Delusions, illusions and hallucinations in epilepsy: 1. Elementary phenomena

Epilepsy Research (2009) 85, 162—171

journal homepage: www.elsevier.com/locate/epilepsyres

REVIEW

Delusions, illusions and hallucinations in epilepsy: 1. Elementary phenomena Brent Elliott a, Eileen Joyce b, Simon Shorvon b,∗ a b

National Hospital for Neurology and Neurosurgery, United Kingdom UCL Institute of Neurology, University College London, United Kingdom

Received 31 October 2008; received in revised form 8 March 2009; accepted 15 March 2009 Available online 6 May 2009

KEYWORDS Epilepsy; Hallucinations; Delusions; Illusions; SEEG; Cerebral localisation

Summary The purpose of this paper and its pair is to provide a comprehensive review, from the different perspectives of neurology and neuropsychiatry, of the phenomenology and mechanisms of hallucinatory experience in epilepsy. We emphasise the clinical and electrophysiological features, and make comparisons with the primary psychoses. In this paper, we consider definitions and elementary hallucinatory phenomena. Regarding definition, there is a clearly divergent evolution in meaning of the terms delusion, illusion and hallucination in the separate traditions of neurology and psychiatry. Psychiatry makes clear distinctions between the terms and has focussed on the empirical use of descriptive psychopathology in order to delineate the various psychiatric syndromes, including those in epilepsy. These distinctions in psychiatry have stood the test of time and are useful in clinical descriptive terms, but do not help to understand the basic mechanisms. The focus of neurology has been to regard delusions, illusions and hallucinations in epilepsy as a result of localised or network based neuronal epileptic activity that can be investigated especially using intracranial stereoelectroencephalography (SEEG). The neurological approach leads to a more synoptical definition of ‘hallucination’ than in psychiatry and to the conclusion that there is little point in differentiating hallucination from illusion or delusion in view of the overlap in the physiological bases of the phenomena. The semiologically derived differentiation of these terms in psychiatry is not supported by similarly discrete electrophysiological signatures. However, as discussed in the second paper, some psychotic states are associated with similar electrophysiological changes. The wide range of hallucinatory symptoms occurring during epileptic seizures recorded during intracranial SEEG and brain stimulation are reviewed here, including: experiential and interpretive phenomena, affective symptoms, as well as auditory, olfactory, gustatory, somatic and visual hallucinatory phenomena. Several conclusions can be drawn. First, it is clear that there is only limited anatomical specificity of many hallucinatory states. Repeated seizures or stimulation of a single area, even within the same patient can produce different psychic responses, whilst stimulation of

∗

Corresponding author at: Department of Clinical and Experimental Epilepsy, National Hospital for Neurology and Neurosurgery, Queen Square, London, WC1N 3BG, United Kingdom. Tel.: +44 845 155 5000x4194; fax: +44 207 676 2155. E-mail address: [email protected] (S. Shorvon). 0920-1211/$ — see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.eplepsyres.2009.03.018

Delusions, illusions and hallucinations in epilepsy

163

widely distinct areas (especially in the limbic system) within the same individual can produce remarkably similar phenomena. This lack of specificity applies particularly to psychic symptoms, including experiential phenomena, and complex hallucinatory states. The most anatomically specific areas from this point of view are the elementary hallucinations arising from primary visual and auditory cortices. Involvement of the limbic cortex is a pre-requisite for the occurrence of complex hallucinatory states. It is clear that on the basis of these findings, as has been recognised at least since the 1960s, that even apparently focal epileptic seizures, (especially in the mesial temporal lobe, insula and limbic cortices), must involve widely distributed neuronal networks. © 2009 Elsevier B.V. All rights reserved.

Contents Definitions ............................................................................................................... Experiential mental phenomena evoked by brain stimulation............................................................. SEEG correlates of hallucinatory symptoms recorded during seizures in patients with epilepsy ........................... Discussion ................................................................................................................ References ...............................................................................................................

