). Epilepsy All of our patients possessed epileptic pathology which might be expected to alter to an unknown degree the activity of their neurons during sleep. Therefore,
Sleep, Vol. 2. No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
L. RA VAGNATJ ET AL.
164
TABLE 1. Neuronal discharge rates in different limbic structures in SWS as compared with REM SWS < REM"
No change b
15
6
12
8
22
13
33 43
6
8
9
23
Structure
SWS > REM"
Hippocampus Hippocampal gyrus Amygdala " p /, p
Total
< 0.01, Student's t-test. > 0.01.
The distribution in this table differs from a random distribution (p < 0.05, X2 = 10.71, df = 4).
units which, according to visual analysis of fast and slow polygraph recordings, grossly increased or decreased their firing rate during or following epileptiform EEG spikes were not included in this quantitative analysis. With these units excluded, there were no obvious differences in the firing patterns during different sleep stages of units recorded either contralateral to an identified temporal lobe focus as compared to those lying ipsilateral to the focus, or of units recorded in patients with foci outside of the temporal lobes. These findings will be described in more detail in a subsequent paper (unpublished data). RESULTS Firing-Rate Hippocampal units and hippocampal gyrus units generally changed firing rates in opposite directions during the sleep cycle. Seventy-one percent of responsive hippocampal units decreased their firing from SWS to REM, whereas 73% of responsive hippocampal gyrus units increased their firing rate (Table 1). The discharge rate of neurons in all structures during stage REM was found to be more similar to that during waking (A W) than during stage 3 or 4 sleep (SWS). Of the 73 units recorded during a complete sleep cycle, 56% showed a significant (p < 0.01) change in firing rate at both the transition from A W to SWS and from SWS to REM. In 78% ofthese responsive units, the change from A W to SWS was in the opposite direction from the change from SWS to REM (Figs. 1 and 2). Plots of the mean firing rates of all recorded units in each structure, individually (Fig. 1) and averaged (Fig. 2), support these same conclusions. Specifically, limbic units recorded showed significant changes in firing rate between SWS and REM, and these changes were usually moderate increases of hippocampal gyrus and amygdala activity and often striking decreases in hippocampal unit activity. Examples of simultaneously recorded changes in neuronal discharges, slow waves, and behavioral measures across representative epochs are shown in Fig. 3. A complete sleep cycle is illustrated in Fig. 4. Although it was not possible, due to the low amplitude or multiunit nature of many of our recordings, to conscientiously determine whether each of our hippocampal recordings displayed bursts of spikes of decreasing amplitude, it should
Sleep, Vol. 2. No.2. 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
HUMAN LIMBIC UNITS DURING SLEEP HC
HCG
10
165 AM
o 1.0 z o
u w
(f)
.....
(f)
w
'"0.::
(f)
0.1
.01
AW
SWS REM
FIG. 1. Mean firing rates for each unit recorded during both SWS and REM (n = 99: not all units were recorded during both periods). Units which decreased firing during this transition are illustrated on the left, together with the firing rates of these same units during waking. Units which increased firing from SWS to REM are illustrated in the right graphs. Transitions which were significant at the 0.01 level (two-tailed I-test) are drawn with solid lines. AW, waking; SWS, slow wave sleep; REM, rapid eye movement sleep; HCG, hippocampal gyrus; HC, hippocampus; AM, amygdala .
.300 HCG .250
.200 "0
c:
0
'-' Q) V>
"en .150 w :> A:::
0
A.
... ...e
t:::l
"
~
~ .N
~
-
•N
~
Q
~
0
'"
~
W~
1-111 Z
0
32
64
US
256
§
TIME IM5EC)
FIG. 5. Autocorrelograms from single units in the left anterior hippocampus ofRG (patient 71) and the left posterior hippocampal gyrus of AJ (patient 64) during stage I, SWS, and stage REM. Note that the autocorrelograms do not differ across stages in RG. A tracing of the typical burst firing of that neuron is shown as an insert above the SWS autocorrelogram. In AJ, a burst firing pattern is evident in the autocorrelogram during SWS but not during stage REM. Average firing rates: RG, stage 1-1.3; SWS, 1.0; REM, 0.5; AJ stage I, 16.6; SWS, 6.5; REM, Il.l spikes/sec .
'0
:(;l
Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
.......
