Amygdala damage in experimental and human temporal lobe epilepsy

Amygdala damage in experimental and human temporal lobe epilepsy

Epilepsy Research 32 (1998) 233 – 253 Review Amygdala damage in experimental and human temporal lobe epilepsy Asla Pitka¨nen a,*, Jarkko Tuunanen a,...

5MB Sizes 0 Downloads 18 Views

Epilepsy Research 32 (1998) 233 – 253

Review

Amygdala damage in experimental and human temporal lobe epilepsy Asla Pitka¨nen a,*, Jarkko Tuunanen a, Reetta Ka¨lvia¨inen b, Kaarina Partanen c, Tuuli Salmenpera¨ b a

A.I.Virtanen Institute, Uni6ersity of Kuopio, P.O. Box 1627, FIN-70 211 Kuopio, Finland Department of Neurology, Kuopio Uni6ersity Hospital, P.O. Box 1777, FIN-70 211 Kuopio, Finland c Department of Clinical Radiology, MRI Unit, Kuopio Uni6ersity Hospital, P.O. Box 1777, FIN-70 211 Kuopio, Finland b

Abstract The amygdala complex is one component of the temporal lobe that may be damaged unilaterally or bilaterally in children and adults with temporal lobe epilepsy (TLE) or following status epilepticus. Most MR (magnetic resonance) imaging studies of epileptic patients have shown that volume reduction of the amygdala ranges from 10 – 30%. In the human amygdala, neuronal loss and gliosis have been reported in the lateral and basal nuclei. Studies in rats have more specifically identified the amygdaloid regions that are sensitive to status epilepticus-induced neuronal damage. These areas include the medial division of the lateral nucleus, the parvicellular division of the basal nucleus, the accessory basal nucleus, the posterior cortical nucleus, and portions of the anterior cortical and medial nuclei. Otherwise, other amygdala nuclei, such as the magnocellular and intermediate divisions of the basal nucleus and the central nucleus, remain relatively well preserved. Amygdala kindling studies in rats have shown that the density of a subpopulation of GABAergic inhibitory neurons that also contain somatostatin may be reduced even after a low number of generalized seizures. While analyses of histological sections and MR images indicate that in approximately 10% of TLE patients, seizure-induced damage is isolated to the amygdala, more often amygdala damage is combined with damage to the hippocampus and/or other brain areas. Moreover, recent data from rodents and nonhuman primates suggest that structural and functional alterations caused by seizure activity originating in the amygdala are not limited to the amygdala itself, but may also affect other temporal lobe structures. The information gathered so far on damage to the amygdala in epilepsy or after status epilepticus suggests that local alterations in inhibitory circuitries may contribute to a lowered seizure threshold and greater excitability within the amygdala. Furthermore, damage to select nuclei in the amygdala may predict impairment of performance in behavioral tasks that depend on the integrity of the amygdaloid circuits. © 1998 Elsevier Science B.V. All rights reserved. Keywords: Amygdaloid complex; Magnetic resonance imaging; Pathology; Primate; Rat; Seizure; T2 relaxometry; Volumetry

* Corresponding author. Tel.: +358 17 163296; fax: +358 17 163025; e-mail: [email protected] 0920-1211/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S0920-1211(98)00055-2

234

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

1. Introduction The temporal lobe is composed of the amygdala, the hippocampus, and the surrounding cortex. These areas are interconnected by a myriad of topographically organized pathways which orchestrate the various functions assigned to the temporal lobe, including the formation of memories and emotions (Amaral, 1987). Damage to the hippocampus and its association with the symptomatology of temporal lobe epilepsy (TLE) have been appreciated since 1825 (Bouchet and Cazauvieilh, 1825). Since then, an increasing volume of data has been collected in the epileptic hippocampus. As a result, our current understanding of the mechanisms underlying epileptogenesis and seizure generation is largely based on information obtained in the hippocampus. Damage to the other components of the temporal lobe network, however, and their contributions to epileptogenesis as well as to the behavioral impairments associated with TLE, have remained relatively unexplored. Interest in TLE-associated structural damage in the amygdala grew in the 1950s when several authors reported damage to the amygdala in patients who had died from status epilepticus (Table 1; for review, see Gloor (1992)). However, it wasn’t until the observations of Feindel and Penfield (1954) and Penfield and Jasper (1954) who stimulated the amygdala of patients undergoing surgery for TLE that the specific involvement of the amygdala in the symptomatology of the seizures of temporal lobe origin was demonstrated. Another landmark discovery for the role of the amygdala in epilepsy was the finding by Goddard et al. (1969), who showed that the amygdala had one of the lowest thresholds for kindling. In recent years, research interest in the amygdala has experienced a renaissance of discoveries, providing us with new insights into its anatomy and function which undoubtedly will have an impact on studies of epilepsy. For example, in humans the amygdala has been shown to be involved in functions such as the recognition of emotion in visual and auditory stimuli (Adolphs et al., 1994; Young et al., 1995; Bonda et al., 1996; Breiter et al., 1996; Irwin et al., 1996; Mor-

ris et al., 1996; Scott et al., 1997), the acquisition of conditioned responses to sensory stimuli (Bechara et al., 1995; LaBar et al., 1995), and the acquisition (Cahill et al., 1996) and retrieval (Rauch et al., 1996) of memories for emotionally arousing events. Using the rat as an experimental system, investigators have found that the functions of the amygdala range from emotion to attention to memory (Davis, 1992; Gallagher and Holland, 1994; Cahill et al., 1996; Rogan and LeDoux, 1996). In this review, we will first summarize the major aspects of the anatomical organization of the amygdala. We will then describe the pattern of amygdala damage that is known to be associated with seizures and epilepsy in primates and rodents, and identify the conditions that are necessary to generate such damage. Finally, we will propose some hypotheses regarding how damage to the amygdala may impair intra-amygdala information processing; we will focus primarily on TLE since most of the information available on this topic comes from epileptic patients of this type.

2. Organization of the connections of the amygdala

2.1. Nuclei and nuclear subdi6isions The amygdala or rather the ‘amygdala complex’ in rat, monkey, and human is composed of more than ten nuclei and their subdivisions (Fig. 1) which have different cytoarchitectonic, chemoarchitectonic, and connectional characteristics. We will consider each subdivision to represent a functional unit in the amygdala (Pitka¨nen et al., 1997).

2.2. Input projections terminate in selecti6e nuclei of the amygdala Tract-tracing studies in rodents and primates have shown that information from different brain areas enters the amygdaloid complex via select nuclei (Amaral et al., 1992). For example, cortical and thalamic sensory information enters the

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

235

Fig. 1. Brightfield photomicrographs from coronal sections of the (A) rat, (B) monkey, and (C) human amygdala. In all species, the amygdaloid complex is composed of more than ten nuclei and their subdivisions. In humans, the amygdaloid complex can be divided into the deep nuclei, which include the lateral (medial and lateral divisions), basal (magnocellular, intermediate, and parvicellular divisions), accessory basal (magnocellular, parvicellular, and ventromedial divisions), and paralaminar nuclei. The superficial nuclei include the anterior cortical nucleus, the medial nucleus, the periamygdaloid cortex (divided into PACs, PACo, PAC1, and PAC3), the posterior cortical nucleus, and the nucleus of the lateral olfactory tract. The other nuclei include the central nucleus (medial and lateral divisions), anterior amygdaloid area, amygdalohippocampal area, and intercalated nuclei (Sorvari et al., 1995). The partitioning of the monkey amygdala was recently presented by Amaral et al. (1992) and the rat amygdala by Pitka¨nen et al. (1997) (see Figs. 5 and 6). Abbreviations: AB, accessory basal nucleus; B, basal nucleus; CE, central nucleus; L, lateral nucleus, M, medial nucleus; PAC, periamygdaloid cortex; PL, paralaminar nucleus. Scale bar of 2 mm applies to all panels.

