- Email: [email protected]
Thalamic kindling: Electrical stimulation of the lateral geniculate nucleus produces photosensitive grand mal seizures
EXPERIMENTAL
NEUROLOGY
86, 18-32 (1984)
Thalamic Kindling: Electrical Stimulation of the Lateral Geniculate Nucleus Produces Photosensitive Grand Mal Seizures M. N. SHOUSE AND W. RYAN’ Research
Service, Veterans Administration Medical Center, Sepulveda, California 91343 and School of Medicine, University of California. Los Angeles, Calijornia 90024 Received
October
5, 1983: revision
received
May
7. 1984
Kindling is traditionally viewed as a chronic, focal epilepsy model which consistently induces complex partial seizures from limbic structures in animals. This study revealed that primary or exceedingly rapid secondary generalized seizures could also be kindled when stimulation was applied to the lateral geniculate nucleus, a thalamic region involved in sleep regulation and possibly also photosensitive epilepsy. Two experiments were conducted in cats. Experiment I compared the development of generalized tonic-clonic convulsions and associated sleep disorders following electrical stimulation of the lateral geniculate nucleus (N = 4) and the amygdala (N = 4). Experiment 2 described the effects of intermittent light stimulation on seizure thresholds in both groups. Three primary findings distinguished the epileptogenic process in those two brain regions. First, generalized electroencephalographic and clinical seizures accompanied the first afterdischarge obtained with thalamic stimulation. In contrast, focal seizures with secondary generalization appeared during a 3to 4-week period of afterdischarge elicitations from the amygdala. Second, amygdalakindled cats showed fewer sleep spindles during slow-wave sleep whereas cats kindled in the lateral geniculate nucleus had abnormal sleep spindles approaching spike wave-like activity. Third, only the latter cats showed reduced seizure thresholds in response to photic stimulation. Based on the anatomic substrates involved, the clinical and electrographic profiles observed during kindling and the type of sleep disturbance shown, we concluded that lateral geniculate nucleus kindling may represent primary generalized epilepsy, possibly of a photosensitive nature; alternatively, the rapid propagation of abnormal discharge was also consistent with the important role of the thalamus in secondary seizure generalization.
Abbreviations: LGN-lateral geniculate nucleus, ILS-intermittent light stimulation, SWSslow-wave sleep, REM-rapid eye-movement. ’ This research was supported by the Veterans Administration, Sepulveda, California.
18 0014-4886/84 $3.00 Copyright 63 1984 by Academic Press, Inc. All rights of reproduction in any form reserved
THALAMIC
19
KINDLING
INTRODUCTION Since its discovery in 1967 by Goddard ( I3), kindling has been construed primarily as a chronic, focal model of epilepsy resembling complex partial seizures in humans (22, 29). This view is supported by the fact that repetitive, electrical or chemical stimulation of virtually all forebrain sites, especially in the limbic system, initially produces focal afterdischarge (AD) and seizures, both of which gradually become secondarily generalized (13, 14, 22, 29). In contrast, the thalamus and the brain stem have been resistant to kindling (7, 13, 14, 29); historically, these structures have been more intimately involved in primary generalized epilepsies (6, 9-l 2, 15- 17, 2628), especially of the “petit mal” variety. A possible explanation for the recalcitrance of brain stem and thalamic nuclei to kindling may reflect different etiologies and anatomic substrates for primary and secondary generalized seizure disorders. A related observation concerns the sleep modulation functions of many brain stem and nonspecific thalamic structures vis-a-vis the stimulus variables used in the kindling procedure. Whereas sleep disruption accompanies limbic system kindling (30, 33, 34, 36, 42) the application of AD-eliciting stimulation to the brain stem, for example, produces normal sleep instead of clinical seizure manifestations (7). We have previously suggested that normal sleep (39) and seizure induction may be incompatible processes (34-36, 4 1). To test these hypotheses, we attempted kindling in the thalamus. The lateral geniculate nucleus (LGN), a specific sensory relay nucleus, was chosen as a target site rather than a nonspecific nucleus in order to avoid the complications cited above. Accordingly, this report documents clinical and electrographic seizure manifestations as well as the sleep anomalies associated with LGN kindling compared with amygdala kindling in cats. METHODOLOGY
AND
RESULTS
Two consecutive experiments were conducted in the same animals. Experiment 1 characterized the development of a seizure condition using elecrical kindling procedures applied to either the LGN or amygdala. In addition, the effects of these seizures on sleep and waking state patterns were also documented. Experiment 2 was undertaken because of the visual functions of the LGN and examined the effects of intermittent light stimulation (ILS) on seizure thresholds in LGN- and amygdala-kindled animals. The procedures and results of these two experiments are described separately except for histologic findings, which are described at the end of this section.
20
SHOUSE
AND
Experiment
RYAN
1
Stereotaxic surgery was carried out on 12 adult cats (1.8 to 4.2 kg) under sodium pentobarbital anesthesia for electrode placements (37) used in kindling and state evaluation. The details of electrode construction and use in kindling and sleep-waking state evaluation were described elsewhere (32-36, 43). The montage included: bilateral, tripolar leads in the LGN (A 6.0; L 10.5; H 2, 2.5, 3.0) and basolateral amygdala (A 11.0, L 9.5, H -5.5, -6.0, -5.5); jeweler’s screws threaded into the bone over sensorimotor cortex (A 23.0; L 8 and 10) bilaterally for surface EEGs and into the frontal sinus for an EOG; and silver wire electrodes inserted into the neck musculature for an EMG. After a 2-week postoperative recovery, all cats underwent two, 12-h polygraphic recordings to establish baseline state measurements. Afterward, the cats were divided into three groups (N = 4 each) according to the type of brain stimulation received: group lamygdala kindling; group 2-LGN kindling; group 3-no brain stimulation (control). At the beginning of kindling, initial AD thresholds were established using a standard method of limits procedure (32-36, 45). A 100~PA stimulus (ls train of 60 Hz, biphasic square waves of I-ms pulse duration) was applied on day 1 and 200 PA on day 2. Subsequently, amperage was increased by 200 FA each day until the AD appeared and then decreased by 100 PA each day until the AD disappeared. Threshold was therefore defined as the minimum intensity eliciting an AD within 100 PA. Afterward, a single daily stimulation was applied at AD threshold until generalized tonicclonic convulsions (stage 6 seizures) were obtained. Stage 6 seizure thresholds were then established using the same method of limits procedure as at the beginning of kindling. Finally, additional 12-h polygraphic recordings were obtained during and at the end of kindling, in relation to initial and final seizure threshold determinations. Statistical analysis consisted of two-way repeated measures analyses of variance (group X time) applied to kindling and sleep state measurements. Dependent variables for kindling consisted of (a) the number of subthreshold vs. threshold stimulations required to obtain stage 6 seizures and (b) initial AD thresholds vs. stage 6 seizure thresholds (PA). Sleep state variables included (a) the mean percentage time spent in slow-wave sleep (SWS) and rapid-eye-movement (REM) sleep and (b) the number of sleep spindles during comparable, 5-min SWS samples derived from the first segment of the second sleep cycle. Post hoc individual comparisons were evaluated with Student’s t tests. Kindling the Lateral Geniculate Nucleus vs. the Amygdala. Table 1 shows
THALAMIC
KINDLING
21
the findings for the kindling variables. More subthreshold stimulations (P < 0.1) were required to elicit initial ADS from the LGN than from the amygdala (Table 1A). Amygdala-stimulated cats initially displayed considerable variability in initial AD (33) as did LGN-stimulated animals. Thus, in spite of an apparently sizeable difference in the number of subthreshold stimulations to initial AD between the two groups, excessive variability in values, as reflected by high standard deviations (Table lA), precluded a statistically significant effect, and only a statistical trend was obtained. However, when AD appeared in the LGN cats, significantly fewer threshold stimulations were needed to obtain generalized seizures (P < 0.001). Compared with the 3- to 4-week interval required to elicit stage 6 convulsions from the amygdala, cats receiving LGN stimulation showed stage 6 seizures on the occasion of their first AD. Figure 1 illustrates the type of EEG generalization obtained after stimulating the two regions. Even at the end of amygdala kindling, predominantly focal AD occurred first in the amygdala ipsilateral to the kindling electrode and was associated with unilateral facial twitching; AD then spread to other regions before culminating in a stage 6 seizure. In contrast, the first threshold stimulation in LGN cats resulted in generalized AD, which approximated bilateral synchrony. Although the amplitude of the ADS was largest in the LGN, AD appeared to occur simultaneously in the LGN, amygdala and over the sensorimotor cortex in 3 of 4 cats; in the remaining animal, cortical AD developed 2 to 3 s after it occurred in deep leads. Typically, these ADS consisted of 4 to 5 c/s spikeand-wave although 3 c/s activity was also observed. The initial seizure manifestation was bilateral eye blinking, which was invariably followed by a generalized tonic-clonic convulsion. Vocalization usually occurred at the end of the seizure, and autonomic symptoms such as salivation and urine incontinence frequently occurred during the seizure. Initial AD thresholds were higher in LGN animals than in the amygdalakindled group (P < 0.1; Table 1B) but were not outside the range of initial AD thresholds obtained from these and larger populations of amygdalakindled cats (32-36, 45). For this reason, only a statistical trend was shown by our data. It is remarkable, however, that no reduction in AD or generalized seizure threshold was seen after LGN kindling, as was customarily observed when initial AD and final stage 6 seizure thresholds were compared in amygdala-kindled cats [P < 0.00 1; Table 1B; ( 14, 2 l-36, 45)]. Stage 6 seizure thresholds were therefore significantly higher in the LGN group compared with values for the amygdala cats (P < 0.001). Further, stable LGN (and amygdala) seizure thresholds were maintained at the same level as before during three additional seizure threshold determinations conducted immediately afterward.