0\ '0
170
L. RA VAGNATI ET AL.
fornix has come to be appreciated as containing HCF afferents as well as efferents. These afferents include cholinergic fibers from the medial septum and noradrenergic fibers from the locus coeruleus that, while small in number, ramify widely throughout the HC (Kuhar, 1975; Moore, 1975; Jones et aI., 1977; Jones and Moore, 1977). Evidence from animals suggests that these fibers may induce a tonic inhibition of HC neurons during REM, overriding their increased excitatory input from the perforant path. REM in rats is characterized by a tonic cholinergic theta rhythm (Robinson et aI., 1977) and when the septal source of this rhythm is removed, the firing rate of HC neurons during REM increases sevenfold (Franzini et aI., 1975). Locus coeruleus cells fire slowly during SWS and in rapid bursts during REM (Chu and Bloom, 1973), and the effects of these volleys appear from stimulation experiments to be inhibitory (Segal and Bloom, 1974; Finch et aI., 1978). In several respects, a recent study by Winson and Abzug (1978) ofthe effectiveness of perforant path stimulation during different stages of sleep and wakefulness is complimentary to the present study. They found that the compound action potential recorded ~t various points within the HC (CAl, CA3, and dentate gyrus) to perforant path stimulation is greatly enhanced in SWS as compared with REM and A W. This effect is not presynaptic, as this potential, recorded before the dentate gyrus synapse, is unaffected by behavioral state. Furthermore, the field potential representing the summed excitatory postsynaptic potentials evoked monosynaptically in the dentate gyrus by perforant path stimulation is actually depressed during SWS. This led Winson and Abzug to suggest that HC neurons are tonically depolarized during SWS and/or hyperpolarized during REM and A W. Our observations of spontaneous firing patterns provide direct support for this hypothesis. Conversely, Winson and Abzug's findings directly confirm our suggestion that the perforant path is ineffective in conveying tonic HCG excitatory influences to the HC during REM and A W. Both studies suggest that tonic levels of HC firing are controlled by subcortical afferents. Comparison with Findings in Animals The large changes in spontaneous firing rate in human mesial temporal lobe neurons correlated with the various phases of sleep conform to the general pattern for brain unit activity described in lower mammals (Jacobs et aI., 1973): the firing rate during REM is more similar to A W than to SWS. In fact, if SWS is considered unaroused and A W moderately aroused, then our study would indicate that the limbic neuronal activity during REM is suggestive of a hyperarousai. Averaged neuronal firing rates that fell during SWS rose to supra waking levels during REM (Fig. 2: HCG and AM). Conversely, increased firing rates in SWS were followed by subwaking levels during REM (Fig. 2: HC). Animal studies of HC firing rates during different stages of sleep have yielded contradictory findings. Olmstead et al. (1973) found that neurons in the regio superior of the rat HC fire more rapidly during SWS than during REM, whereas regio inferior and fascia dentata cells fire most rapidly during REM. In contrast, Ranck (1973) found that rat HC neurons in all areas responded similarly, the "complex spike" cells (probably pyramidal or granule cells: Fox and Ranck, 1975) firing
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
HUMAN LIMBIC UNITS DURING SLEEP
171
most rapidly during SWS and the "theta" cells (possibly interneurons) firing most rapidly during REM and waking. Noda et al. (1969) reported yet a third result, in cats: all neurons fired more rapidly in REM than in SWS. As Ranck pointed out, the discrepancy between his findings and those of Noda et al. may only be apparent, inasmuch as Noda et al. apparently preselected for HC neurons with consistently high firing rates, which Ranck has shown in rats are usually theta cells. Our findings are most in accord with Ranck's: the majority of our HC units (and all those displaying complex spikes) declined during REM. However, since we lacked fine histological localization and most of our recordings were not single unit, it is possible that, in fact, our findings were due to a fortuitous concentration of our microwires in a particular architectonic field. The mixed response of our amygdala units, located mainly or entirely in the basolateral complex, correspond~ with the mixed response reported in the cat basolateral amygdala by Jacobs and McGinty (1971). However, our lack of single-unit recordings prevented us from testing their suggestion that the changes in firing rate were strongly correlated with the absolute level of spontaneous firing. Finally, although studies describing the firing rate during sleep of mammalian HCG neurons appear to be lacking, our HCG units conform to the decrease in SWS followed by an increase during REM that is generally characteristic of mammalian neocortical neurons (Hobson, 1972). Auto- and Cross-Correlation Inasmuch as, in general, our recordings were of small groups of neurons rather than individual cells, interpretation of the auto- and cross-correlation findings is somewhat problematical. Suffice it to say that the cross-correlation findings indicated, in all sleep stages, that the degree to which nearby neuronal groups fired in relation to each other was greater for the subcortical structures (HC and amygdala) than for the HCG. The autocorrelation data suggest that in all structures the degree to which each neuron in a group tended to fire in close synchrony with the other neurons or with itself (in a burst) was greater during SWS than during A W and REM, providing a cellular correlate for the EEG synchrony present during SWS (unpublished data). ACKNOWLEDGMENT We thank Jeff Lieb, Charles Wilson, and Jonathan Winson for critical comments, Elmo Mariani and especially Everett Carr for technical assistance, Steven Woods for computational assistance, Joan Hopgood and Shirley Jennings for typing the manuscript, and our patients for their active cooperation. This work was supported by National Science Foundation Grant BNS 77-17070 and by V.S.P.H.S. Grant NS 02808. Additionally, E.H. was supported by the Ralph Smith Foundation and L.R. by the Italian CNR. REFERENCES Andersen, P. Organization of hippocampal neurons and their interconnections. In: RL Isaacson and KH Pribram (Eds), The Hippocampus, Vol. /: Structure and Development. Plenum Press, New York, 1975, pp 155-175.