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

236

Table 1 Histopathological studies on the amygdala damage in humans Author

Specimen

Patients

Observation

Brockhaus (1938)

Autopsy

Two children

Sano and Malamud (1953) Meyer et al. (1955)

Autopsy

29 adults

Autopsy

One child

Cavanagh and Meyer (1956)

Surgery

Six children/34 adults

Fowler (1957)

Autopsy

Four children

Falconer et al. (1964) Norman (1964)

Surgery

100 patients (majority adults) 11 children

‘Status marmoratus’ of the amygdala. No clear association of damage with epilepsy. Amygdala gliosis in 11/29 patients with HC sclerosis. Amygdala damage was bilateral in 8/11 cases Bilateral amygdala damage in a 9-year old boy who died from SE (seizures and several episodes of SE over a period of 1 year). Damage was located in ‘the ventral half of the amygdala’. Amygdala was analyzed in 10/17 patients with HC sclerosis. Six cases had amygdala damage characterized as a focal cell loss in ‘the basal group’. Other cases had severe gliosis in the amygdala. Three of four children with SE associated with fever had amygdala damage analyzed within 18 weeks after the initial insult. In one case the amygdala damage was bilateral. Patchy nerve cell loss and gliosis in the amygdala.

Autopsy

Margerison and Corsellis (1966)

Autopsy

55 patients (adults/children)

Ounstedt et al. (1966)

Autopsy

One child

Bruton (1988)

Surgery

249 patients (adults/children)

Hudson et al. (1993) Miller et al. (1994) Cendes et al. (1995)

Surgery

16 adults

Surgery

113 patients (adults/children) One adult

Autopsy

Amygdala was damaged in 6/11 cases (4/6 52 years of age) who had experienced SE associated with fever within the past 2 weeks before autopsy. Two patients with amygdala damage had no prior history of epilepsy. Amygdala damage was found in 15/55 patients. It was bilateral in five cases. Amygdala damage was always associated with HC damage. In 13/15 cases with amygdala damage also the thalamus, cerebellum or cortex was damaged. Patchy nerve cell loss and gliosis were most clear in ‘the basolateral nuclear group’. A child with recurrent generalized seizures and mental retardation died at the age of 12 months (case M.H.). Bilateral neuronal loss and fibrous gliosis were found in ‘the ventral part of the amygdaloid nucleus’. Amygdala damage was found in 81/92 patients with HC sclerosis. Depending on the severity of the HC damage they report amygdala damage in 1/3 patients with ‘end folium sclerosis’, 43/52 patients with ‘classical’ Ammon’s horn sclerosis, and in 37/37 patients with ‘total’ Ammon’s horn sclerosis. Neuronal loss and gliosis was found in the lateral nucleus also without concomittant hippocampal pathology. Isolated amygdala sclerosis in 10% (11/113) of the patients. In 53% of cases, both the amygdala and HC were damaged. Patchy neuronal loss in ‘the medial and basal portions’ of the amygdala in a man who died 3 1/4 years after developing TLE due to domoic acid intoxication.

HC, hippocampus; SE, status epilepticus, TLE, temporal lobe epilepsy.

amygdala largely, though not exclusively, via the lateral nucleus. Projections from the frontal cortex terminate primarily in the lateral, basal, and accessory basal nuclei. Inputs from the hippocampal formation terminate in the basal nucleus, those from the hypothalamus in the accessory basal, medial, and central nuclei, and projections from the brainstem in the central nucleus (for

details, see Price et al. (1987), Amaral et al. (1992)).

2.3. Three le6els of intra-amygdaloid connecti6ity The major principles of the organization of the intra-amygdaloid circuitries in rat are summarized in Fig. 2 (Pitka¨nen et al., 1997). After entering the

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

237

Fig. 2. Summary of the major principles of the organization of the amygdaloid circuitry in rat. The inputs enter the amygdala via select nuclei or nuclear subdivisions. From these nuclei, the information is distributed to different locations within the amygdala by intra-amygdaloid circuitries. Finally, the outputs originating in select amygdaloid areas convey the information to other functional systems in the brain like the temporal lobe memory system, autonomic centers in the brainstem, or to the motor system. Abbreviations: AB, accessory basal nucleus; B, basal nucleus; CE, central nucleus; COa, anterior cortical nucleus; Ldl, dorsolateral division of the lateral nucleus; Lm, medial division of the lateral nucleus; Lvl, ventrolateral division of the lateral nucleus; M, medial nucleus; PAC, periamygdaloid cortex.

amygdala, the information may travel to other locations within the amygdala via intra-amygdala connections. These local pathways have three different levels. Intradivisional connections transfer information within a subdivision. For example, the rostral portion of the magnocellular division is heavily connected with other neurons in that division. Interdivisional connections are links between regions of a particular nucleus. For example, the dorsolateral division of the lateral nucleus projects to the medial division of the lateral nucleus. Internuclear pathways connect various amygdaloid nuclei with each other. For example, the lateral nucleus, which provides the most substantial intra-amygdala connections, innervates the basal nucleus, the accessory basal nucleus, the medial nucleus, the amygdalohippocampal area, the central nucleus, the posterior cortical nucleus, and the periamygdaloid cortex. Via these different intra-amygdala connections, information entering one nucleus of the amygdala may have representations in various locations within the amygdala,

and consequently become associated with input from other functional systems of the brain.

2.4. Reciprocal intra-amygdala connections Many of the intra-amygdala connections are reciprocal (Fig. 2). For example, most of the amygdala nuclei that receive input from the lateral nucleus (e.g. the basal nucleus, the accessory basal nucleus, the periamygdaloid cortex) project back to the lateral nucleus. These pathways may provide routes by which target neurons can regulate the responsiveness of their input regions.

2.5. Con6ergence of intra-amygdala connections in select regions There are a few nuclei, such as the central nucleus and the amygdalohippocampal area, that receive convergent inputs from several amygdala nuclei but do not send any substantial inputs back to the other amygdala areas (Fig. 2). These nuclei

238

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

are presumed to act primarily as output stations from the amygdala to other brain regions, to evoke the appropriate behavioral responses to stimuli entering the amygdala.

2.6. Output projections to other brain areas originate in select regions of the amygdala The outputs to various functional systems of the brain leave the amygdala from different locations. For example, the lateral nucleus projects heavily to the entorhinal and perirhinal cortices. The basal nucleus projects to the basal forebrain, hippocampus, striatum and frontal cortex. The accessory basal nucleus provides substantial projection to the temporal and orbitofrontal cortices as well as to the striatum, hypothalamus, and basal forebrain. The medial nucleus projects heavily to the hypothalamus, and the central nucleus sends prominent projections to the brainstem (for details of these connections in different species, see Price et al. (1987) for rat and monkey; Amaral et al. (1992) for monkey).

2.7. Amygdala anatomy and seizure generation What is there in the anatomy of the amygdala that would link it so closely with epileptogenesis and seizure generation in TLE? First, as tracttracing studies have shown, the amygdala complex receives monosynaptic inputs from large areas of the frontal and temporal cortices that may generate and propagate seizure activity to the amygdala from foci located in these regions. Second, the smallest functional units of the amygdala, the nuclear subdivisions, often have a dense intradivisional network of connections. This suggests that activation of a small portion of a division by afferent inputs could rapidly recruit a large number of neurons within that division. Third, via intra-amygdala connections, the seizure activity may become monosynaptically distributed in parallel to various amygdala nuclei. Fourth, outputs from the amygdala to the extrapyramidal system, cortex, and hippocampal formation are even more widespread than the inputs from these areas to the amygdala; these pathways may provide routes by which the amygdala activity can

rapidly recruit other regions of the brain. Fifth, in rodents the two amygdalae are interconnected monosynaptically, which may explain the rapid contralateral activation produced by seizures elicited in one amygdala (Savander et al., 1997). In primates, however, monosynaptic inter-amygdala connections have not been described (Amaral et al., 1992). Sixth, recent electrophysiological studies have proposed that interconnections between the amygdala and the entorhinal cortex underlie the coherent oscillations observed in amygdalahippocampal circuitries (Pare´ and Gaudreau, 1996). And finally, the fact that each of the amygdala nuclei has unique anatomical characteristics suggests that the functional consequences of seizure-induced neuronal damage to the amygdala are largely dependent on the nuclear location of the damage within the amygdala.