SHOUSE AND RYAN
22
TABLE I Kindling Values for Two Groups of Cats Receiving Stimulation in the Lateral Geniculate Nucleus (LGN) or Amygdala” A. Number of stimulations to kindling Experimental groups
Subthreshold stimuli
Threshold (AD) stimuli
B. Seizure thresholds (PA) Initial AD threshold
Stage 6 seizure threshold
Amygdala kindling
(N= 4)
3.25f 2
24.2k 3***
600+100
275+ 70***
LGN kindling
(N= 4)
6.5 + 3**
1.0 f o*s***
1275f 618**
1275+ 618***
’ A-the means f SD number of subthreshold and threshold [after discharge (AD)-eliciting] stimuli required to obtain stage 6 or generalized tonic-clonic convulsions. B-the means f SD stimulus threshold-inducing AD at the beginning of kindling and stage 6 seizures at the end of kindling. Significant difference between groups: *P < 0.001; **P < 0.1. Significant difference within cats in the same group: ***P c 0.001.
Kindling E&cts on Sleep States. An analysis of sleep state parameters (Table 2) revealed a progressive and significant reduction in SWS and REM sleep time during amygdala kindling. This was accompanied by a reduction in the density of sleep spindles (number of sleep spindles per unit time, visually counted by two independent observers), which were of normal configuration before and after kindling. These postkindling changes confirm previous work (32-36, 40) and showed sleep spindle reduction only during the generalized phases of amygdala kindling. Postkindling reductions in sleep spindles were statistically significant compared with initial baselines in amygdala-kindled cats as well as with values derived from control animals. In contrast, cats receiving LGN stimulation showed a decrement only in REM sleep. It was difficult to compare sleep spindle incidence in the baseline and postkindling records in LGN-stimulated cats. In postkindling records sleep spindles were abnormal in appearance; the 12 to 15 c/s spindle activity of the prekindling records was intermixed with slower activity (8 to 10 c/s) in postkindling records. This altered rhythmic activity may reflect an intermediate form between sleep spindles and a spike wave-pattern even though no interictal discharge was apparent in the postkindling sleep records (Fig. 2). Experiment 2 Experiment 2 examined the effects of photic stimulation on seizure thresholds in both groups of kindled animals and began 10 days after the
THALAMIC
KINDLING
A. AMYGDALA
STIMULATION
6. LGN
23 %cL”
STIMULATION
LT AMYG LT LGN LT LGN FIT LGN LT SM CX RT SM CX RT LGN
t STIMULUS
25
FIG. 1. Representative electroencephalographic responses to electrical stimulation (l-s, 60Hz, biphasic square waves) at intensities (PA) just sufficient to elicit afterdischarge (AD) and generalized tonic-clonic convulsions (stage 6 seizures) from the left amygdala (A) and the left lateral geniculate nucleus (B). Amygdala (AMYG) stimulation evoked AD first in the amygdala ipsilateral to the kindling electrode followed by secondary generalization to other brain areas. Note some initial AD also over the left sensorimotor cortex (LT SM CX). In contrast, stimulation of the lateral geniculate nucleus (LGN) produced generalized AD accompanied by a stage 6 seizure.
last seizure threshold determination was made at the end of Experiment 1. In this experiment, stage 6 seizure thresholds were reestablished using the same procedures as in Experiment 1 except that all stimuli were applied on the same day, with a 5-min interstimulus interval. This modified procedure was used in previous studies of seizure threshold in kindled cats (32, 35). A within-subjects paradigm was used in which seizure thresholds were established on three consecutive occasions, with a lo-day intertrial interval. The first experimental condition (A,) provided an initial baseline seizure threshold without photic stimulation. During the second condition, (B,),
SHOUSE AND RYAN
24
TABLE 2 Means + SD Percentage Time Spent in Slow-Wave Sleep (SWS) and Rapid Eye-Movement (REM) Sleep before and at the End ofAmygdala and Lateral Geniculate Nucleus (LGN) Kindling sws Experimental groups Amygdala kindling (N= 4) LGN kindling (N= 4) Control (N=
4)
REM sleep
Before kindling
After kindling
Before kindling
After kindling
33 f 8
24.0 2 3*,**
15 *2
10 * 4****
31*7
30
*8
16 +4
10 + 5****
30 + 6
35
+2
14+
I
16 -t 2
* P < 0.05 compared with control values. ** P < 0.05 compared with initial baseline values.
photic stimulation (1.5 Hz) was administered through the window of the recording chamber 2 min prior to and throughout the threshold testing procedure. The last condition (AZ) provided a final baseline seizure threshold without photic stimulation. The results of Experiment 2 are summarized in Fig. 3, which shows stage 6 seizure thresholds in the two groups. There were two primary findings. First, LGN-stimulated cats continued to show higher seizure thresholds than amygdala-kindled cats regardless of the experimental condition. Further,
A. BEFORE
B. AFTER
LGN
LGN
KINDLING
KINDLING
FIG. 2. Representative samples of slow-wave sleep before (A) and after (B) lateral geniculate nucleus (LGN) kindling. It is more difficult to identify the sleep spindles in the postkindling record because their frequency is unusually slow. This abnormal configuration may represent an intermediate form between sleep spindles and spike wave-like discharge.
THALAMIC
NO
PHOTIC
PHOTIC
25
KINDLING
PHOTIC
STIMULATION
NO PHOTlC
CONDITIONS
FIG. 3. Mean seizure threshold (PA) and standard deviations obtained from cats kindled through leads in the lateral geniculate nucleus (LGN) and basolateral amygdala (N = 4 each). Threshold determinations were made on three consecutive occasions with a IO-day interval elapsing between each test. The A, and A2 conditions provided baseline values obtained without photic stimulation. During the intervening B, condition, the animals were treated identically except that intermittent light stimulation (ILS at 1.5 Hz) was presented 2 min prior to and throughout the method of limits threshold procedure. The two primary findings were (i) that seizure thresholds were consistently higher in LGN cats compared with amygdala-kindled animals and (P < 0.001) and (ii) that ILS significantly reduced seizure thresholds only in LGNkindled cats when B, values were compared with initial (A,) and (A*) baselines (P < 0.001).