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
172
L. RA VAGNATI ET AL.
Babb TL and Crandall PH. Epileptogenesis of human limbic neurons in psychomotor epileptics. Electroencephalogr Clin Neurophysiol 90:225-243, 1976. Calkins NW. Statistics of dreams. Am J Psychol 5:311-343, 1893. Chu N-S and Bloom PE. Norepinephrine-containing neurons: Changes in spontaneous discharge patterns during sleeping and waking. Science 179:908-918, 1973. Ferguson SM, Rayport M, Gardner R, Kass W, Weiner H, and Reiser MF. Similarities in mental content of psychotic states, spontaneous seizures, dreams, and response to electrical brain stimulation in patients with temporal lobe epilepsy. Psychosom Med 31:479-498, 1969. Finch DM, Feld RE, and Babb TL. Effects of mesencephalic and pontine electrical stimulation on hippocampal neuronal activity in drug-free cat. Exp Neurol 61:318-336, 1978. Fox SE and Ranck JB, Jr. Localization and anatomical identification of theta and complex spike cells in dorsal hippocampal formation of rats. Exp Neurol 49:299- 3\3, 1975. Franzini C, Calasso M, and Parmeggiani PL. Septal influences on the discharge of hippocampal CA-1 complex spike neurones during sleep. In: P Levin and WP Koella (Eds), Sleep 1974. Karger, Basel, 1975, pp 259-262. Freud S. The Interpretation of Dreams. Modern Library, 1950. Halgren E, Babb TL, and Crandall PH. Human hippocampal EEG desynchronizes during attentiveness and movement. Electroencephalogr Clin Neurophysiol 44:778-781, 1978a. Halgren E, Babb FL, and Crandall PH. Responses of human limbic neurons to induced changes in blood gases. Brain Res 132:43-63, 1977a. Halgren E, Babb TL, Rausch R, and Crandall PH. Neurons in the human basolateral amygdala and hippocampal formation do not respond to odors. Neurosci Lett 4:331-335, 1977b. Halgren E, Walter RD, Cherlow DG, and Crandall PH. Mental phenomena evoked by electrical stimulation of the human hippocampal formation and amygdala. Brain 101:83-117, 1978b. Hobson JA. Cellular neurophysiology and sleep. In: MH Chase (Ed), The Sleeping Brain. UCLA Brain Information Service, Los Angeles, 1972, pp 59-84. Jackson BL and Colman WS. Case of epilepsy with tasting movements and "dreamy-state"-very small patch of softening in the left uncinate gyrus. Brain 21:580-590, 1898. Jacobs BL and McGinty DJ. Amygdala unit activity during sleep and waking. Exp Neurol 33: 1-15, 1971. Jacobs BL, McGinty DJ, and Harper RM. Brain single unit activity during sleep-wakefulness-A review. In: MI Phillips (Ed), Brain Unit Activity During Behavior. Charles C Thomas, Springfield, 111.,1973, pp 165-178. . Jones BE, Halaris AE, Mcllhany M, and Moore RY. Ascending projections of the locus coeruleus in the rat. I. Axonal transport in central noradrenergic neurons. Brain Res 127:1-21,1977. Jones BE and Moore RY. Ascending projections of the locus coeruleus in the rat. II. Autoradiographic study. Brain Res 127:23-53, 1977. Kuhar MJ. Cholinergic neurons: Septal-hippocampal relationships. In: RL Isaacson and KH Pribram (Eds), The Hippocampus, Vol. I: Structure and Development. Plenum Press, New York, 1975, pp 269-283. Meibach RC and Siegel A. Thalamic projections of the hippocampal formation: Evidence for an alternate pathway involving the internal capsule. Brain Res 134:1-12, 1977. Milner B (Ed). Disorders of memory after brain lesions in man. Neuropsychology 6: 175-291, 1968. Moiseeva NI and Aleksanyan ZA. Activity of neuronal populations of human subcortical structures during sleep. Electroencephalogr Clin Neurophysiol 41:467-475, 1976. Moore RY. Monoamine neurons innervating the hippocampal formation and septum: Organization and response to injury. In: RL Isaacson and KH Pribram (Eds), The Hippocampus, Vol. I: Structure and Development. Plenum Press, New York, 1975, pp 215-237. Noda H, Manohar S, and Adey WR. Spontaneous activity of cat hippocampal neurons in sleep and wakefulness. Exp NeuroI24:217-231, 1969. Olmstead CE, Best Pl, and Mays LE. Neural activity in the dorsal hippocampus during paradoxical sleep, slow wave sleep, and waking. Brain Res 60:381-391, 1973. Penfield WP and Perot P. The brain's record of auditory and visual experience. A final summary. Brain 86:595-696, 1963. Ranck JB Jr. Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats. I. Behavioral correlates and firing repertoires. Exp Neurol 41:461-531,1973. Rechtschaffen A and Kales A (Eds). A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. UCLA Brain Information Service/Brain Research Institute, Los Angeles, 1968. Robinson TE, Kramis RC, and Vanderwolf CH. Two types of cerebral activation during active sleep: Relations to behavior. Brain Res 124:544-549, 1977.
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
i
.'