3. Amygdaloid damage in epilepsy

3.1. Human studies 3.1.1. Histopathology Neuronal loss and gliosis in the amygdala have been reported in a large number of histopathological studies in which amygdala tissue from humans with chronic epilepsy or status epilepticus was available for analysis either from autopsy or epilepsy surgery (Table 1). Amygdala damage was found both in adults (Sano and Malamud, 1953; Cavanagh and Meyer, 1956; Falconer et al., 1964; Margerison and Corsellis, 1966; Bruton, 1988; Hudson et al., 1993; Miller et al., 1994; Cendes et al., 1995) as well as in children, some of whom were under 2 years of age (Meyer et al., 1955; Fowler, 1957; Norman, 1964; Ounstedt et al., 1966). In many of the early studies, amygdala damage was found in patients who had experienced recent status epilepticus (Meyer et al., 1955; Fowler, 1957; Norman, 1964; Fujikawa and Itabashi, 1994), which in some studies was associated with fever (Fowler, 1957; Norman, 1964). Interestingly, some of these patients did not have epilepsy prior to the episode of status (Norman, 1964). Based on the data available, amygdala damage may become apparent over a period of a

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

few days or a few weeks following status epilepticus (Fowler, 1957; Norman, 1964). However, many studies show that amygdala damage may also occur in patients with TLE but who have no prior history of status epilepticus (Margerison and Corsellis, 1966; Bruton, 1988; Hudson et al., 1993). Amygdala damage may be either unilateral or bilateral (Sano and Malamud, 1953; Fowler, 1957; Margerison and Corsellis, 1966). Most often, it has been reported to occur in combination with hippocampal damage or with damage to the cerebral cortex, cerebellum, or thalamus. In material analyzed by Bruton (1988), the percentage of patients with amygdala damage increased with the severity of the hippocampal damage (Table 1). More recently, however, Hudson et al. (1993) described eight patients with amygdala damage who did not have any apparent neuronal loss in the hippocampus. Miller et al. (1994) investigated a series of 113 patients undergoing temporal lobe surgery and also found isolated amygdala sclerosis in approximately 10% of the patients. Even though we do not have a detailed analysis available of the distribution of damage to various nuclei of the human amygdala, there is some evidence showing nuclear specificity. Early reports mention that the ‘ventral part of the amygdala’ or ‘the basal group of the amygdala’ was the most damaged portion of the amygdala (Meyer et al., 1955; Cavanagh and Meyer, 1956; Ounstedt et al., 1966). Moreover, Margerison and Corsellis (1966) found neuronal loss and/or gliosis in ‘the basolateral nuclear group’. In Fig. 11 of Meyer et al. (1955) and Fig. 25 of Ounstedt et al. (1966), it appears that most of the gliosis is in a region that involves the lateral (medial division) and basal (parvicellular division) nuclei. In a study by Hudson et al. (1993), neuronal damage and gliosis were identified in the lateral nucleus. Also in our material, we found substantial gliosis in the medial division of the lateral nucleus in patients with TLE (Fig. 3B–D).

3.1.2. MR 6olumetry Detection of amygdala damage in vivo using volumetric measurements of the amygdala by MR (magnetic resonance) imaging has provided us

239

with a new tool to investigate in more detail the factors that lead to amygdala damage. One example of the appearance of the amygdala damage in MR image is shown in Fig. 3A. Caution should be taken when interpreting data from volumetry studies because the relative sensitivity of this method in detecting amygdala damage, compared to histological analysis, has not yet been established. For example, in a study by Miller et al. (1994), none of the 11 patients with histologically verified amygdala damage had definitive atrophy or an increased T2 signal in the amygdala. MR volumetric measurements of the amygdala revealed that the reduction in the amygdala volume in patients with drug-refractory TLE varies between 10–30% (Cendes et al., 1993a,b,c; Saukkonen et al., 1994; Bronen et al., 1995; Ka¨lvia¨inen et al., 1997). In our ongoing study where we have measured the volume of the amygdala by MR imaging in 147 patients with TLE, the volume of the smallest amygdala was 57% of that in control subjects (Salmenpera¨, Ka¨lvia¨inen, Partanen, Vainio and Pitka¨nen, unpublished data). In a study by Cendes et al. (1993a), the most pronounced atrophy of the amygdala (a 30% volume reduction) was found in drug-refractory patients with TLE who had experienced prolonged febrile convulsions in childhood. In 26% (10/39) of all patients, the amygdala damage (] 2 S.D. reduction)1 was bilateral (i.e. in 45% of patients with any amygdala damage). As in histopathological studies, MR volumetry has demonstrated that amygdala damage typically appears in combination with hippocampal damage. For example, in the patients of Cendes et al. (1993a), a large majority with amygdala damage also had hippocampal damage. More recently, Bronen et al. (1995) reported by visually inspecting the MR images that amygdala damage was In MR volumetry, \1 S.D. or \2 S.D. reduction in the volume of the amygdala refers to the volume of the amygdala that was one or two standard deviations smaller than the mean amygdala volume in controls, respectively. In our material, a 1 S.D. volume reduction equals a 16% volume reduction on the left and a 11% volume reduction on the right, compared to the mean volume on the left or right amygdala in controls, respectively. A 2 S.D. volume reduction equals a 32% volume reduction on the left and a 22% volume reduction on the right. 1

240

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

Fig. 3. Amygdala damage in MR image and histological sections. (A) A coronal MR image from a 46-year-old male who was operated due to drug-refractory TLE. The right amygdala (open arrow) is atrophied and its normalized volume is 75% of that in control subjects. Also, the T2 relaxation time of the right amygdala was prolonged (108 ms). The patient had a 4-year history of epilepsy, during which he was estimated to have experienced 1200 seizures. Also, the volume of his right hippocampus was reduced (55% of that in controls, not shown). The asterisk indicates the location of the lateral nucleus. (B) Brightfield photomicrograph of a thionin-stained section from the lateral nucleus of the amygdala of the patient illustrated in panel A. (C) Brightfield photomicrograph from an adjacent section stained with an antibody raised against parvalbumin (PARV) or (D) glial fibrillary acid protein (GFAP). The level of histological section is rostral to that in the MR image. Note the increased astrocytosis in the ventromedial aspect of the lateral nucleus in panel D (area between the arrowheads) which corresponds to a region that is lightly stained in parvalbumin preparations in panel C. Abbreviations: L, lateral nucleus; ec, external capsule. Scale bars: A, 10 mm; B – D, 2 mm.

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

241

Table 2 MR volumetry of the amygdala in patients with TLE Patients with volume reduction of the amygdala (%) ]1 S.D.

]2 S.D.

All patients with TLE (147)

29

6

Etiology From patients with cryptogenic etiology From patients with symptomatic etiology

28 31

5 8

Duration of epilepsy From newly diagnosed patients From chronic patients

18 33

3 7

Amygdala damage bilateral Associated with ]1 S.D. volume reduction of the HC ]2 S.D. volume reduction of the HC

8

0

56 42

89 67

Isolated volume reduction of the amygdala (simultaneous HC damage 51 S.D.)