baseline LGN seizure thresholds (A, and Az) were not significantly different, substantiating the rather unique stability of the high kindled seizure thresholds in these animals. Second, ILS markedly and selectively reduced seizure thresholds in LGN cats, comparing B, values with initial (A,) and final (A*) baselines. This 50% reduction in LGN seizure thresholds was
26
SHOUSE AND RYAN
comparable in the four animals tested and occurred regardless of whether the animal looked directly at the flashing light or had its eyes open or closed. Histology
The accuracy of the LGN placements was verified by selective spiking in that nucleus during REM sleep and during ILS; LGN and amygdala placements were also confirmed by subsequent histological examination. Figure 4 illustrates electrode sites in the LGN. Placements were centrally situated in the nucleus in three cats. Stimulating electrodes were situated somewhat more ventrally in the remaining cat, which showed a delay between subcortical and cortical AD onset. DISCUSSION The primary finding was the discovery of generalized seizures systematically induced by kindling the LGN, a specific thalamic relay nucleus. In addition, REM sleep loss and atypical sleep spindles were observed after LGN kindling, an outcome which is consistent with the hypothesis that sleep and seizure disturbances are related phenomena (32, 34-36). Finally, the sensitivity to ILS shown after LGN kindling is compatible with the visual functions of this nucleus. These observations with cats extend those of Caine (2), who first attempted kindling of specific thalamic nuclei, including the lateral and medial geniculate nuclei. Caine hypothesized that kindling stimulation might significantly alter specific sensory function, a view which our ILS findings support. Further, he also noted high AD thresholds as well as rapid seizure generalization but was able to obtain clinical seizures in only 1 of 10 subjects (medial geniculate placement). Noteworthy differences between Caine’s study and ours include the species investigated and, perhaps more important, the stimulus variables used. Caine used rats as subjects and imposed a 450~PA ceiling on stimulus intensity. Had more flexible stimulus variables been used in that early study, the two experiments might have had equivalent results. A second account for the immediate seizure generalization seen with LGN kindling also concerns the magnitude of stimulation required to elicit AD from the region. Because of the anatomic proximity of the LGN and hippocampus, it is possible that volume conduction from thalamus to hippocampus may have contributed to our finding. Contrary to this view is the important observation of Sato and Nakashima (3 1) that hippocampal kindling required an average of 5 1.8 AD elicitations to establish generalized
THALAMIC
KINDLING
27
FIG. 4. Coronal section through the lateral geniculate nucleus at A 6.5 (from 37). Note the central location of the LGN stimulating electrodes (0) in three of four cats. The electrode site in the remaining cat was situated more ventrally.
tonic-clonic convulsions in cats. Moreover, suprathreshold stimulation of limbic structures did not facilitate the kindling process (29). Finally, in spite of the absence of hippocampal electrode placements in our study, it is well known that hippocampal kindling in rats and cats is a gradual process reflected electrographically by increasing cortical and subcortical (e.g., amygdala) involvement and by corresponding, slowly developing clinical accompaniment (14, 29, 3 1). Similarly, amygdala kindling also develops gradually but requires fewer AD elicitations (e.g., 21 to 28 ADS in cats) to establish generalized tonic-clonic convulsions (14, 29, 3 1, 33, 34, 45). None of these progressive features characterized LGN kindling, which required an average of 7.5 stimulations (6.5 subthreshold and 1 AD threshhold) in
28
SHOUSE
AND
RYAN
our population. A volume conduction hypothesis therefore appears untenable, at least with respect to limbic system structures. Volume conduction to other thalamic or cortical regions is also a potential complicating factor. Primary generalized epilepsy can be recruited by stimulation of thalamic nuclei, as evidenced by ample experimental support dating from the 1940s (6, 9, 16, 17, 28). In this regard, it is important to note that positive results were largely confined to the midline thalamic nuclei and elicited a primary generalized “petit mal” epilepsy. In the present study, only grand ma1 seizures were observed. Finally, efforts to kindle cortical regions involved in photosensitive grand ma1 epilepsy have been successful, for example, in Pupio pupio (44) but, unlike our findings, generalized seizures developed gradually from an initial, well defined, focal onset. Thus, although volume conduction may not be excluded as a factor in LGN kindling, our observations are not consistent with the development and/or types of seizures induced by stimulation of adjacent or distant anatomic sites. In addition to demonstrating seizures after LGN stimulation, a second objective of this study was to compare the characteristics of LGN and amygdala kindling. In that regard, a number of distinctions were noted. For example, the data on LGN kindling are in many ways consistent with a primary generalized grand mal seizure disorder, possibly of a photosensitive nature. On the other hand, LGN kindling may represent a focal epilepsy, with the differences from amygdala kindling attributed to the distinctive functions and interconnections of the two nuclei. These two interpretations are discussed separately. It is conceivable that LGN kindling qualifies as a primary generalized grand ma1 epilepsy based on certain characteristics of its development, symptomatology and associated sleep anomalies (8, 10, 18, 23-28). First, threshold stimulation produced a sudden, bilaterally synchronous generalization of 3 or 4 to 5 c/s spike and wave activity with clinical accompaniment. Although no isolated minor seizures were ever observed, the initial ictal event involved the eyelids bilaterally, and the entire clinical prodrome remarkably resembled the photoconvulsive response in P. papio (18, 2325). Second, the sleep characteristics after LGN kindling also characterize primary generalized human and experimental seizure conditions (11, 12, 18, 41). For example, REM sleep loss is common in P. pupio (18) and the slowing of EEG sleep spindles without interictal discharge also occurs in the early stages of feline generalized penicillin epilepsy; the latter has been interpreted as an intermediate form between normal sleep spindle and spike-and-wave activity by Kostopoulos, Gloor, and colleagues ( 11, 19). These sleep abnormalities are distinguished from those accompanying complex partial seizures, where sleep spindles are reduced in incidence (46)
THALAMIC
KINDLING
29
but are of essentially normal configuration (4 1). Finally, the visual functions of the LGN suggest the photosensitive nature of these seizures, and the selective reduction in seizure threshold with ILS also supports this perspective (1). Collectively, then, the anatomic substrate, the electrographic and clinical features of the seizure disorder, the sleep state alterations and the ILS results all suggest primary grand ma1 epilepsy of a photosensitive nature. On the other hand, rapid secondary seizure generalization is consistent with historic views of the thalamus (38). This structure has always been a logical choice for the propagation of generalized seizure manifestations because of its extensive modulation of both sensory and motor functions. Specific evidence for widespread thalamic involvement in secondary seizure generalization from the amygdala has been provided by Engel and colleagues (5), whose 2-deoxyglucose radiographs showed increased metabolic rates in all specific and nonspecific thalamic nuclei during the transition between focal to generalized seizures. In this context, one could construe the latent period between focal discharge in most limbic structures and the generalization of AD to the thalamus as representing endogenous differences in seizure susceptibility between limbic and thalamic structures. Certainly, a host of studies have documented unusually low thresholds in the limbic systems of human and infrahuman species (14, 29, 46). This conceptualization could account for the immediate onset of generalized seizures when threshold is attained in the thalamus, compared with their gradual appearance after high-frequency [(3) vs. (4)] stimulation of the amygdala. Rapid secondary generalization is also consonant with the lag time in the appearance of AD that occasionally occurred in cortical electrodes compared with deep leads and with the findings from traditional models of primary generalized photosensitive epilepsy, notably P. pupio, where the photoconvulsive response is presumably of frontooccipital cortical origin (23-25). The findings with LGN stimulation extend the flexibility and generality of the kindling model of epilepsy. In contrast to amygdala kindling, which resembles complex partial seizures, LGN kindling appears to reflect a different process, which represents a primary or exceedingly rapid secondary generalized epilepsy. The sleep disturbances observed in these two types of kindled seizures correspond to those seen in human and other animal analogues (18, 41). Although the neurochemical basis of both sleep and epilepsy remain enigmatic, there is some evidence in the kindling literature which could account for both abnormalities. For example, both short- and long-term reductions in catecholamines have been documented after limbic system kindling in rats and cats (4, 42). and diminished acetylcholine receptors have also been observed in structures utilizing that neurotransmitter (22, 29). Neurons in the LGN are thought to use acetylcholine (20).