HUMAN LIMBIC UNITS DURING SLEEP
173
Rosene DL and Van Hoesen OW, Hippocampal efferents reach widespread areas of cerebral cortex and amygdala in the Rhesus monkey, Science 198:315-317, 1977. Segal M and Bloom FE. The action of norepinephrine in the rat hippocampus. II. Activation of the input pathway. Brain Res 72:99-114,1974. Seltzer B and Pandya DN. Some cortical projections to the parahippocampal area in the Rhesus monkey. Exp Neurol 50: 146-160, 1976. Swanson LW and Cowan WM. An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat. J Comp Neurol 172:49-84, 1977. Umbach W. Long-term results offornicotomy for temporal epilepsy. Conjin Neurol27: 121-213, 1966. Van Hoesen OW and Pandya DN. Some connections of the entorhinal (area 28) and perirhinal cortices of the Rhesus monkey. I. Temporal lobe afferents. Brain Res 95: 1-24, 1975. Van Hoesen OW, Pandya DN, and Butters N. Some connections of the entorhinal (area 28) and perirhinal (area 35) cortices of the Rhesus monkey. II. Frontal lobe afferents. Brain Res 95:25-38, 1975. Winson J and Abzug C. Neuronal transmission through hippocampal pathways dependent on behavior..1 NeurophysioI41:716-732, 1978, Woods SC, Babb TL, Halgren E, and Perga A. TROSYS: A multichannel fast point process data acquisition system. Digital Equipment Corp Users Society Program No. 12-170, 1975a. Woods SC, Williams JT, Halgren E, and Babb TL. HSTPL T: A versatile program for cross-correlation of point-process data on a PDP-12. Decuscope 14: 12-13, 1975b. Woolsey RM and Nelson JS. Asymptomatic destruction of the fornix in man. Arch Neural 32:566-568, 1975. Wyss UR and Handwerker H. ST AP- 12: A library system for on-line assimilation and off-line analysis of event/time data. Comput Programs Biomed 1:309- 138, 1971.
Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
© 1979 Raven Press, New York
Activity of Human Hippocampal Formation and Amygdala Neurons During Sleep Luigi Ravagnati, Eric Halgren, Thomas L. Babb, and Paul H. Crandall Reed Neurological Research Center, Brain Research Institute and Department of SlIrgery/Nellrological, Unil'ersity of Calij()rnia School of Medicine, Los Angeles, California
Summary: Fine wire microelectrodes were implanted for diagnostic purposes in 17 patients suffering from psychomotor epilepsy, Single- and multiunit activity during waking and natural nocturnal slow wave sleep and REM sleep was recorded in the hippocampus (n = 42), hippocampal gyrus (n = 53), and amygdala (n = 32). The firing rates of hippocampal gyrus units usually decreased during slow wave sleep and then increased to levels equal to or above waking during REM. In contrast, the firing rates of hippocampal neurons generally increased during slow wave sleep and fell to very low levels during REM. The amygdala presented a more mixed pattern. Since the projection from the hippocampal gyrus to hippocampus is excitatory, their opposite patterns during sleep suggest that the tonic firing patterns of He neurons may be mainly the result of subcortical afferents. Key Words: Sleep-HippocampusAmygdala-Unit activity-REM sleep.
The human hippocampal formation is necessary for the recall of recent events (Milner, 1968), an essential component ofthe manifest content of dreams (Calkins, 1893; Freud, 1950). Electrical stimulation or pathological irritation of this region in the waking human occasionally evokes vivid experiences similar to dreams in manifest and latent content and in subjective quality (Ferguson et al., 1969; Jackson and Colman, 1898; Penfield and Perot, 1963). More intense stimulation evokes more bizarre and more dreamlike hallucinations (Halgren et al., 1978b). These facts suggest the hypothesis that the human hippocampal formation becomes highly activated during REM sleep as compared with slow wave sleep (SWS) or with a waking period not involving recent memory or bizarre thought processes. Direct evidence for this hypothesis is sparse. Previous observations in humans Accepted for publication September 1979. Dr. Ravagnati's present address is University of Milan, Milan, Italy. Address reprint requests to Dr. Halgren at Brain Research Institute 73-364 CHS, UCLA, Los Angeles, California 90024.
161 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
162
L. RAVAGNATI ET AL.
is limited to Moiseeva and Aleksanyan's (1976) report that multiunit activity in the hippocampus is greatest during waking, least during SWS, and intermediate during REM. However, this conclusion is apparently based on only a single or a few· observations. More extensive animal studies have failed to yield consistent findings (see Discussion), and, in any case, their applicability to humans is questioned by the strong species differences in the behavioral physiology of the hippocampal formation (Halgren et aI., 1977b, 1978a). The present paper appears to be the first full study of human hippocampal multi- and single-unit activity recorded during different stages of natural sleep and wakefulness. We have found that the neural activity of the hippocampus appears to be greatly reduced during REM, whereas the activity of its primary afferent structure, the hippocampal gyrus, is generally enhanced. METHODS Recording Procedures Patient characteristics, surgical technique, electrode localization, and recording methods have been described in detail in previous papers (Babb and Crandall, 1976; Halgren et aI., 1977a, 1978b). Briefly, 17 psychomotor epileptics (5 female) without severe intellectual or personality handicaps were each chronically implanted with 14 gross electrodes in order to localize their epileptic foci (6 right temporal lobe, 5 left TL, 3 bitemporal, 3 extra temporal).! The stereotactic procedure, based on direct visualization of the lateral ventricle, possessed an overall accuracy of approximately 88% according to histological examination of excised lobes from 13 patients.ofthis series (Halgren et aI., 1978b). Electrodes directed toward the hippocampal gyrus usually terminated in the entorhinal cortex, and amygdala electrodes lay in the basolateral nuclear complex. Hippocampal electrodes were in all fields of Ammon's horn and in the dentate gyrus. Some of the electrodes were hollow cannulae through which were inserted bundles of 40 ILm wires, insulated except for the tip. The activity from 106 of these microelectrodes displaying action potentials with the greatest signal-to-noise ratio was recorded on a wide-band 14-track tape recorder for off-line analysis. Direct confirmation ofthe location of the microelectrodes was not possible because their tracks were not visible histologically. Sleep Recordings Sleep recordings began at approximately 11 p.m. and continued until the patient had completed one or two complete sleep cycles (4-8 hr). Active awake periods were obtained by engaging the patient in mental arithmetic, spelling, or divergent production tasks for 3 - 5 min immediately preceding "lights-out" or immediately on awakening. Eye movements and submental muscle activity were recorded
I The procedures were approved by the Human Subjects Protection Committee of this School of Medicine under guidelines established by the National Institutes of Health. Patients were fully informed of the steps in the research procedures both during interviews and in explicit written form.