13

1

The data comes from an ongoing study, in which we have analyzed the amygdala volumes of 147 patients with temporal lobe epilepsy by using MR volumetry (Salmenpera¨, Ka¨lvia¨inen, Partanen, Vainio, and Pitka¨nen, in preparation). For methodology and clinical criteria for classification of patients, see Ka¨lvia¨inen et al. (1997). A 1 S.D. volume reduction equals a 16% volume reduction on the left and a 11% volume reduction on the right compared to the mean volume of the left or right amygdala in controls, respectively. A 2 S.D. volume reduction equals a 32% volume reduction on the left and a 22% volume reduction on the right. HC, hippocampus; MR, magnetic resonance; S.D., standard deviation of the mean; TLE, temporal lobe epilepsy.

present in 12% (7/52) of the patients with pathologically proven hippocampus sclerosis. However, amygdala damage was apparent in MR images also in cases where no evidence of hippocampal atrophy was detected. Cendes et al. (1993a) reported three patients out of 39 (8%) in which amygdala damage was more pronounced than the hippocampal damage. In our series of 147 patients with TLE, isolated ] 1 S.D. reduction in the amygdala volume was found in 13% of the patients with TLE, whereas isolated ]2 S.D. reduction in the amygdala volume was found in less than 1% of the patients (Table 2) (Salmenpera¨, Ka¨lvia¨inen, Partanen, Vainio and Pitka¨nen, unpublished data). Tract-tracing studies have shown that the amygdala receives monosynaptic inputs from various extratemporal cortical regions in primates (Amaral et al., 1992). This connectivity leads to the question of whether the amygdala is damaged in extratemporal epilepsy. Cendes et al. (1993b) found no amygdala damage in six patients with

an extratemporal seizure focus. However, their original finding was challenged in a later study (Cendes et al., 1993c) in which they described bilateral amygdala damage in two out of seven patients with an extratemporal focus. In our own studies, ] 1 S.D. reduction in the amygdala volume was found in 3% (1/36) of the patients with an extratemporal seizure focus (Salmenpera¨, Ka¨lvia¨inen, Partanen, Vainio and Pitka¨nen, unpublished data). Is amygdala damage already present at the time of epilepsy diagnosis or does it appear later? In our series of patients, which included both newly diagnosed and chronic patients, we found that 18% of the newly diagnosed patients had ] 1 S.D. reduction in the amygdala volume, whereas 33% of the patients with chronic TLE showed a similar volume reduction. Over 2 S.D. volume reduction was found in only 3% of the newly diagnosed and in 7% of the chronic patients (Table 2) (Salmenpera¨, Ka¨lvia¨inen, Partanen, Vainio and Pitka¨nen, unpublished data). Interest-

242

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

ingly, there was a significant correlation between lifetime seizure number and volume reduction in the amygdala (Fig. 4) (Ka¨lvia¨inen et al., 1997), which suggests that amygdala damage progresses as the number of seizures over a lifetime increases. Another interesting question is whether amygdala damage is associated with the etiology of TLE. Previous MR imaging studies have shown that, for example, herpes encephalitis, which may be a preceding cerebral insult leading to epilepsy, causes widespread damage to various brain areas including the amygdala (Kapur et al., 1994). Moreover, according to Cendes et al. (1993a), prolonged febrile seizures in childhood may be associated with pronounced damage to the amygdala. This contrasts with the observations of Miller et al. (1994) who found that in patients with isolated amygdala sclerosis, their first seizure appeared later than in patients with amygdalahippocampus sclerosis, and they had no clinical history of seizures in early childhood. In our series of patients, amygdala damage ] 1 S.D.

Fig. 4. The reduction of the right amygdala volume as a function of the logarithm of lifetime seizure number. It can be estimated that a 20% volume reduction of the amygdala required approximately 4000 seizures. Abbreviations: C, controls; n, number of patients with TLE; P, statistical significance (Pearson’s correlation test); r, correlation coefficient. Original data was presented in Ka¨lvia¨inen et al. (1997).

over control was found in 28% of patients with cryptogenic TLE and in 31% of patients with symptomatic TLE, which suggests that amygdala damage is not any more common in patients with a defined seizure etiology (Table 2) (Salmenpera¨, Ka¨lvia¨inen, Partanen, Vainio and Pitka¨nen, unpublished data). What are the symptoms associated with amygdala damage? Cendes et al. (1994) studied 50 patients with drug-refractory TLE, and found that 34% of the patients recognized fear as a component of their seizures. The volume of the amygdala in patients experiencing fear was 75% of that in control subjects, whereas the patients having no fear had an amygdala volume 91% of that in controls. Miller et al. (1994) investigated patients with isolated amygdala sclerosis and found that amygdala damage was associated with a more widespread EEG abnormality and with a higher tendency of seizures to become generalized compared to patients with amygdala-hippocampal sclerosis. Moreover, the patients with amygdala sclerosis did not show improvement in IQ following surgery, and they had a greater tendency towards postoperative seizures than did those with amygdala-hippocampal sclerosis.

3.1.3. T2 relaxometry T2 relaxation time is another measure obtained by MR imaging that is used to locate structural changes in epileptic brain. Van Paesshen et al. (1996) investigated T2 relaxation times in 82 patients with intractable TLE. They found that 54% (44/82) of the patients had an abnormal amygdala T2 time which was bilateral in 22% (18/82) of the patients. Interestingly, of the patients with unilateral prolongation of T2 in the hippocampus, 57% also had an abnormal amygdala T2 time. Moreover, isolated prolongation of amygdala T2 was found unilaterally in 18% (15/82) and bilaterally in 8% (7/82) of the patients. These authors found that patients with isolated amygdala T2 prolongation were older at the onset of epilepsy and rarely experienced febrile convulsions. We recently found ]2 S.D. prolongation of amygdala T2 relaxation time in approximately 20% of patients with TLE (Ka¨lvia¨inen et al., 1997). Like the volumetry findings, T2 prolongation occurred both

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

in newly diagnosed and in chronic patients. Moreover, T2 prolongation was also not clearly associated with either cryptogenic or symptomatic etiology. To summarize, histological, MR volumetry, and MR T2 relaxometry studies on humans show that unilateral or bilateral amygdala damage may occur both in adults and children with epilepsy. Seizure-associated neuronal loss and gliosis have been generally described in humans to be in the ventral portion (the lateral and basal nuclei) of the amygdala complex. Several types of brain insults may cause amygdala damage, including status epilepticus, encephalitis, and prolonged febrile seizures, each of which has been reported to be associated with a volume reduction of the amygdala. In line with these observations, our cross-sectional study showed that the amygdala volume may be reduced in patients with symptomatic epilepsy, though this was shown to be no different from patients with cryptogenic epilepsy. TLE-associated damage is isolated to the amygdala in approximately 10% of the patients. More typically it occurs in combination with damage to other structures of the brain. In extratemporal epilepsy where the seizure focus lies outside of the temporal lobe, amygdala damage is rare.

3.2. Nonhuman primates Studies in nonhuman primates have provided us with more detailed information on the pattern of seizure-induced damage to the amygdala. Wasterlain et al. (1996) described damage to the basolateral amygdala in 3 – 4 week-old marmoset monkeys following status epilepticus induced with lithium chloride and pilocarpine. Meldrum and Brierley (1973) induced seizures in adult baboons with bicuculline. They found damage to ‘the basolateral portions of the amygdala’ in seven of ten animals. ‘The centromedial portion’ was also damaged in four of the seven baboons. The duration of the seizures leading to damage in the temporal lobe of these animals lasted for 82 – 299 min. In another study, Meinini et al. (1980) injected kainic acid into the amygdala of adult Papio papio baboons. They found that damage to the amygdala was unilateral, but the accompany-

243

ing mild hippocampal damage was bilateral in some cases. It should be noted, that in all cases damage was also observed in several extra-amygdaloid brain areas.