SHOUSE AND RYAN
30
Although other mechanisms cannot be excluded (22), an imbalance in these important neurohumoral substances could account not only for the two seizure conditions but also for the sleep deficits observed. Further investigation of this new kindling phenomenon could therefore contribute not only to our understanding of epilepsy but also of its important relation to hypnogenic mechanisms. REFERENCES 1. BICKFORD, R. G., AND D. W. KLASS. 1969. Sensory precipitation and reflex mechanisms. Pages 543-564 in A. A. WARD, H. H. JASPER,AND A. POPE, eds., Basic Mechanisms of the Epilepsies. Little, Brown, Boston. 2. CAINE, D. P. 1977. Seizure development following repeated electrical stimulation of central olfactory structures. Ann. N. Y. Acad. Sci. 290: 200-2 16. 3. CAINE, D. P., AND M. E. CORCORAN. 1981. Kindling with low-frequency stimulation: generality, transfer and recruiting effects. Exp. Neural. 73: 2 19-232. 4. ENGEL, J., AND N. S. SHARPLESS.1977. Long lasting depletion of catecholamines in the rat amygdala induced by kindling stimulation. Brain Res. 136: 38 I-386. 5. ENGEL, J., JR., L. WOLFSON, AND L. BROWN. 1978. Anatomical correlates of electrical and behavioral events related to amygdaloid kindling. Ann. Neurol. 3: 538-544. 6. FEENEY, 0. M., AND F. P. GULLOTTA. 1972. Suppression of seizure discharges and sleep spindles by lesions of rostra1 thalamus. Brain Res. 45: 254-259. 7. FERNANDEZ-GUARDIOLA, A., J. L. JURADO, AND J. M. CALVO. 1981. Repetitive lowintensity electrical stimulation of cat’s nonlimbic brain structures: dorsal raphe nucleus kindling, Pages 123-135 in J. WADA, Ed., Kindling 2. Raven Press, New York. 8. GASTAUT, H. 1968. Clinical and electroencephalographic correlates of generalized spike and wave bursts occurring spontaneously in man. Epilepsia 9: 79-84. 9. GASTAUT, H., AND M. FISCHER-WILLIAMS. 1959. The physiopathology of epileptic seizures, pages 329-363 in J. FIELD, Ed., Handbook OfPhysiology. Sect. I, Vol. I. Amer. Physiol. Sot., Washington, D.C. 10. GIBBS, E. L., H. H. MERRITT, AND F. A. GIBBS, 1943. Electroencephalographic foci associated with epilepsy. Arch. Neurol. Psychiafry 49: 793-80 1. I I. GLOOR, P. 1968. Generalized cortico-reticular epilepsies: some considerations in the pathophysiology of generalized bilaterally synchronous spike and wave discharge. Epiiepsia 9: 249-263. 12. GLOOR, P., L. F. QUESNEY, AND H. ZUMSTEIN. 1977. Pathophysiology of generalized epilepsy in the cat: the role of cortical and subcortical structures. II. Topical application of penicillin to the cerebral cortex and to subcortical structures. Electroenceph. Clin. Neurophysiol.
43: 79-94.
13. GODDARD, G. V. 1967. The development of epileptic seizures through brain stimulation at low intensity. Nature (London) 214: 1020. 14. GODDARD, G. V., D. MCINTYRE, AND C. LEACH. 1969. A permanent change in brain function resulting from daily electrical stimulation. Exp. Neurol. 25: 295-330. 15. HALASZ, P. 198 1, Generalized epilepsy with spike-wave paroxysms as an epileptic disorder of the function of sleep promotion. Acta Physiol. Acad. Sci. Hung. 57: 5 l-86. 16. INGVAR, D. H. 1955. Reproduction of the 3 per second spike-wave EEG pattern by subcortical electrical stimulation in cats. Acta Physiol. &and. 33: I-14. 17. JASPER, H. H., AND J. DROOGLEEVER-FORTUYN. 1947. Experimental studies in the
THALAMIC
18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.
KINDLING
31
functional anatomy of petit ma1 epilepsy. Rex Publ. Assoc. Res. Nerv. Ment. Dis. 26: 272-298. KILLAM, K. F., E. K. KILLAM, AND R. NAQUET. 1967. An animal model of light sensitive epilepsy. Eleetroenceph. Clin. Neurophysiol. 22: 497-5 13. KOSTOFQULOS, G., AND P. GLOOR. 1982. A mechanism for spike-wave discharge in feline epilepsy and its relation to spindle generation, Pages 29-42 in M. B. STERMAN, M. N. SHOLJSE,AND P. PASSOUANT, eds., Sleep and Epilepsy. Academic Press, New York. LEWIS, P. R., AND C. C. D. SHUTE. 1967. The cholinergic limbic system: projections to hippocampal formation, medial cortex, nuclei of the ascenting choline+ reticular system and the subfomical organ and supraoptic crest. Brain 90: 52 I-540. MCGINTY, D. J., AND R. R. DRUCKER-COLIN. 1982. Sleep mechanisms: biology and control of REM sleep. Int. Rev. Neurosci. 23: 391-436. MCNAMARA, J. O., M. C. BYRNE, R. M. DASHEIFF, AND J. G. FRITZ. 1980. Kindling model of epilepsy: a review. Prog. Neurobiol. 15: 139- 157. NAQUET, R. 1973. L’epilepsie photosensible du Papio papio un modele de I’epilepsie photosensible de l’homme. Arch. Ital. Biol. 111: 5 16-526. NAQUET, R. 1969. Photogenic seizures in the baboon, pages 565-574 in A. A. WARD, H. H. JASPER,AND A. POPE, Eds., Basic Mechanisms of the Epilepsies. Little, Brown, Boston. NAQUET, R. 1977. Sleep and experimental epilepsy, Pages 37-43 in Sleep 1976, 3rd European Congress of Sleep Research, Montpellier. Karger, Basel. NIEDERMEYER, E. 1972. The Generalized Epilepsies. Thomas, Springfield, IL. NIEDERMEYER, E. 1982. Petit mal, primary generalized epilepsy and sleep. Pages 19 I-207 in M. B. STERMAN, M. N. SHOUSE, AND P. PASSOUANT, Eds.. Sleep and Epilepsy, Academic Press, New York. PENFIELD, W., AND H. H. JASPER. 1954. Epilepsy and the Functional Anatomy of the Brain. Little, Brown, Boston. RACINE, R. 1978. Kindling: the first decade. Neurosurgery 3: 234-252. RONDOUIN, A., M. BALDY-MOULINIER, AND P. PASSOUANT. 1980. The influence of hippocampal kindling on sleep organization in cats. Effects of alpha-methylparatyrosine. Brain
Res. 181: 413-424.
31. SATO, M., AND T. NAKASHIMA. 1975. Kindling: secondary epileptogenesis, sleep and catecholamines. Can. .I. Neurol. Sci. 2, 439-446. 32. SHOUSE,M. N. 1982. Acute effects of pyridoxine hydrochloride on monomethylhydrazine (MMH) seizure latency and amygdaloid kindled seizure thresholds in cats. Exp. Neural 75: 79-88. 33. SHOUSE, M. N., AND M. B. STERMAN. 1981. Sleep and kindling. I. Effects of initial afterdischarge threshold determination. Exp. Neurol. 71: 550-562. 34. SHOUSE,M. N., and M. B. STERMAN, 1981. Sleep and Kindling: II. Effects of generalized seizure induction. Exp. Neurol. 71: 563-580. 35. SHOUSE, M. N., AND M. B. STERMAN, 1982. Acute sleep deprivation reduces amygdalakindled seizure thresholds in cats. Exp. Neurol. 78: 716-727. 36. SHOUSE, M. N., AND M. B. STERMAN. 1983. “Kindling” a sleep disorder. Degree of sleep pathology predicts kindled seizure susceptibility in cats. Brain Res. 271: 186-200. 37. SNIDER, R. S., AND W. T. NIEMER. 1961. Stereotaxic Atlas of the Cat Brain. Univ. of Chicago Press, Chicago. 38. STERIADE, M. 1970. Ascending control of thalamic and cortical responsiveness. Int. Rev. Neurobiol.
12: 87-144.
39. STERMAN, M. B., AND C. D. CLEMENTE. 1962. Forebrain inhibitory mechanisms: sleep
SHOUSE AND RYAN
32
40. 41. 42. 43. 44. 45.
patterns induced by basal forebrain stimulation in the behaving cat. Exp. Neurol. 6: 103-I 17. STERMAN, M. B., and M. N. SHOUSE. 1983. Basal forebrain mechanisms in sleep. Pages 277-300 in D. J. MCGINTY, P. L. PARMEGGIANI, R. DRUCKER-COLIN, AND A. MORRISON, Eds., Brain Mechanisms of Sleep. Raven Press, New York. STERMAN, M. B., M. N. SHOUSE, AND P. PASSOUANT. Eds. 1982. Sleep and Epilepsy. Academic Press, New York. TANAKA, R., H. LANGE, AND R. NAQUET. 1975. Sleep, subcortical stimulation and kindling in the cat. Can. J. Neurol. Sci. 2: 447-455. URSIN, R., AND M. B. STERMAN. 1982. A Manual for Recording and Scoring Sleep Stages in Ihe Cat. Brain Information Service, UCLA, Los Angeles. WADA, J. A., T. OSAWA, AND MIZOGUCHI. 1975. Recurrent spontaneous seizure state induced by prefrontal kindling in Senegalese baboons, Papio papio. Can. J. Neural. Sci. 2: 477-499. WADA, J. A., AND M. SATO. 1974. Generalized convulsive seizure induced by stimulation of the amygdala in cats: correlative electroencephalographic and behavioral features. Neurology
(New
York)
24: 565-574.
46. WALTER, R. D. 1973. Tactical considerations leading to surgical treatment of limbic epilepsy. Pages 99-l 18 in M. A. B. BRAZER ed., Epilepsy; 11s Phenomena in Man. Academic Press, New York/London.