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
HUMAN LIMBIC UNITS DURING SLEEP
/63
according to the standard procedures of Rechtschaffen and Kales (1968). In addition to limbic depth activity, cortical electroencephalogram (EEG) was recorded from nails placed during surgery into the outer table of the skull and referenced to each other (e.g., T3-T5), or, in later patients, to the contralateral earlobe (C4- AI). Behavioral state was also monitored by recording the patient's voice and, in later patients, via a closed-circuit low-illumination television system. In all cases, sleep staging according to the guidelines of Rechtschaffen and Kales (1968) was attempted. However, due primarily to the epileptic pathology, but also probably to their varying medications and to the inevitably disruptive, uncomfortable, and stressful hospital routine, it was not always possible to unambiguously distinguish between different stages of NREM sleep. For example, it was not always possible to distinguish slow waves due to sleep from those due to medication or pathology. Therefore, for quantitative analysis we considered only one stage of NREM sleep: "slow wave sleep" with the apparent behavioral and electrographic characteristics of stage 3 and/or stage 4 sleep. Data Analysis Neural activity was recorded simultaneously from 6 to 10 microelectrodes in each patient. This activity was passed through a high-pass filter into a pulse height window discriminator. From 1 to 3 nonoverlapping windows were utilized per electrode. Only about 1110 of the electrodes recorded action potentials whose constant waveform and amplitude, and clear separation from lower amplitude action potentials, permitted triggering action potentials from single neurons. In all cases, the minimum level of the lowest window was greater than twice the amplitude of the background multiunit "hash." The output of each window discriminator therefore represents the action potentials of one to a few neurons and is termed here a "unit." Forty-two hippocampal, 53 hippocampal gyrus, and 32 amygdala units were recorded. The outputs of the window discriminators (1 msec pulses) were played out together with the recorded EEG, electro-oculogram, and electromyogram onto a pen-writing oscillograph running at a slow chart speed (0.3-1.5 mmlsec) for direct visual analysis. These records plus the 15 mmlsec tracings acquired on-line were used to choose periods representative of the various sleep stages for quantitative analysis. The analyzed periods were 3-10 min in duration, lacked movement artifact, and were homogeneous in depth of sleep. The unit activity during each period was analyzed with the aid ofa PDP-12 computer, which provided the mean and standard deviation of the interspike intervals, an interval histogram, and autoand cross-correlograms (Woods et aI., 1975a,b; Wyss and Handwerker, 1971). The means and standard deviations of the interspike intervals from different periods of sleep were compared using Student's I-test (two-tailed). Epilepsy All of our patients possessed epileptic pathology which might be expected to alter to an unknown degree the activity of their neurons during sleep. Therefore,
Sleep, Vol. 2. No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
L. RA VAGNATJ ET AL.
164
TABLE 1. Neuronal discharge rates in different limbic structures in SWS as compared with REM SWS < REM"
No change b
15
6
12
8
22
13
33 43
6
8
9
23
Structure
SWS > REM"
Hippocampus Hippocampal gyrus Amygdala " p /, p
Total
< 0.01, Student's t-test. > 0.01.
The distribution in this table differs from a random distribution (p < 0.05, X2 = 10.71, df = 4).