3.3. Rats Studies of a number of rat seizure models have provided us with even more details of the specific pattern of seizure-induced damage to different amygdaloid nuclei and neuronal populations. In adult rats, amygdala damage has been reported in various epilepsy models, including those where seizures were induced by either kainic acid (Schwob et al., 1980; Tuunanen et al., 1996) or pilocarpine (Houser and Obenaus, 1994) injection or by electrical stimulation of the perforant pathway (Tuunanen et al., 1996) or the lateral nucleus of the amygdala (Nissinen et al., 1996). In addition, Baram and Ribak (1995) recently reported amygdala damage in 10–13 day-old infant rats that had experienced status epilepticus induced by intracerebroventricular injection of corticotropinreleasing hormone. Electrophysiological studies have also shown that the amygdala is prone to seizure-induced alterations. For example, the amygdala has one of the lowest thresholds for kindled seizures (Goddard et al., 1969). Moreover, after kainate-induced status epilepticus, spontaneous bursting activity first appears in the basal nucleus of the amygdala (Smith and Dudek, 1996). White and Price (1993) also showed that the activation of the basal nucleus is primarily responsible for the generation of widespread status epilepticus activity in models where seizures were evoked even in extra-amygdaloid regions. Furthermore, feed-forward GABAergic inhibition was found to be impaired in the kindled amygdala (Rainnie et al., 1992). What have we learned about the nucleus specificity of seizure-induced damage in the amygdala? Studies in the hippocampus have shown that the hilar neurons are among the first to die, whereas the cells in the CA2 region are relatively resistant to seizure-induced damage (Babb and Pretorius, 1993). Similarly in the amygdala, some regions seem to be more resistant to seizure damage than are others which is illustrated in Figs. 5–7A–C.

244

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

Fig. 5. Brightfield photomicrographs of thionin-stained sections showing the distribution of damage in the rat amygdala after status epilepticus was induced with kainic acid. (A) Rostral section from the normal rat amygdala. (B) Rostral section from the amygdala of a rat that was injected with kainic acid two days earlier. Note the damage to layer III of the anterior cortical nucleus (arrow). Otherwise, the magnocellular divisions of the basal nucleus and the central nucleus are well preserved. Abbreviations; Bmc, magnocellular division of the basal nucleus; CE, central nucleus (c, capsular division; CEl, lateral division; m, medial division); Ldl, dorsolateral division of the lateral nucleus; M, medial nucleus (Mcd, dorsal portion of the central division; Mr, rostral division), PAC, periamygdaloid cortex. Scale bar: 500 mm.

In models where status epilepticus was induced by systemically injecting kainic acid or electrically stimulating the perforant pathway, the amygdala regions most seriously damaged were the deep layers of the anterior cortical and medial nuclei,

the medial division of the lateral nucleus, the parvicellular division of the basal nucleus, the accessory basal nucleus, and the posterior cortical nucleus (Tuunanen et al., 1996). The most preserved regions were the magnocellular division of

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

245

Fig. 6. See legend to Fig. 5. (A) Caudal section from the normal rat amygdala. (B) Caudal section from a rat injected with kainic acid 2 days earlier. Note the preservation of the dorsolateral division of the lateral nucleus. Otherwise, the medial division of the lateral nucleus (arrow), the parvicellular division of the basal nucleus, the accessory basal nucleus, and the posterior cortical nucleus are heavily damaged. Abbreviations; ABmc, magnocellular division of the accessory basal nucleus; ABpc, parvicellular division of the accessory basal nucleus; AHAl, lateral division of the amygdalohippocampal area; AHAm, medial division of the amygdalohippocampal area; Bpc, parvicellular division of the basal nucleus; COp, posterior cortical nucleus; PAC, periamygdaloid cortex (PACm, medial division; PACs, sulcal division). Scale bar: 500 mm.

the basal nucleus, the dorsolateral division of the lateral nucleus, and the central nucleus (Tuunanen et al., 1996). How much seizure activity is sufficient to cause damage to the amygdala? Generally, seizure-induced structural damage has been assessed in

models of status epilepticus. Whether a single epileptic seizure can cause structural damage to the amygdala or to any other region of the brain is under dispute. Callahan et al. (1991) showed that amygdala-kindled rats perfused for histological purposes 2–6 months after experiencing three

246

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

Fig. 7. (A – C) Schematic drawings summarizing the distribution of damage in various nuclei and nuclear subdivisions of the amygdala in rat two weeks after the induction of status epilepticus with kainic acid. Panel A is the most rostral and Panel C is the most caudal. The darkness of the shading indicates the severity of damage (dotted pattern = mild, grey = moderate, black = severe damage). Panel D shows the correlation of the number of TUNEL positive neurons in the medial division of the lateral nucleus with the duration of epileptic EEG seizure activity (generalized rhythmic high-voltage sharp waves). Abbreviations: n, number of rats; P, statistical significance (Pearson’s correlation test); r, correlation coefficient. Anatomic abbreviations as in Figs. 5 and 6.

to five generalized seizures (not including the seizures that occurred during the induction of kindling) had a 37 – 64% loss of GABA-immunoreactive neurons in ‘the basolateral’ amygdala. We recently reinvestigated amygdala damage in an amygdala kindling model in which the animals had experienced five class 5 seizures by the end of the kindling procedure, which required eight to 11 stimulations. The total duration of the after-discharges varied between 343 – 628 s. We could not find any reduction in the total number of neurons in the lateral, basal, or accessory basal nuclei (Tuunanen and Pitka¨nen, un-

published data) or in the overall density of GABA-immunoreactive neurons (Fig. 8) (Tuunanen et al., 1997). Surprisingly, however, we found over a 35% decrease in the density of somatostatin-immunoreactive neurons, presumably GABAergic inhibitory neurons (McDonald and Pearson, 1989), in the medial division of the lateral nucleus and in the magnocellular division of the basal nucleus in the amygdala contralateral to the stimulation site (Tuunanen et al., 1997). These observations support the idea that a relatively low number of seizures may damage a subpopulation of inhibitory neurons in the amyg-

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

247

Fig. 8. (A) The percentage of GABA-immunoreactive (GABA-ir) neurons remaining in the lateral nucleus of the rat amygdala after various treatments. Note that kainate-treated animals had a substantial decrease in the density of GABA-ir neurons. (B) The percentage of somatostatin-immunoreactive neurons (SOM-ir) remaining in the lateral nucleus after various treatments. Groups Kind and KA correspond to those in panel A. Group WDS: rats that experienced wet-dog shakes (WDS) after a low dose of kainic acid was injected (5 mg/kg, intraperitoneally); Group Kind: rats that were amygdala-kindled 6 months earlier; Group PP: rats that had experienced status epilepticus induced by electrical stimulation of the perforant pathway 2 weeks earlier; Group KA: rats that had experienced kainate-induced status epilepticus (9 mg/kg, intraperitoneally) 2 weeks earlier. Statistical significances: * P B0.05, ** PB 0.01 compared to controls (Mann-Whitney U-test). Details of the studies can be found in the original publications (Tuunanen et al., 1996, 1997).

dala, which could contribute to the low kindling threshold in the amygdala. However, the more severe seizure activity that is associated with status epilepticus may cause damage not only to the somatostatin-ir neurons, but also to the general GABA-ir neuron population and to the pyramidal cells as well (see Fig. 8) (Tuunanen et al., 1996). What is the mechanism of seizure-induced neuronal damage in the amygdala? Data from recent studies of kainic acid model indicate that, in addition to necrotic cell death, apoptosis may contribute to neuronal damage. In our hands, rats that were killed 8 h after kainate injection had TUNEL positive neurons (an indicator of DNA fragmentation) in amygdaloid nuclei that are sensitive to seizure-induced damage. Moreover, the number of TUNEL positive neurons correlated with the duration of epileptiform activity in electroencephalogram (Fig. 7D) (Tuunanen and

Pitka¨nen, unpublished data). In silver-stained sections, damaged (presumably necrotic) neurons can be seen as early as 4 h after kainate injection (Tuunanen and Pitka¨nen, unpublished data).