NEUROLOGY
86, 18-32 (1984)
Thalamic Kindling: Electrical Stimulation of the Lateral Geniculate Nucleus Produces Photosensitive Grand Mal Seizures M. N. SHOUSE AND W. RYAN’ Research
Service, Veterans Administration Medical Center, Sepulveda, California 91343 and School of Medicine, University of California. Los Angeles, Calijornia 90024 Received
October
5, 1983: revision
received
May
7. 1984
Kindling is traditionally viewed as a chronic, focal epilepsy model which consistently induces complex partial seizures from limbic structures in animals. This study revealed that primary or exceedingly rapid secondary generalized seizures could also be kindled when stimulation was applied to the lateral geniculate nucleus, a thalamic region involved in sleep regulation and possibly also photosensitive epilepsy. Two experiments were conducted in cats. Experiment I compared the development of generalized tonic-clonic convulsions and associated sleep disorders following electrical stimulation of the lateral geniculate nucleus (N = 4) and the amygdala (N = 4). Experiment 2 described the effects of intermittent light stimulation on seizure thresholds in both groups. Three primary findings distinguished the epileptogenic process in those two brain regions. First, generalized electroencephalographic and clinical seizures accompanied the first afterdischarge obtained with thalamic stimulation. In contrast, focal seizures with secondary generalization appeared during a 3to 4-week period of afterdischarge elicitations from the amygdala. Second, amygdalakindled cats showed fewer sleep spindles during slow-wave sleep whereas cats kindled in the lateral geniculate nucleus had abnormal sleep spindles approaching spike wave-like activity. Third, only the latter cats showed reduced seizure thresholds in response to photic stimulation. Based on the anatomic substrates involved, the clinical and electrographic profiles observed during kindling and the type of sleep disturbance shown, we concluded that lateral geniculate nucleus kindling may represent primary generalized epilepsy, possibly of a photosensitive nature; alternatively, the rapid propagation of abnormal discharge was also consistent with the important role of the thalamus in secondary seizure generalization.
Abbreviations: LGN-lateral geniculate nucleus, ILS-intermittent light stimulation, SWSslow-wave sleep, REM-rapid eye-movement. ’ This research was supported by the Veterans Administration, Sepulveda, California.
18 0014-4886/84 $3.00 Copyright 63 1984 by Academic Press, Inc. All rights of reproduction in any form reserved
THALAMIC
19
KINDLING
INTRODUCTION Since its discovery in 1967 by Goddard ( I3), kindling has been construed primarily as a chronic, focal model of epilepsy resembling complex partial seizures in humans (22, 29). This view is supported by the fact that repetitive, electrical or chemical stimulation of virtually all forebrain sites, especially in the limbic system, initially produces focal afterdischarge (AD) and seizures, both of which gradually become secondarily generalized (13, 14, 22, 29). In contrast, the thalamus and the brain stem have been resistant to kindling (7, 13, 14, 29); historically, these structures have been more intimately involved in primary generalized epilepsies (6, 9-l 2, 15- 17, 2628), especially of the “petit mal” variety. A possible explanation for the recalcitrance of brain stem and thalamic nuclei to kindling may reflect different etiologies and anatomic substrates for primary and secondary generalized seizure disorders. A related observation concerns the sleep modulation functions of many brain stem and nonspecific thalamic structures vis-a-vis the stimulus variables used in the kindling procedure. Whereas sleep disruption accompanies limbic system kindling (30, 33, 34, 36, 42) the application of AD-eliciting stimulation to the brain stem, for example, produces normal sleep instead of clinical seizure manifestations (7). We have previously suggested that normal sleep (39) and seizure induction may be incompatible processes (34-36, 4 1). To test these hypotheses, we attempted kindling in the thalamus. The lateral geniculate nucleus (LGN), a specific sensory relay nucleus, was chosen as a target site rather than a nonspecific nucleus in order to avoid the complications cited above. Accordingly, this report documents clinical and electrographic seizure manifestations as well as the sleep anomalies associated with LGN kindling compared with amygdala kindling in cats. METHODOLOGY
AND
RESULTS
Two consecutive experiments were conducted in the same animals. Experiment 1 characterized the development of a seizure condition using elecrical kindling procedures applied to either the LGN or amygdala. In addition, the effects of these seizures on sleep and waking state patterns were also documented. Experiment 2 was undertaken because of the visual functions of the LGN and examined the effects of intermittent light stimulation (ILS) on seizure thresholds in LGN- and amygdala-kindled animals. The procedures and results of these two experiments are described separately except for histologic findings, which are described at the end of this section.
20
SHOUSE
AND
Experiment
RYAN
1
Stereotaxic surgery was carried out on 12 adult cats (1.8 to 4.2 kg) under sodium pentobarbital anesthesia for electrode placements (37) used in kindling and state evaluation. The details of electrode construction and use in kindling and sleep-waking state evaluation were described elsewhere (32-36, 43). The montage included: bilateral, tripolar leads in the LGN (A 6.0; L 10.5; H 2, 2.5, 3.0) and basolateral amygdala (A 11.0, L 9.5, H -5.5, -6.0, -5.5); jeweler’s screws threaded into the bone over sensorimotor cortex (A 23.0; L 8 and 10) bilaterally for surface EEGs and into the frontal sinus for an EOG; and silver wire electrodes inserted into the neck musculature for an EMG. After a 2-week postoperative recovery, all cats underwent two, 12-h polygraphic recordings to establish baseline state measurements. Afterward, the cats were divided into three groups (N = 4 each) according to the type of brain stimulation received: group lamygdala kindling; group 2-LGN kindling; group 3-no brain stimulation (control). At the beginning of kindling, initial AD thresholds were established using a standard method of limits procedure (32-36, 45). A 100~PA stimulus (ls train of 60 Hz, biphasic square waves of I-ms pulse duration) was applied on day 1 and 200 PA on day 2. Subsequently, amperage was increased by 200 FA each day until the AD appeared and then decreased by 100 PA each day until the AD disappeared. Threshold was therefore defined as the minimum intensity eliciting an AD within 100 PA. Afterward, a single daily stimulation was applied at AD threshold until generalized tonicclonic convulsions (stage 6 seizures) were obtained. Stage 6 seizure thresholds were then established using the same method of limits procedure as at the beginning of kindling. Finally, additional 12-h polygraphic recordings were obtained during and at the end of kindling, in relation to initial and final seizure threshold determinations. Statistical analysis consisted of two-way repeated measures analyses of variance (group X time) applied to kindling and sleep state measurements. Dependent variables for kindling consisted of (a) the number of subthreshold vs. threshold stimulations required to obtain stage 6 seizures and (b) initial AD thresholds vs. stage 6 seizure thresholds (PA). Sleep state variables included (a) the mean percentage time spent in slow-wave sleep (SWS) and rapid-eye-movement (REM) sleep and (b) the number of sleep spindles during comparable, 5-min SWS samples derived from the first segment of the second sleep cycle. Post hoc individual comparisons were evaluated with Student’s t tests. Kindling the Lateral Geniculate Nucleus vs. the Amygdala. Table 1 shows
THALAMIC
KINDLING
21
the findings for the kindling variables. More subthreshold stimulations (P < 0.1) were required to elicit initial ADS from the LGN than from the amygdala (Table 1A). Amygdala-stimulated cats initially displayed considerable variability in initial AD (33) as did LGN-stimulated animals. Thus, in spite of an apparently sizeable difference in the number of subthreshold stimulations to initial AD between the two groups, excessive variability in values, as reflected by high standard deviations (Table lA), precluded a statistically significant effect, and only a statistical trend was obtained. However, when AD appeared in the LGN cats, significantly fewer threshold stimulations were needed to obtain generalized seizures (P < 0.001). Compared with the 3- to 4-week interval required to elicit stage 6 convulsions from the amygdala, cats receiving LGN stimulation showed stage 6 seizures on the occasion of their first AD. Figure 1 illustrates the type of EEG generalization obtained after stimulating the two regions. Even at the end of amygdala kindling, predominantly focal AD occurred first in the amygdala ipsilateral to the kindling electrode and was associated with unilateral facial twitching; AD then spread to other regions before culminating in a stage 6 seizure. In contrast, the first threshold stimulation in LGN cats resulted in generalized AD, which approximated bilateral synchrony. Although the amplitude of the ADS was largest in the LGN, AD appeared to occur simultaneously in the LGN, amygdala and over the sensorimotor cortex in 3 of 4 cats; in the remaining animal, cortical AD developed 2 to 3 s after it occurred in deep leads. Typically, these ADS consisted of 4 to 5 c/s spikeand-wave although 3 c/s activity was also observed. The initial seizure manifestation was bilateral eye blinking, which was invariably followed by a generalized tonic-clonic convulsion. Vocalization usually occurred at the end of the seizure, and autonomic symptoms such as salivation and urine incontinence frequently occurred during the seizure. Initial AD thresholds were higher in LGN animals than in the amygdalakindled group (P < 0.1; Table 1B) but were not outside the range of initial AD thresholds obtained from these and larger populations of amygdalakindled cats (32-36, 45). For this reason, only a statistical trend was shown by our data. It is remarkable, however, that no reduction in AD or generalized seizure threshold was seen after LGN kindling, as was customarily observed when initial AD and final stage 6 seizure thresholds were compared in amygdala-kindled cats [P < 0.00 1; Table 1B; ( 14, 2 l-36, 45)]. Stage 6 seizure thresholds were therefore significantly higher in the LGN group compared with values for the amygdala cats (P < 0.001). Further, stable LGN (and amygdala) seizure thresholds were maintained at the same level as before during three additional seizure threshold determinations conducted immediately afterward.