units which, according to visual analysis of fast and slow polygraph recordings, grossly increased or decreased their firing rate during or following epileptiform EEG spikes were not included in this quantitative analysis. With these units excluded, there were no obvious differences in the firing patterns during different sleep stages of units recorded either contralateral to an identified temporal lobe focus as compared to those lying ipsilateral to the focus, or of units recorded in patients with foci outside of the temporal lobes. These findings will be described in more detail in a subsequent paper (unpublished data). RESULTS Firing-Rate Hippocampal units and hippocampal gyrus units generally changed firing rates in opposite directions during the sleep cycle. Seventy-one percent of responsive hippocampal units decreased their firing from SWS to REM, whereas 73% of responsive hippocampal gyrus units increased their firing rate (Table 1). The discharge rate of neurons in all structures during stage REM was found to be more similar to that during waking (A W) than during stage 3 or 4 sleep (SWS). Of the 73 units recorded during a complete sleep cycle, 56% showed a significant (p < 0.01) change in firing rate at both the transition from A W to SWS and from SWS to REM. In 78% ofthese responsive units, the change from A W to SWS was in the opposite direction from the change from SWS to REM (Figs. 1 and 2). Plots of the mean firing rates of all recorded units in each structure, individually (Fig. 1) and averaged (Fig. 2), support these same conclusions. Specifically, limbic units recorded showed significant changes in firing rate between SWS and REM, and these changes were usually moderate increases of hippocampal gyrus and amygdala activity and often striking decreases in hippocampal unit activity. Examples of simultaneously recorded changes in neuronal discharges, slow waves, and behavioral measures across representative epochs are shown in Fig. 3. A complete sleep cycle is illustrated in Fig. 4. Although it was not possible, due to the low amplitude or multiunit nature of many of our recordings, to conscientiously determine whether each of our hippocampal recordings displayed bursts of spikes of decreasing amplitude, it should
Sleep, Vol. 2. No.2. 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
HUMAN LIMBIC UNITS DURING SLEEP HC
HCG
10
165 AM
o 1.0 z o
u w
(f)
.....
(f)
w
'"0.::
(f)
0.1
.01
AW
SWS REM
FIG. 1. Mean firing rates for each unit recorded during both SWS and REM (n = 99: not all units were recorded during both periods). Units which decreased firing during this transition are illustrated on the left, together with the firing rates of these same units during waking. Units which increased firing from SWS to REM are illustrated in the right graphs. Transitions which were significant at the 0.01 level (two-tailed I-test) are drawn with solid lines. AW, waking; SWS, slow wave sleep; REM, rapid eye movement sleep; HCG, hippocampal gyrus; HC, hippocampus; AM, amygdala .
.300 HCG .250
.200 "0
c:
0
'-' Q) V>
"en .150 w :> A:::
0
A.
... ...e
t:::l
"
~
~ .N
~
-
•N
~
Q
~
0
'"
~
W~
1-111 Z
0
32
64
US
256
§
TIME IM5EC)
FIG. 5. Autocorrelograms from single units in the left anterior hippocampus ofRG (patient 71) and the left posterior hippocampal gyrus of AJ (patient 64) during stage I, SWS, and stage REM. Note that the autocorrelograms do not differ across stages in RG. A tracing of the typical burst firing of that neuron is shown as an insert above the SWS autocorrelogram. In AJ, a burst firing pattern is evident in the autocorrelogram during SWS but not during stage REM. Average firing rates: RG, stage 1-1.3; SWS, 1.0; REM, 0.5; AJ stage I, 16.6; SWS, 6.5; REM, Il.l spikes/sec .
'0
:(;l
Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
.......
0\ '0
170
L. RA VAGNATI ET AL.
fornix has come to be appreciated as containing HCF afferents as well as efferents. These afferents include cholinergic fibers from the medial septum and noradrenergic fibers from the locus coeruleus that, while small in number, ramify widely throughout the HC (Kuhar, 1975; Moore, 1975; Jones et aI., 1977; Jones and Moore, 1977). Evidence from animals suggests that these fibers may induce a tonic inhibition of HC neurons during REM, overriding their increased excitatory input from the perforant path. REM in rats is characterized by a tonic cholinergic theta rhythm (Robinson et aI., 1977) and when the septal source of this rhythm is removed, the firing rate of HC neurons during REM increases sevenfold (Franzini et aI., 1975). Locus coeruleus cells fire slowly during SWS and in rapid bursts during REM (Chu and Bloom, 1973), and the effects of these volleys appear from stimulation experiments to be inhibitory (Segal and Bloom, 1974; Finch et aI., 1978). In several respects, a recent study by Winson and Abzug (1978) ofthe effectiveness of perforant path stimulation during different stages of sleep and wakefulness is complimentary to the present study. They found that the compound action potential recorded ~t various points within the HC (CAl, CA3, and dentate gyrus) to perforant path stimulation is greatly enhanced in SWS as compared with REM and A W. This effect is not presynaptic, as this potential, recorded before the dentate gyrus synapse, is unaffected by behavioral state. Furthermore, the field potential representing the summed excitatory postsynaptic potentials evoked monosynaptically in the dentate gyrus by perforant path stimulation is actually depressed during SWS. This led Winson and Abzug to suggest that HC neurons are tonically depolarized during SWS and/or hyperpolarized during REM and