4. Functional aspects

4.1. Nuclei targeted in seizure-induced amygdala damage Studies in rats have shown that the amygdala is a critical component of a network which generates appropriate behavioral responses to emotionally significant sensory stimuli (Rogan and LeDoux, 1996). In humans, the amygdala is also involved in the interpretation of emotional aspects of sensory stimuli. For example, patients with amygdala lesions have difficulties in recognizing fear in facial expressions (Adolphs et al., 1994) and in

248

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

perceiving changes in intonation in speech patterns (Scott et al., 1997). Anatomical studies have shown that the lateral nucleus of the amygdala is the major recipient of sensory information directed to the amygdala (Amaral et al., 1992). As an example, the dorsolateral division of the lateral nucleus that is relatively undamaged after status epilepticus, receives auditory information from the thalamus (Price et al., 1987). Via interdivisional connections the dorsolateral division conveys the representation of a tone to the medial division of the lateral nucleus, which is one of the most sensitive regions of the amygdala to seizureinduced neuronal damage. The medial division in turn provides substantial projections to the other amygdaloid nuclei (Pitka¨nen et al., 1995). Therefore, even though the neurons receiving the sensory information may reside in unaffected portions of the lateral nucleus, damage to the medial division interferes with the connections of the lateral nucleus with the rest of the amygdala, and consequently impairs processing of sensory information within the amygdaloid circuits. A schematic diagram summarizing the presumed pattern of disrupted amygdaloid circuits is presented in Fig. 9. The pattern of amygdaloid damage that can be extracted from the little data available in humans suggests that it is the ventral portion of the lateral nucleus that is damaged in human epilepsy (Ounstedt et al., 1966; Hudson et al., 1993). Interest-

ingly, previous anatomic studies suggest that the chemoarchitectonic and connectional characteristics of the ventral portion of the lateral nucleus in primates resemble that of the medial division of the lateral nucleus in rodents (Amaral et al., 1992; Pitka¨nen et al., 1997). For example, like in rodents, also in nonhuman primates the ventral portion of the lateral nucleus gives origin to widespread intra-amygdala projections (Pitka¨nen and Amaral, 1991). It remains to be determined whether the recognition of emotions in sensory signals is impaired in patients with epilepsy and amygdala damage. The preservation of the magnocellular division of the basal nucleus and of the central nucleus is of interest since they project to the striatum and the brainstem autonomic areas, respectively (Amaral et al., 1992). Preservation of the amygdala input to the striatum would suggest that the pathways mediating the spread of the motor component of behavioral seizures remain intact within the epileptic amygdala (Fig. 9). Motor and autonomic components of the fear response should also remain, since the central nucleus mediates these behaviors via its projections to the brainstem and hypothalamus. Therefore, though some of the amygdaloid nuclei may be damaged by seizures, the remaining areas may still convey the behavioral manifestations of seizure activity to wide areas of the motor and autonomic centers in the brain.

Fig. 9. Schematic diagram summarizing the seizure-induced disruption of the amygdaloid circuitries in the lateral, basal, accessory basal, and central nuclei and how the damage may compromise the functions of the amygdala. Panel A shows how in the normal amygdaloid complex the sensory information enters the amygdala via the lateral nucleus. A substantial portion of the lateral nucleus outputs are directed to the other amygdaloid nuclei, including the basal nucleus (B), accessory basal nucleus (AB), and central nucleus (CE). These nuclei, on the other hand, send outputs to the extra-amygdaloid areas where many of the symptoms of the amygdaloid seizures are generated. The basal nucleus projects to the striatum. It is also reciprocally connected with the subiculum/CA1 border of the hippocampus and the ventral subiculum. The accessory basal nucleus also projects to the hippocampus. The central nucleus projects to the brainstem and to the hypothalamic autonomic and endocrine centers (data taken from Price et al. (1987), Pitka¨nen et al. (1997). In Panel B the dark shading indicates the seizure-induced damage to selective regions of the lateral, basal, and accessory basal nuclei. Dashed lines indicate impaired connectivity. We hypothesize that damage to the ventrolateral and medial divisions of the lateral nucleus may impair access of sensory information from the lateral nucleus to the rest of the amygdala. Damage to the parvicellular division of the basal nucleus and to the accessory basal nucleus may interfere with the information flow between the amygdala and the hippocampal formation, and thus, impair the memory processing. Lighter shading shows the amygdaloid regions that were found to be relatively well preserved after status epilepticus that was induced by kainic acid injection or by electrical stimulation of the perforant pathway (Tuunanen et al., 1996). The magnocellular and intermediate divisions of the basal nucleus project to the striatum, which may be a pathway by which amygdaloid seizures manifest themselves as secondarily generalized convulsions. On the other hand, outputs from the almost undamaged central nucleus may mediate the autonomic and endocrine symptoms of fear associated with amygdaloid seizures. Abbreviations as in Figs. 5 and 6.

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

Fig. 9.

249

250

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

4.2. Epileptic acti6ity originating in the amygdala may cause damage to the other regions of the temporal lobe The loss of inhibitory neurons in select subdivisions of the amygdala nuclei may underlie the low seizure threshold that is observed in this region in amygdala-kindling or kainate-induced status epilepticus in rat. The remaining excitatory neurons presumably transfer the seizure activity from one amygdaloid nucleus to another as well as when the seizures spread from the amygdala to a variety of other brain areas monosynaptically. In fact, although this is not always appreciated, the seizure activity that is initiated by the stimulation of the amygdala has on several occasions been reported to cause secondary damage in the hippocampus. For example, we found that loss of hilar neurons and sprouting of mossy fibers in the dentate gyrus occurred in rats that developed spontaneous seizures after a self-sustained status epilepticus was induced by electrical stimulation of the lateral nucleus of the amygdala (Nissinen et al., 1996). Damage was also found in the entorhinal cortex. Loss of hilar cells and mossy fiber sprouting has also been reported in amygdala-kindled rats (Sutula et al., 1988; Cavazos et al., 1991) or after injecting kainic acid into the rat amygdala (Mascott et al., 1994). Similarly in nonhuman primates, pathological findings in the hippocampus were reported after injecting kainic acid (Meinini et al., 1980) or alumina gel (Ribak et al., 1995) into the amygdala. More recently, Behr et al. (1996) reported an altered electrophysiological responsiveness of CA1 pyramidal cells to the stimulation of the entorhinal cortex in slices that were prepared from amygdala-kindled rats. Also, in humans, there is evidence that seizure activity may spread from the epileptic amygdala to the hippocampus or to the surrounding cortex (Quesney, 1986; So et al., 1989; Wilson et al., 1990; Gotman and Levtova, 1996; Bertram, 1997). These observations raise the question of what the specific role is of each of the temporal lobe structures in the generation and propagation of seizure activity among the myriad of connections which link these areas together.

5. Conclusions The amygdaloid complex is one component of the temporal lobe that is damaged in a large subpopulation of patients with TLE. Future studies will show if amygdala pathology plays a critical role in epileptogenesis, and how amygdala interacts with other components of the temporal lobe network to generate temporal lobe seizures.

Acknowledgements This study was supported by the Academy of Finland, the Vaajasalo Foundation, and the Sigrid Juselius Foundation. The help from Dr David G. Amaral (UC at Davis, CA) in the preparation of Fig. 1 is greatly appreciated.

References Adolphs, R.D., Tranel, H., Damasio, H., Damasio, A.R., 1994. Impaired recognition of emotion in facial expressions following bilateral damage to the human amygdala. Nature 372, 669 – 672. Amaral, D.G., 1987. Memory: anatomical organization of candidate brain regions. In: Mountcastle, V.B. (Ed.). Williams and Wilkins, Baltimore, pp. 211 – 294. Amaral, D.G., Price, J.L., Pitka¨nen, A., Carmichael, S.T., 1992. Anatomical organization of the primate amygdaloid complex. In: Aggleton J.P. (Ed.), The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction. Wiley-Liss, New York, pp. 1 – 66. Babb, T.L., Pretorius, J.K., 1993. Pathologic substrates of epilepsy. In: Wylie, E. (Ed.), The Treatment of Epilepsy: Principles and Practice. Lea&Febiger, Philadelphia, pp. 55 – 70. Baram, T.Z., Ribak, C.E., 1995. Peptide-induced infant status epilepticus causes neuronal death and synaptic reorganization. Neuroreport 6, 277 – 280. Bechara, A., Tranel, D., Damasio, H., Adolphs, R., Rockland, C., Damasio, A.R., 1995. Double dissociation of conditioning and declarative knowledge relative to the amygdala and huippocampus in humans. Science 269, 111 – 115. Behr, J., Gloveli, R., Gutierrez, R., Heinemann, U., 1996. Spread of low Mg2 + induced epileptiform activity from the entorhinal cortex to hippocampus after kindling studied in vitro. Neurosci. Lett. 216, 41 – 44. Bertram, E.H., 1997. Functional anatomy of spontaneous seizures in a rat model of limbic epilepsy. Epilepsia 38, 95 – 105.