SHOUSE AND RYAN
22
TABLE I Kindling Values for Two Groups of Cats Receiving Stimulation in the Lateral Geniculate Nucleus (LGN) or Amygdala” A. Number of stimulations to kindling Experimental groups
Subthreshold stimuli
Threshold (AD) stimuli
B. Seizure thresholds (PA) Initial AD threshold
Stage 6 seizure threshold
Amygdala kindling
(N= 4)
3.25f 2
24.2k 3***
600+100
275+ 70***
LGN kindling
(N= 4)
6.5 + 3**
1.0 f o*s***
1275f 618**
1275+ 618***
’ A-the means f SD number of subthreshold and threshold [after discharge (AD)-eliciting] stimuli required to obtain stage 6 or generalized tonic-clonic convulsions. B-the means f SD stimulus threshold-inducing AD at the beginning of kindling and stage 6 seizures at the end of kindling. Significant difference between groups: *P < 0.001; **P < 0.1. Significant difference within cats in the same group: ***P c 0.001.
Kindling E&cts on Sleep States. An analysis of sleep state parameters (Table 2) revealed a progressive and significant reduction in SWS and REM sleep time during amygdala kindling. This was accompanied by a reduction in the density of sleep spindles (number of sleep spindles per unit time, visually counted by two independent observers), which were of normal configuration before and after kindling. These postkindling changes confirm previous work (32-36, 40) and showed sleep spindle reduction only during the generalized phases of amygdala kindling. Postkindling reductions in sleep spindles were statistically significant compared with initial baselines in amygdala-kindled cats as well as with values derived from control animals. In contrast, cats receiving LGN stimulation showed a decrement only in REM sleep. It was difficult to compare sleep spindle incidence in the baseline and postkindling records in LGN-stimulated cats. In postkindling records sleep spindles were abnormal in appearance; the 12 to 15 c/s spindle activity of the prekindling records was intermixed with slower activity (8 to 10 c/s) in postkindling records. This altered rhythmic activity may reflect an intermediate form between sleep spindles and a spike wave-pattern even though no interictal discharge was apparent in the postkindling sleep records (Fig. 2). Experiment 2 Experiment 2 examined the effects of photic stimulation on seizure thresholds in both groups of kindled animals and began 10 days after the
THALAMIC
KINDLING
A. AMYGDALA
STIMULATION
6. LGN
23 %cL”
STIMULATION
LT AMYG LT LGN LT LGN FIT LGN LT SM CX RT SM CX RT LGN
t STIMULUS
25
FIG. 1. Representative electroencephalographic responses to electrical stimulation (l-s, 60Hz, biphasic square waves) at intensities (PA) just sufficient to elicit afterdischarge (AD) and generalized tonic-clonic convulsions (stage 6 seizures) from the left amygdala (A) and the left lateral geniculate nucleus (B). Amygdala (AMYG) stimulation evoked AD first in the amygdala ipsilateral to the kindling electrode followed by secondary generalization to other brain areas. Note some initial AD also over the left sensorimotor cortex (LT SM CX). In contrast, stimulation of the lateral geniculate nucleus (LGN) produced generalized AD accompanied by a stage 6 seizure.
last seizure threshold determination was made at the end of Experiment 1. In this experiment, stage 6 seizure thresholds were reestablished using the same procedures as in Experiment 1 except that all stimuli were applied on the same day, with a 5-min interstimulus interval. This modified procedure was used in previous studies of seizure threshold in kindled cats (32, 35). A within-subjects paradigm was used in which seizure thresholds were established on three consecutive occasions, with a lo-day intertrial interval. The first experimental condition (A,) provided an initial baseline seizure threshold without photic stimulation. During the second condition, (B,),
SHOUSE AND RYAN
24
TABLE 2 Means + SD Percentage Time Spent in Slow-Wave Sleep (SWS) and Rapid Eye-Movement (REM) Sleep before and at the End ofAmygdala and Lateral Geniculate Nucleus (LGN) Kindling sws Experimental groups Amygdala kindling (N= 4) LGN kindling (N= 4) Control (N=
4)
REM sleep
Before kindling
After kindling
Before kindling
After kindling
33 f 8
24.0 2 3*,**
15 *2
10 * 4****
31*7
30
*8
16 +4
10 + 5****
30 + 6
35
+2
14+
I
16 -t 2
* P < 0.05 compared with control values. ** P < 0.05 compared with initial baseline values.
photic stimulation (1.5 Hz) was administered through the window of the recording chamber 2 min prior to and throughout the threshold testing procedure. The last condition (AZ) provided a final baseline seizure threshold without photic stimulation. The results of Experiment 2 are summarized in Fig. 3, which shows stage 6 seizure thresholds in the two groups. There were two primary findings. First, LGN-stimulated cats continued to show higher seizure thresholds than amygdala-kindled cats regardless of the experimental condition. Further,
A. BEFORE
B. AFTER
LGN
LGN
KINDLING
KINDLING
FIG. 2. Representative samples of slow-wave sleep before (A) and after (B) lateral geniculate nucleus (LGN) kindling. It is more difficult to identify the sleep spindles in the postkindling record because their frequency is unusually slow. This abnormal configuration may represent an intermediate form between sleep spindles and spike wave-like discharge.
THALAMIC
NO
PHOTIC
PHOTIC
25
KINDLING
PHOTIC
STIMULATION
NO PHOTlC
CONDITIONS
FIG. 3. Mean seizure threshold (PA) and standard deviations obtained from cats kindled through leads in the lateral geniculate nucleus (LGN) and basolateral amygdala (N = 4 each). Threshold determinations were made on three consecutive occasions with a IO-day interval elapsing between each test. The A, and A2 conditions provided baseline values obtained without photic stimulation. During the intervening B, condition, the animals were treated identically except that intermittent light stimulation (ILS at 1.5 Hz) was presented 2 min prior to and throughout the method of limits threshold procedure. The two primary findings were (i) that seizure thresholds were consistently higher in LGN cats compared with amygdala-kindled animals and (P < 0.001) and (ii) that ILS significantly reduced seizure thresholds only in LGNkindled cats when B, values were compared with initial (A,) and (A*) baselines (P < 0.001).