A W. Our observations of spontaneous firing patterns provide direct support for this hypothesis. Conversely, Winson and Abzug's findings directly confirm our suggestion that the perforant path is ineffective in conveying tonic HCG excitatory influences to the HC during REM and A W. Both studies suggest that tonic levels of HC firing are controlled by subcortical afferents. Comparison with Findings in Animals The large changes in spontaneous firing rate in human mesial temporal lobe neurons correlated with the various phases of sleep conform to the general pattern for brain unit activity described in lower mammals (Jacobs et aI., 1973): the firing rate during REM is more similar to A W than to SWS. In fact, if SWS is considered unaroused and A W moderately aroused, then our study would indicate that the limbic neuronal activity during REM is suggestive of a hyperarousai. Averaged neuronal firing rates that fell during SWS rose to supra waking levels during REM (Fig. 2: HCG and AM). Conversely, increased firing rates in SWS were followed by subwaking levels during REM (Fig. 2: HC). Animal studies of HC firing rates during different stages of sleep have yielded contradictory findings. Olmstead et al. (1973) found that neurons in the regio superior of the rat HC fire more rapidly during SWS than during REM, whereas regio inferior and fascia dentata cells fire most rapidly during REM. In contrast, Ranck (1973) found that rat HC neurons in all areas responded similarly, the "complex spike" cells (probably pyramidal or granule cells: Fox and Ranck, 1975) firing
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
HUMAN LIMBIC UNITS DURING SLEEP
171
most rapidly during SWS and the "theta" cells (possibly interneurons) firing most rapidly during REM and waking. Noda et al. (1969) reported yet a third result, in cats: all neurons fired more rapidly in REM than in SWS. As Ranck pointed out, the discrepancy between his findings and those of Noda et al. may only be apparent, inasmuch as Noda et al. apparently preselected for HC neurons with consistently high firing rates, which Ranck has shown in rats are usually theta cells. Our findings are most in accord with Ranck's: the majority of our HC units (and all those displaying complex spikes) declined during REM. However, since we lacked fine histological localization and most of our recordings were not single unit, it is possible that, in fact, our findings were due to a fortuitous concentration of our microwires in a particular architectonic field. The mixed response of our amygdala units, located mainly or entirely in the basolateral complex, correspond~ with the mixed response reported in the cat basolateral amygdala by Jacobs and McGinty (1971). However, our lack of single-unit recordings prevented us from testing their suggestion that the changes in firing rate were strongly correlated with the absolute level of spontaneous firing. Finally, although studies describing the firing rate during sleep of mammalian HCG neurons appear to be lacking, our HCG units conform to the decrease in SWS followed by an increase during REM that is generally characteristic of mammalian neocortical neurons (Hobson, 1972). Auto- and Cross-Correlation Inasmuch as, in general, our recordings were of small groups of neurons rather than individual cells, interpretation of the auto- and cross-correlation findings is somewhat problematical. Suffice it to say that the cross-correlation findings indicated, in all sleep stages, that the degree to which nearby neuronal groups fired in relation to each other was greater for the subcortical structures (HC and amygdala) than for the HCG. The autocorrelation data suggest that in all structures the degree to which each neuron in a group tended to fire in close synchrony with the other neurons or with itself (in a burst) was greater during SWS than during A W and REM, providing a cellular correlate for the EEG synchrony present during SWS (unpublished data). ACKNOWLEDGMENT We thank Jeff Lieb, Charles Wilson, and Jonathan Winson for critical comments, Elmo Mariani and especially Everett Carr for technical assistance, Steven Woods for computational assistance, Joan Hopgood and Shirley Jennings for typing the manuscript, and our patients for their active cooperation. This work was supported by National Science Foundation Grant BNS 77-17070 and by V.S.P.H.S. Grant NS 02808. Additionally, E.H. was supported by the Ralph Smith Foundation and L.R. by the Italian CNR. REFERENCES Andersen, P. Organization of hippocampal neurons and their interconnections. In: RL Isaacson and KH Pribram (Eds), The Hippocampus, Vol. /: Structure and Development. Plenum Press, New York, 1975, pp 155-175.
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
172
L. RA VAGNATI ET AL.
Babb TL and Crandall PH. Epileptogenesis of human limbic neurons in psychomotor epileptics. Electroencephalogr Clin Neurophysiol 90:225-243, 1976. Calkins NW. Statistics of dreams. Am J Psychol 5:311-343, 1893. Chu N-S and Bloom PE. Norepinephrine-containing neurons: Changes in spontaneous discharge patterns during sleeping and waking. Science 179:908-918, 1973. Ferguson SM, Rayport M, Gardner R, Kass W, Weiner H, and Reiser MF. Similarities in mental content of psychotic states, spontaneous seizures, dreams, and response to electrical brain stimulation in patients with temporal lobe epilepsy. Psychosom Med 31:479-498, 1969. Finch DM, Feld RE, and Babb TL. Effects of mesencephalic and pontine electrical stimulation on hippocampal neuronal activity in drug-free cat. Exp Neurol 61:318-336, 1978. Fox SE and Ranck JB, Jr. Localization and anatomical identification of theta