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253 Bonda, E., Petrides, M., Ostry, D., Evans, A., 1996. Specific involvement of human parietal systems and the amygdala in perception of biological motion. J. Neurosci. 16, 3737– 3744. Bouchet, C., Cazauvieilh, C., 1825. De l’epilepsie consideree dans ses rapports avec la l’alienation mentale. Arch. Gen. Med. 9, 510 – 542. Breiter, H.C., Etcoff, N.L., Whalen, P.J., Kennedy, W.A., Rauch, S.L., Buckner, R.L., Strauss, M.M., Hyman, S.E., Rosen, B.R., 1996. Response and habituation of the human amygdala during visual processing of facial expression. Neuron 17, 875–887. Brockhaus, H., 1938. Zur normalen und patologischen Anatomie des Mandelkerngebietes. J. Psychol. Neurol. 49, 1–136. Bronen, R.A., Fulbright, R.K., Kim, J.H., Spencer, S.S., Spencer, D.D., Al-Rodham, N.F.R., 1995. Regional distribution of MR findings in hippocampal sclerosis. Am. J. Neuroradiol. 16, 1193–1200. Bruton, C.J., 1988. The Neuropathology of Temporal Lobe Epilepsy. Oxford University Press, Oxford. Cahill, L., Haier, R.J., Fallon, J., Alkire, M.T., Tang, C., Keator, D., Wu, J., McGaugh, J.L., 1996. Amygdala activity at encoding correlated with with long-term, free recall of emotional information. Proc. Natl. Acad. Sci. USA 93, 8016 – 8021. Callahan, P.M., Paris, J.M., Cunningham, K.A., ShinnickGallagher, P., 1991. Decrease of GABA-immunoreactive neurons in the amygdala after electrical kindling in the rat. Brain Res. 555, 335 –339. Cavanagh, J.B., Meyer, A., 1956. Aetiological aspects of Ammon’s horn sclerosis associated with temporal lobe epilepsy. Br. Med. J. 2, 1403–1407. Cavazos, J., Golarai, G., Sutula, T., 1991. Mossy fiber synaptic reorganization induced by kindling: time course of development, progression and permanence. J. Neurosci. 11, 2795 – 2803. Cendes, F., Anderman, F., Dubeau, F., Gloor, P., Evans, A., Jones-Gotman, M., Olivier, A., Anderman, E., Robitaille, Y., Lopes-Cendes, I., Peters, T., Melanson, D., 1993a. Early childhood prolonged febrile convulsions, atrophy and sclerosis of mesial structures, and temporal lobe epilepsy: a MRI volumetric study. Neurology (NY) 43, 1083 – 1087. Cendes, F., Leproux, F., Melanson, D., Ethier, R., Evans, A., Peters, T., Anderman, F., 1993b. MRI of amygdala and hippocampus in temporal lobe epilepsy. J. Comp. Assist. Tomogr. 17, 206 – 210. Cendes, F., Anderman, F., Gloor, P., et al., 1993c. MRI volumetric measurements of amygdala and hippocampus in temporal lobe epilepsy. Neurology (NY) 43, 719–725. Cendes, F., Anderman, F., Gloor, P., Gambardella, A., LopesCendes, I., Watson, G., Evans, A., Carpenter, S., Olivier, A., 1994. Relationship between atrophy of the amygdala and ictal fear in temporal lobe epilepsy. Brain 117, 739– 746.

251

Cendes, F., Anderman, F., Carpenter, S., Zatorre, R.J., Cashman, N.R., 1995. Temporal lobe epilepsy caused by domoic acid intoxication: evidence for glutamate receptor mediated excitotoxicity in humans. Ann. Neurol. 37, 123 – 126. Davis, M., 1992. The role of the amygdala in conditioned fear. In: Aggleton, J.P. (Ed.), The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction. Wiley-Liss, New York, pp. 255 – 305. Falconer, M.A., Serafetinides, E.A., Corsellis, J.A.N., 1964. Etiology and pathogenesis of temporal lobe epilepsy. Arch. Neurol. 10, 233 – 248. Feindel, W., Penfield, W., 1954. Localization of discharge in temporal lobe automatism. Arch. Neurol. Psychiatr. 72, 605 – 630. Fowler, M., 1957. Brain damage after febrile convulsions. Arch. Dis. Child. 32, 67 – 76. Fujikawa, D.G., Itabashi, H.H., 1994. Status epilepticus-induced brain damage in humans. Epilepsia 35 (Suppl. 8), 9. Gallagher, M., Holland, P.C., 1994. The amygdala complex: multiple roles in associative learning and emotion. Proc. Natl. Acad. Sci. USA 91, 11771 – 11776. Gloor, P., 1992. Role of amygdala in temporal lobe epilepsy. In: Aggleton, J.P. (Ed.), The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction. Wiley-Liss, New York, pp. 505 – 538. Goddard, G.V., McIntyre, D.C., Leech, C.K., 1969. A permanent change in brain function resulting from daily electrical stimulation. Exp. Neurol. 25, 295 – 330. Gotman, J., Levtova, V., 1996. Amygdala-hippocampus relationships in temporal lobe seizures. A phase-coherence study. Epilepsy Res. 25, 51 – 57. Houser, C.R., Obenaus, A., 1994. Hippocampal neurons loss following pilocarpine-induced seizures in developing rats. Epilepsia 35 (Suppl. 8), 9. Hudson, L.P., Monoz, D.G., Miller, L., McLachlan, R.S., Girvin, J.P., Blume, W.T., 1993. Amygdaloid sclerosis in temporal lobe epilepsy. Ann. Neurol. 33, 622 – 631. Irwin, W., Davidson, R.J., Lowe, M.J., Mock, B.J., Sorenson, J.A., Turski, P.A., 1996. Human amygdala activation detected with echo-planar functional magnetic resonance imaging. NeuroReport 7, 1765 – 1769. Kapur, N., Barker, S., Burrows, E.H., Ellison, D., Brice, J., Illis, L.S., Scholey, K., Colbourn, C., Wilson, B., Loates, M., 1994. Herpes simplex encephalitis: long term magnetic resonance imaging and neuropsychological profile. J. Neurol. Neurosurg. Psychiatr. 57, 1334 – 1342. Ka¨lvia¨inen, R., Salmenpera¨, T., Partanen, K., Vainio, P., Riekkinen, P., Pitka¨nen, A., 1997. MRI volumetry and T2 relaxometry of the amygdala in newly diagnosed and chronic temporal lobe epilepsy. Epilepsy Res. 28, 39 – 50. LaBar, K.S., LeDoux, J.E., Spencer, D.D., Phelps, E., 1995. Impaired fear conditioning following unilateral temporal lobectomy in humans. J. Neurosci. 15, 6846 – 6855. Margerison, J.H., Corsellis, J.A.N., 1966. Epilepsy and the temporal lobes. A clinical, electroencephalographic and neuropathological study of the brain in epilepsy with particular reference to the temporal lobes. Brain 89, 499 – 530.