baseline LGN seizure thresholds (A, and Az) were not significantly different, substantiating the rather unique stability of the high kindled seizure thresholds in these animals. Second, ILS markedly and selectively reduced seizure thresholds in LGN cats, comparing B, values with initial (A,) and final (A*) baselines. This 50% reduction in LGN seizure thresholds was
26
SHOUSE AND RYAN
comparable in the four animals tested and occurred regardless of whether the animal looked directly at the flashing light or had its eyes open or closed. Histology
The accuracy of the LGN placements was verified by selective spiking in that nucleus during REM sleep and during ILS; LGN and amygdala placements were also confirmed by subsequent histological examination. Figure 4 illustrates electrode sites in the LGN. Placements were centrally situated in the nucleus in three cats. Stimulating electrodes were situated somewhat more ventrally in the remaining cat, which showed a delay between subcortical and cortical AD onset. DISCUSSION The primary finding was the discovery of generalized seizures systematically induced by kindling the LGN, a specific thalamic relay nucleus. In addition, REM sleep loss and atypical sleep spindles were observed after LGN kindling, an outcome which is consistent with the hypothesis that sleep and seizure disturbances are related phenomena (32, 34-36). Finally, the sensitivity to ILS shown after LGN kindling is compatible with the visual functions of this nucleus. These observations with cats extend those of Caine (2), who first attempted kindling of specific thalamic nuclei, including the lateral and medial geniculate nuclei. Caine hypothesized that kindling stimulation might significantly alter specific sensory function, a view which our ILS findings support. Further, he also noted high AD thresholds as well as rapid seizure generalization but was able to obtain clinical seizures in only 1 of 10 subjects (medial geniculate placement). Noteworthy differences between Caine’s study and ours include the species investigated and, perhaps more important, the stimulus variables used. Caine used rats as subjects and imposed a 450~PA ceiling on stimulus intensity. Had more flexible stimulus variables been used in that early study, the two experiments might have had equivalent results. A second account for the immediate seizure generalization seen with LGN kindling also concerns the magnitude of stimulation required to elicit AD from the region. Because of the anatomic proximity of the LGN and hippocampus, it is possible that volume conduction from thalamus to hippocampus may have contributed to our finding. Contrary to this view is the important observation of Sato and Nakashima (3 1) that hippocampal kindling required an average of 5 1.8 AD elicitations to establish generalized
THALAMIC
KINDLING
27
FIG. 4. Coronal section through the lateral geniculate nucleus at A 6.5 (from 37). Note the central location of the LGN stimulating electrodes (0) in three of four cats. The electrode site in the remaining cat was situated more ventrally.
tonic-clonic convulsions in cats. Moreover, suprathreshold stimulation of limbic structures did not facilitate the kindling process (29). Finally, in spite of the absence of hippocampal electrode placements in our study, it is well known that hippocampal kindling in rats and cats is a gradual process reflected electrographically by increasing cortical and subcortical (e.g., amygdala) involvement and by corresponding, slowly developing clinical accompaniment (14, 29, 3 1). Similarly, amygdala kindling also develops gradually but requires fewer AD elicitations (e.g., 21 to 28 ADS in cats) to establish generalized tonic-clonic convulsions (14, 29, 3 1, 33, 34, 45). None of these progressive features characterized LGN kindling, which required an average of 7.5 stimulations (6.5 subthreshold and 1 AD threshhold) in
28
SHOUSE
AND
RYAN
our population. A volume conduction hypothesis therefore appears untenable, at least with respect to limbic system structures. Volume conduction to other thalamic or cortical regions is also a potential complicating factor. Primary generalized epilepsy can be recruited by stimulation of thalamic nuclei, as evidenced by ample experimental support dating from the 1940s (6, 9, 16, 17, 28). In this regard, it is important to note that positive results were largely confined to the midline thalamic nuclei and elicited a primary generalized “petit mal” epilepsy. In the present study, only grand ma1 seizures were observed. Finally, efforts to kindle cortical regions involved in photosensitive grand ma1 epilepsy have been successful, for example, in Pupio pupio (44) but, unlike our findings, generalized seizures developed gradually from an initial, well defined, focal onset. Thus, although volume conduction may not be excluded as a factor in LGN kindling, our observations are not consistent with the development and/or types of seizures induced by stimulation of adjacent or distant anatomic sites. In addition to demonstrating seizures after LGN stimulation, a second objective of this study was to compare the characteristics of LGN and amygdala kindling. In that regard, a number of distinctions were noted. For example, the data on LGN kindling are in many ways consistent with a primary generalized grand mal seizure disorder, possibly of a photosensitive nature. On the other hand, LGN kindling may represent a focal epilepsy, with the differences from amygdala kindling attributed to the distinctive functions and interconnections of the two nuclei. These two interpretations are discussed separately. It is conceivable that LGN kindling qualifies as a primary generalized grand ma1 epilepsy based on certain characteristics of its development, symptomatology and associated sleep anomalies (8, 10, 18, 23-28). First, threshold stimulation produced a sudden, bilaterally synchronous generalization of 3 or 4 to 5 c/s spike and wave activity with clinical accompaniment. Although no isolated minor seizures were ever observed, the initial ictal event involved the eyelids bilaterally, and the entire clinical prodrome remarkably resembled the photoconvulsive response in P. papio (18, 2325). Second, the sleep characteristics after LGN kindling also characterize primary generalized human and experimental seizure conditions (11, 12, 18, 41). For example, REM sleep loss is common in P. pupio (18) and the slowing of EEG sleep spindles without interictal discharge also occurs in the early stages of feline generalized penicillin epilepsy; the latter has been interpreted as an intermediate form between normal sleep spindle and spike-and-wave activity by Kostopoulos, Gloor, and colleagues ( 11, 19). These sleep abnormalities are distinguished from those accompanying complex partial seizures, where sleep spindles are reduced in incidence (46)
THALAMIC
KINDLING
29
but are of essentially normal configuration (4 1). Finally, the visual functions of the LGN suggest the photosensitive nature of these seizures, and the selective reduction in seizure threshold with ILS also supports this perspective (1). Collectively, then, the anatomic substrate, the electrographic and clinical features of the seizure disorder, the sleep state alterations and the ILS results all suggest primary grand ma1 epilepsy of a photosensitive nature. On the other hand, rapid secondary seizure generalization is consistent with historic views of the thalamus (38). This structure has always been a logical choice for the propagation of generalized seizure manifestations because of its extensive modulation of both sensory and motor functions. Specific evidence for widespread thalamic involvement in secondary seizure generalization from the amygdala has been provided by Engel and colleagues (5), whose 2-deoxyglucose radiographs showed increased metabolic rates in all specific and nonspecific thalamic nuclei during the transition between focal to generalized seizures. In this context, one could construe the latent period between focal discharge in most limbic structures and the generalization of AD to the thalamus as representing endogenous differences in seizure susceptibility between limbic and thalamic structures. Certainly, a host of studies have documented unusually low thresholds in the limbic systems of human and infrahuman species (14, 29, 46). This conceptualization could account for the immediate onset of generalized seizures when threshold is attained in the thalamus, compared with their gradual appearance after high-frequency [(3) vs. (4)] stimulation of the amygdala. Rapid secondary generalization is also consonant with the lag time in the appearance of AD that occasionally occurred in cortical electrodes compared with deep leads and with the findings from traditional models of primary generalized photosensitive epilepsy, notably P. pupio, where the photoconvulsive response is presumably of frontooccipital cortical origin (23-25). The findings with LGN stimulation extend the flexibility and generality of the kindling model of epilepsy. In contrast to amygdala kindling, which resembles complex partial seizures, LGN kindling appears to reflect a different process, which represents a primary or exceedingly rapid secondary generalized epilepsy. The sleep disturbances observed in these two types of kindled seizures correspond to those seen in human and other animal analogues (18, 41). Although the neurochemical basis of both sleep and epilepsy remain enigmatic, there is some evidence in the kindling literature which could account for both abnormalities. For example, both short- and long-term reductions in catecholamines have been documented after limbic system kindling in rats and cats (4, 42). and diminished acetylcholine receptors have also been observed in structures utilizing that neurotransmitter (22, 29). Neurons in the LGN are thought to use acetylcholine (20).