and complex spike cells in dorsal hippocampal formation of rats. Exp Neurol 49:299- 3\3, 1975. Franzini C, Calasso M, and Parmeggiani PL. Septal influences on the discharge of hippocampal CA-1 complex spike neurones during sleep. In: P Levin and WP Koella (Eds), Sleep 1974. Karger, Basel, 1975, pp 259-262. Freud S. The Interpretation of Dreams. Modern Library, 1950. Halgren E, Babb TL, and Crandall PH. Human hippocampal EEG desynchronizes during attentiveness and movement. Electroencephalogr Clin Neurophysiol 44:778-781, 1978a. Halgren E, Babb FL, and Crandall PH. Responses of human limbic neurons to induced changes in blood gases. Brain Res 132:43-63, 1977a. Halgren E, Babb TL, Rausch R, and Crandall PH. Neurons in the human basolateral amygdala and hippocampal formation do not respond to odors. Neurosci Lett 4:331-335, 1977b. Halgren E, Walter RD, Cherlow DG, and Crandall PH. Mental phenomena evoked by electrical stimulation of the human hippocampal formation and amygdala. Brain 101:83-117, 1978b. Hobson JA. Cellular neurophysiology and sleep. In: MH Chase (Ed), The Sleeping Brain. UCLA Brain Information Service, Los Angeles, 1972, pp 59-84. Jackson BL and Colman WS. Case of epilepsy with tasting movements and "dreamy-state"-very small patch of softening in the left uncinate gyrus. Brain 21:580-590, 1898. Jacobs BL and McGinty DJ. Amygdala unit activity during sleep and waking. Exp Neurol 33: 1-15, 1971. Jacobs BL, McGinty DJ, and Harper RM. Brain single unit activity during sleep-wakefulness-A review. In: MI Phillips (Ed), Brain Unit Activity During Behavior. Charles C Thomas, Springfield, 111.,1973, pp 165-178. . Jones BE, Halaris AE, Mcllhany M, and Moore RY. Ascending projections of the locus coeruleus in the rat. I. Axonal transport in central noradrenergic neurons. Brain Res 127:1-21,1977. Jones BE and Moore RY. Ascending projections of the locus coeruleus in the rat. II. Autoradiographic study. Brain Res 127:23-53, 1977. Kuhar MJ. Cholinergic neurons: Septal-hippocampal relationships. In: RL Isaacson and KH Pribram (Eds), The Hippocampus, Vol. I: Structure and Development. Plenum Press, New York, 1975, pp 269-283. Meibach RC and Siegel A. Thalamic projections of the hippocampal formation: Evidence for an alternate pathway involving the internal capsule. Brain Res 134:1-12, 1977. Milner B (Ed). Disorders of memory after brain lesions in man. Neuropsychology 6: 175-291, 1968. Moiseeva NI and Aleksanyan ZA. Activity of neuronal populations of human subcortical structures during sleep. Electroencephalogr Clin Neurophysiol 41:467-475, 1976. Moore RY. Monoamine neurons innervating the hippocampal formation and septum: Organization and response to injury. In: RL Isaacson and KH Pribram (Eds), The Hippocampus, Vol. I: Structure and Development. Plenum Press, New York, 1975, pp 215-237. Noda H, Manohar S, and Adey WR. Spontaneous activity of cat hippocampal neurons in sleep and wakefulness. Exp NeuroI24:217-231, 1969. Olmstead CE, Best Pl, and Mays LE. Neural activity in the dorsal hippocampus during paradoxical sleep, slow wave sleep, and waking. Brain Res 60:381-391, 1973. Penfield WP and Perot P. The brain's record of auditory and visual experience. A final summary. Brain 86:595-696, 1963. Ranck JB Jr. Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats. I. Behavioral correlates and firing repertoires. Exp Neurol 41:461-531,1973. Rechtschaffen A and Kales A (Eds). A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. UCLA Brain Information Service/Brain Research Institute, Los Angeles, 1968. Robinson TE, Kramis RC, and Vanderwolf CH. Two types of cerebral activation during active sleep: Relations to behavior. Brain Res 124:544-549, 1977.
Sleep, Vol. 2, No.2, 1979 Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
i
.'
HUMAN LIMBIC UNITS DURING SLEEP
173
Rosene DL and Van Hoesen OW, Hippocampal efferents reach widespread areas of cerebral cortex and amygdala in the Rhesus monkey, Science 198:315-317, 1977. Segal M and Bloom FE. The action of norepinephrine in the rat hippocampus. II. Activation of the input pathway. Brain Res 72:99-114,1974. Seltzer B and Pandya DN. Some cortical projections to the parahippocampal area in the Rhesus monkey. Exp Neurol 50: 146-160, 1976. Swanson LW and Cowan WM. An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat. J Comp Neurol 172:49-84, 1977. Umbach W. Long-term results offornicotomy for temporal epilepsy. Conjin Neurol27: 121-213, 1966. Van Hoesen OW and Pandya DN. Some connections of the entorhinal (area 28) and perirhinal cortices of the Rhesus monkey. I. Temporal lobe afferents. Brain Res 95: 1-24, 1975. Van Hoesen OW, Pandya DN, and Butters N. Some connections of the entorhinal (area 28) and perirhinal (area 35) cortices of the Rhesus monkey. II. Frontal lobe afferents. Brain Res 95:25-38, 1975. Winson J and Abzug C. Neuronal transmission through hippocampal pathways dependent on behavior..1 NeurophysioI41:716-732, 1978, Woods SC, Babb TL, Halgren E, and Perga A. TROSYS: A multichannel fast point process data acquisition system. Digital Equipment Corp Users Society Program No. 12-170, 1975a. Woods SC, Williams JT, Halgren E, and Babb TL. HSTPL T: A versatile program for cross-correlation of point-process data on a PDP-12. Decuscope 14: 12-13, 1975b. Woolsey RM and Nelson JS. Asymptomatic destruction of the fornix in man. Arch Neural 32:566-568, 1975. Wyss UR and Handwerker H. ST AP- 12: A library system for on-line assimilation and off-line analysis of event/time data. Comput Programs Biomed 1:309- 138, 1971.
Downloaded from https://academic.oup.com/sleep/article-abstract/2/2/161/2749904 by guest on 24 January 2018