252

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253

Mascott, C.R., Gotman, J., Beaudet, A., 1994. Automated EEG monitoring in defining a chronic epilepsy model. Epilepsia 35, 895 – 902. McDonald, A.J., Pearson, J.C., 1989. Coexistence of GABA and peptide immunoreactivity in non-pyramidal neurons of the basolateral amygdala. Neurosci. Lett. 100, 499–530. Meinini, C., Meldrum, B.S., Riche, D., Silva-Comte, C., Stutzmann, J.M., 1980. Sustained limbic seizures induced by intra-amygdaloid kainic acid in the baboon: symptomatology and neuropathological consequences. Ann. Neurol. 8, 501 – 509. Meldrum, B.S., Brierley, J.B., 1973. Prolonged epileptic seizures in primates. Ischemic cell change and its relation to ictal physiological events. Arch. Neurol. 28, 10–17. Meyer, A., Beck, E., Shepherd, M., 1955. Unusually severe lesions in the brain following status epilepticus. J. Neurol. Neurosurg. Psychiat. 18, 24–33. Miller, L.A., McLachlan, R.S., Bouwer, M.S., Hudson, L.P., Munoz, D.G., 1994. Amygdalar sclerosis: preoperative indicators and outcome after temporal lobectomy. J. Neurol. Neurosurg. Psychiatr. 57, 1099–1105. Morris, J.S., Frith, C.D., Perret, D.I., Rowland, D., Young, A.W., Calder, A.J., Dolan, R.J., 1996. A differential neural response in the human amygdala to fearful and happy facial expressions. Nature 383, 812–815. Nissinen, J.T.P., Pitka¨nen, A., Koivisto, E., Halonen, T.O., 1996. Appearance of spontaneous seizures in a new experimental model of temporal lobe epilepsy in rats. Epilepsia 37 (Suppl. 5), 76. Norman, R.M., 1964. The neuropathology of status epilepticus. Med. Sci. Law 4, 46–51. Ounstedt, C., Lindsay, J., Norman, R., 1966. Biological factors in temporal lobe epilepsy. Clinics in Developmental Medicine, vol. 22. Medical Education and Information Unit. The Spastic Society (in association with William Heinemann Medical Books), Suffork, UK. Pare´, D., Gaudreau, H., 1996. Projection cells and interneurons of the lateral and basolateral amygdala: Distinct firing patterns and differential relation to theta and delta rhythms in conscious cats. J. Neurosci. 16, 3334–3350. Penfield, W., Jasper, H.H., 1954. Epilepsy and the functional anatomy of the human brain. Little Brown, Boston. Pitka¨nen, A., Amaral, D.G., 1991. Demonstration of projection from the lateral nucleus to the basal nucleus of the amygdala: a PHA-L study in the monkey. Exp. Brain Res. 83, 465 – 470. Pitka¨nen, A., Stefanacci, L., Farb, C., Go, G., LeDoux, J.E., Amaral, D.G., 1995. Intrinsic connections of the rat amygdaloid complex: projections originating in the lateral nucleus. J. Comp. Neurol. 356, 288–310. Pitka¨nen, A., Savander, V., LeDoux, J.E., 1997. Organization of intra-amygdaloid circuitries in the rat: an emerging framework for understanding functions of the amygdala. Trends Neurosci. 20, 517–523. Price, J.L., Russchen, F.T., Amaral, D.G., 1987. The Limbic Region. II. The Amygdaloid Complex. In: Bjo¨rklund, A., Ho¨kfelt, T., Swanson, L.W. (Eds.), Handbook of Chemical

Neuroanatomy, vol. 5, Integrated Systems of the CNS, part I. Elsevier, Amsterdam, pp. 279 – 388. Quesney, L.F., 1986. Clinical and EEG features of complex partial seizures of temporal lobe origin. Epilepsia 27 (Suppl. 2), S27 – S45. Rainnie, D.G., Asprodini, E.K., Shinnick-Gallagher, P., 1992. Kindling-induced long-lasting changes in synaptic transmission in the basolateral amygdala. J. Neurophysiol. 67, 443 – 454. Rauch, S.L., van der Kolk, B.A., Fisler, R.E., Alpert, N.M., Orr, S.P., Savage, C.R., 1996. A symptom provocation study of posttraumatic stress disorder using positron emission tomography and script-driven imagery. Arch. Gen. Psychiatr. 53, 380 – 387. Ribak, C.E., Seress, L., Weber, P., Bakay, R.A.E., 1995. Alumina gel injections into the amygdala of monkeys cause behavioral and pathological changes found in temporal lobe epilepsy. Soc. Neurosci. Abstr. 21, 1964. Rogan, M.T., LeDoux, J.E., 1996. Emotion: systems, cells, synaptic plasticity. Cell 85, 469 – 475. Sano, K., Malamud, N., 1953. Clinical significance of sclerosis of the Cornu Ammonis. Arch. Neurol. Psychiatr. (Chicago) 70, 40 – 53. Saukkonen, A., Ka¨lvia¨inen, R., Partanen, K., Vainio, P., Riekkinen, P., Pitka¨nen, A., 1994. Do seizures cause neuronal damage? A MRI study comparing the volumes of the hippocampus, amygdala and parahippocampal gyrus in newly diagnosed and chronic epilepsy. NeuroReport 6, 219 – 223. Savander, V., Miettinen, R., LeDoux, J.E., Pitka¨nen, A., 1997. Lateral nucleus of the rat amygdala is reciprocally connected with the basal and accessory basal nuclei: A light and electron microscopic study. Neuroscience 77, 767 – 781. Schwob, J.E., Fuller, T., Price, J.L., Olney, J.W., 1980. Widespread patterns of neuronal damage folowing systemic or intracerebral injections of kainic acid: a histological study. Neuroscience 5, 991 – 1014. Scott, S.K., Young, A.W., Calder, A.J., Hellawell, D.J., Aggleton, J.P., Johnson, M., 1997. Impaired auditory recognition of fear and anger following bilateral amygdala lesions. Nature 385, 254 – 257. Smith, B.N., Dudek, F.E., 1996. Enhanced population responses in the basolateral amygdala of kainate-treated, epileptic rats in vitro. Neurosci. Lett. 222, 1 – 4. So, N., Quesney, L.F., Jones-Gotman, M., Olivier, A., Andermann, F., 1989. Depth electrode investications in patients with bitemporal abnormalities. Ann. Neurol. 25, 423 – 431. Sorvari, H., Soininen, H., Palja¨rvi, L., Karkola, K., Pitka¨nen, A., 1995. Distribution of parvalbumin-immunoreactive cells and fibers in the human amygdaloid complex. J. Comp. Neurol. 360, 185 – 212. Sutula, T., He, X.X., Cavazos, J., Scott, G., 1988. Synaptic reorganization in the hippocampus induced by abnormal functional activity. Science 239, 1147 – 1150. Tuunanen, J., Halonen, T., Pitka¨nen, A., 1996. Seizures cause selective regional damage and loss of GABAergic neurons in the rat amygdaloid complex. Eur. J. Neurosci. 8, 2711 – 2725.

A. Pitka¨nen et al. / Epilepsy Research 32 (1998) 233–253 Tuunanen, J., Halonen, T., Pitka¨nen, A., 1997. Decrease in somatostatin-immunoreactive neurons in the rat amygdaloid complex in a kindling model of temporal lobe epilepsy. Epilepsy Res. 26, 315–327. Van Paesshen, W., Connelly, A., Johnson, C.L., Duncan, J.S., 1996. The amygdala and intractable temporal lobe epilepsy: a quantative magnetic resonance imaging study. Neurology (NY) 47, 1021–1031. Wasterlain, C., Baldwin, R., Itabashi, H., 1996. Status epilepticus induces widespread neuronal injury in newborn marmosets. Epilepsia 37 (Suppl. 5), 141.

253

White, L., Price, J.L., 1993. The functional anatomy of limbic status epilepticus in the rat. I. Patterns of [14C]2-deoxyglucose uptake and fos immunocytochemistry. J. Neurosci. 13, 4787 – 4809. Wilson, C.L., Isokawa, M., Babb, T.L., Crandall, P.H., 1990. Functional connections in the human temporal lobe. I. Analysis of limbic system pathways using neuronal responses evoked by electrical stimulation. Exp. Brain Res. 82, 279 – 292. Young, A.W., Aggleton, J.P., Hellawell, D.J., Johnson, M., Broks, P., Hanley, J.R., 1995. Face processing impairments after amygdalatomy. Brain 118, 15 – 24.

.

.