SHOUSE AND RYAN
30
Although other mechanisms cannot be excluded (22), an imbalance in these important neurohumoral substances could account not only for the two seizure conditions but also for the sleep deficits observed. Further investigation of this new kindling phenomenon could therefore contribute not only to our understanding of epilepsy but also of its important relation to hypnogenic mechanisms. REFERENCES 1. BICKFORD, R. G., AND D. W. KLASS. 1969. Sensory precipitation and reflex mechanisms. Pages 543-564 in A. A. WARD, H. H. JASPER,AND A. POPE, eds., Basic Mechanisms of the Epilepsies. Little, Brown, Boston. 2. CAINE, D. P. 1977. Seizure development following repeated electrical stimulation of central olfactory structures. Ann. N. Y. Acad. Sci. 290: 200-2 16. 3. CAINE, D. P., AND M. E. CORCORAN. 1981. Kindling with low-frequency stimulation: generality, transfer and recruiting effects. Exp. Neural. 73: 2 19-232. 4. ENGEL, J., AND N. S. SHARPLESS.1977. Long lasting depletion of catecholamines in the rat amygdala induced by kindling stimulation. Brain Res. 136: 38 I-386. 5. ENGEL, J., JR., L. WOLFSON, AND L. BROWN. 1978. Anatomical correlates of electrical and behavioral events related to amygdaloid kindling. Ann. Neurol. 3: 538-544. 6. FEENEY, 0. M., AND F. P. GULLOTTA. 1972. Suppression of seizure discharges and sleep spindles by lesions of rostra1 thalamus. Brain Res. 45: 254-259. 7. FERNANDEZ-GUARDIOLA, A., J. L. JURADO, AND J. M. CALVO. 1981. Repetitive lowintensity electrical stimulation of cat’s nonlimbic brain structures: dorsal raphe nucleus kindling, Pages 123-135 in J. WADA, Ed., Kindling 2. Raven Press, New York. 8. GASTAUT, H. 1968. Clinical and electroencephalographic correlates of generalized spike and wave bursts occurring spontaneously in man. Epilepsia 9: 79-84. 9. GASTAUT, H., AND M. FISCHER-WILLIAMS. 1959. The physiopathology of epileptic seizures, pages 329-363 in J. FIELD, Ed., Handbook OfPhysiology. Sect. I, Vol. I. Amer. Physiol. Sot., Washington, D.C. 10. GIBBS, E. L., H. H. MERRITT, AND F. A. GIBBS, 1943. Electroencephalographic foci associated with epilepsy. Arch. Neurol. Psychiafry 49: 793-80 1. I I. GLOOR, P. 1968. Generalized cortico-reticular epilepsies: some considerations in the pathophysiology of generalized bilaterally synchronous spike and wave discharge. Epiiepsia 9: 249-263. 12. GLOOR, P., L. F. QUESNEY, AND H. ZUMSTEIN. 1977. Pathophysiology of generalized epilepsy in the cat: the role of cortical and subcortical structures. II. Topical application of penicillin to the cerebral cortex and to subcortical structures. Electroenceph. Clin. Neurophysiol.
43: 79-94.
13. GODDARD, G. V. 1967. The development of epileptic seizures through brain stimulation at low intensity. Nature (London) 214: 1020. 14. GODDARD, G. V., D. MCINTYRE, AND C. LEACH. 1969. A permanent change in brain function resulting from daily electrical stimulation. Exp. Neurol. 25: 295-330. 15. HALASZ, P. 198 1, Generalized epilepsy with spike-wave paroxysms as an epileptic disorder of the function of sleep promotion. Acta Physiol. Acad. Sci. Hung. 57: 5 l-86. 16. INGVAR, D. H. 1955. Reproduction of the 3 per second spike-wave EEG pattern by subcortical electrical stimulation in cats. Acta Physiol. &and. 33: I-14. 17. JASPER, H. H., AND J. DROOGLEEVER-FORTUYN. 1947. Experimental studies in the
THALAMIC
18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.
KINDLING
31
functional anatomy of petit ma1 epilepsy. Rex Publ. Assoc. Res. Nerv. Ment. Dis. 26: 272-298. KILLAM, K. F., E. K. KILLAM, AND R. NAQUET. 1967. An animal model of light sensitive epilepsy. Eleetroenceph. Clin. Neurophysiol. 22: 497-5 13. KOSTOFQULOS, G., AND P. GLOOR. 1982. A mechanism for spike-wave discharge in feline epilepsy and its relation to spindle generation, Pages 29-42 in M. B. STERMAN, M. N. SHOLJSE,AND P. PASSOUANT, eds., Sleep and Epilepsy. Academic Press, New York. LEWIS, P. R., AND C. C. D. SHUTE. 1967. The cholinergic limbic system: projections to hippocampal formation, medial cortex, nuclei of the ascenting choline+ reticular system and the subfomical organ and supraoptic crest. Brain 90: 52 I-540. MCGINTY, D. J., AND R. R. DRUCKER-COLIN. 1982. Sleep mechanisms: biology and control of REM sleep. Int. Rev. Neurosci. 23: 391-436. MCNAMARA, J. O., M. C. BYRNE, R. M. DASHEIFF, AND J. G. FRITZ. 1980. Kindling model of epilepsy: a review. Prog. Neurobiol. 15: 139- 157. NAQUET, R. 1973. L’epilepsie photosensible du Papio papio un modele de I’epilepsie photosensible de l’homme. Arch. Ital. Biol. 111: 5 16-526. NAQUET, R. 1969. Photogenic seizures in the baboon, pages 565-574 in A. A. WARD, H. H. JASPER,AND A. POPE, Eds., Basic Mechanisms of the Epilepsies. Little, Brown, Boston. NAQUET, R. 1977. Sleep and experimental epilepsy, Pages 37-43 in Sleep 1976, 3rd European Congress of Sleep Research, Montpellier. Karger, Basel. NIEDERMEYER, E. 1972. The Generalized Epilepsies. Thomas, Springfield, IL. NIEDERMEYER, E. 1982. Petit mal, primary generalized epilepsy and sleep. Pages 19 I-207 in M. B. STERMAN, M. N. SHOUSE, AND P. PASSOUANT, Eds.. Sleep and Epilepsy, Academic Press, New York. PENFIELD, W., AND H. H. JASPER. 1954. Epilepsy and the Functional Anatomy of the Brain. Little, Brown, Boston. RACINE, R. 1978. Kindling: the first decade. Neurosurgery 3: 234-252. RONDOUIN, A., M. BALDY-MOULINIER, AND P. PASSOUANT. 1980. The influence of hippocampal kindling on sleep organization in cats. Effects of alpha-methylparatyrosine. Brain
Res. 181: 413-424.
31. SATO, M., AND T. NAKASHIMA. 1975. Kindling: secondary epileptogenesis, sleep and catecholamines. Can. .I. Neurol. Sci. 2, 439-446. 32. SHOUSE,M. N. 1982. Acute effects of pyridoxine hydrochloride on monomethylhydrazine (MMH) seizure latency and amygdaloid kindled seizure thresholds in cats. Exp. Neural 75: 79-88. 33. SHOUSE, M. N., AND M. B. STERMAN. 1981. Sleep and kindling. I. Effects of initial afterdischarge threshold determination. Exp. Neurol. 71: 550-562. 34. SHOUSE,M. N., and M. B. STERMAN, 1981. Sleep and Kindling: II. Effects of generalized seizure induction. Exp. Neurol. 71: 563-580. 35. SHOUSE, M. N., AND M. B. STERMAN, 1982. Acute sleep deprivation reduces amygdalakindled seizure thresholds in cats. Exp. Neurol. 78: 716-727. 36. SHOUSE, M. N., AND M. B. STERMAN. 1983. “Kindling” a sleep disorder. Degree of sleep pathology predicts kindled seizure susceptibility in cats. Brain Res. 271: 186-200. 37. SNIDER, R. S., AND W. T. NIEMER. 1961. Stereotaxic Atlas of the Cat Brain. Univ. of Chicago Press, Chicago. 38. STERIADE, M. 1970. Ascending control of thalamic and cortical responsiveness. Int. Rev. Neurobiol.
12: 87-144.
39. STERMAN, M. B., AND C. D. CLEMENTE. 1962. Forebrain inhibitory mechanisms: sleep
SHOUSE AND RYAN
32
40. 41. 42. 43. 44. 45.
patterns induced by basal forebrain stimulation in the behaving cat. Exp. Neurol. 6: 103-I 17. STERMAN, M. B., and M. N. SHOUSE. 1983. Basal forebrain mechanisms in sleep. Pages 277-300 in D. J. MCGINTY, P. L. PARMEGGIANI, R. DRUCKER-COLIN, AND A. MORRISON, Eds., Brain Mechanisms of Sleep. Raven Press, New York. STERMAN, M. B., M. N. SHOUSE, AND P. PASSOUANT. Eds. 1982. Sleep and Epilepsy. Academic Press, New York. TANAKA, R., H. LANGE, AND R. NAQUET. 1975. Sleep, subcortical stimulation and kindling in the cat. Can. J. Neurol. Sci. 2: 447-455. URSIN, R., AND M. B. STERMAN. 1982. A Manual for Recording and Scoring Sleep Stages in Ihe Cat. Brain Information Service, UCLA, Los Angeles. WADA, J. A., T. OSAWA, AND MIZOGUCHI. 1975. Recurrent spontaneous seizure state induced by prefrontal kindling in Senegalese baboons, Papio papio. Can. J. Neural. Sci. 2: 477-499. WADA, J. A., AND M. SATO. 1974. Generalized convulsive seizure induced by stimulation of the amygdala in cats: correlative electroencephalographic and behavioral features. Neurology
(New
York)
24: 565-574.
46. WALTER, R. D. 1973. Tactical considerations leading to surgical treatment of limbic epilepsy. Pages 99-l 18 in M. A. B. BRAZER ed., Epilepsy; 11s Phenomena in Man. Academic Press, New York/London.