‘If sensitive nerves are enough to make a poet I should be worth more than Shakespeare and Homer. . . I who have heard through closed doors people talking in low tones thirty paces away, across whose abdomen one may see the viscera throbbing, and who have sometimes felt in the space of a minute a million thoughts, images and combinations of all kinds throwing themselves into my brain at once, as it were the lighted squibs of fireworks.’ Correspondence of Flaubert, (Lombroso 1891). The term ‘lighted squibs of fireworks’ is a pertinent way of describing the disordered bursts of energy which underpin at least some of the hallucinatory experiences seen in epilepsy. Historically, much of the research in this area derives from the field of psychiatry, with its emphasis on the empirical use of descriptive psychopathology in order to delineate the various psychiatric syndromes and determine to what extent the psychopathology (i.e., abnormalities of affect, thought, and perception) seen in epilepsy is similar to, or differs from that seen in the major mental disorders, for example schizophrenia. By contrast, there has been an almost entirely separate tradition within neurology. The central theorem of this school is that hallucinations occur as a consequence of the activation of a localised group of neurones which can be investigated by cerebral recording and cerebral stimulation, and the ‘Gold Standard’ investigation has been intracranial stereoelectroencephalography (SEEG). Over the past 50 years this application has greatly clarified some of the details regarding the anatomical basis and physiological mechanisms underpinning hallucinations in epilepsy. Throughout this period psychiatry has tended to echo the view of Hughlings Jackson that compound mental states ‘cannot be owing to an epileptic discharge’ (Jackson, 1958), however, whilst it is true that complex psychic states as seen in the ‘functional’ psychoses have less often been correlated with abnormalities on SEEG, as Trimble (1991) points out ‘specific Schneiderian phenomena have not been recorded but in all probability have not been examined for.’ There is now a considerable body of SEEG evidence which calls this

164 164 166 170 170

view into question and it appears, in at least some cases, that complex psychotic symptoms are directly due to the effects of non-convulsive epileptic activity (limbic status) or to the indirect after-effects of chronic epileptic discharges. The purpose of this set of two papers is to provide a comprehensive review of the phenomenonology of hallucinations in epilepsy from a specifically neurological and neuropsychiatric standpoint, and to draw distinctions between this approach and that of clinical psychiatry. The term ‘hallucination’ is here taken to encompass a range of phenomena giving it a very different usage to that seen in psychiatry. Based as it is on electrophysiological findings, our main thrust is to present SEEG evidence which demonstrates that hallucinations often have an electrophysiological basis, usually involving widely disseminated limbic structures. In this first paper, we begin with a brief review of definitions and of the different meanings assigned to key terms such as ‘hallucination’ and ‘psychosis’ that have emerged as a result of the divergent evolution of very separate neurological and psychiatric traditions. We then consider the elementary hallucinatory states which are observed during brain stimulation, and during spontaneous brief epileptic seizures. In the second paper we consider the more complex and prolonged states associated with complex partial status epilepticus, postictal and interictal psychosis. The similarity of these latter states to the primary psychoses raises interesting questions about the pathophysiology of psychosis. It is important to recognise that SEEG does have a major limitation, namely sampling bias, which we would like to mention at the outset. The electrodes record activity from only a very small area in the vicinity of the electrode (measured in millimetres), and activity in areas beyond this will be overlooked. This is important as there are many indications that limbic functional (and dysfunctional) activity is often a widely distributed network phenomenon involving disparate interconnected neuronal areas. Gloor for instance, has shown repeatedly that the concept of a small discrete limbic epileptic focus is inaccurate and that limbic seizures (even with focal pathology, such as hippocampal sclerosis) may involve simultaneously a wide network of neuronal activation, not necessarily even contiguous. This

164 Table 1

B. Elliott et al. Summary of the conventional definitions as used commonly in psychiatry.

Hallucination (David, 2004)

Illusion

Delusion

‘A sensory experience which occurs in the absence of corresponding external stimulation of the relevant sensory organ, has sufficient sense of reality resemble a veridical perception, over which the subject does not feel direct and voluntary control, and which occurs in the awake state’. These are false perceptions of a real external stimulus, for example a change in shape, size, colour or texture. In some cases, where the external stimulus is minimal, the differentiation nosologically from hallucination can be difficult, although illusions carry different aetiological and diagnostic implications. Delusions are abnormalities of thought rather than perception (although they may develop from the latter) and may be defined as ‘fixed false beliefs, strongly held and immutable in the face of refuting evidence, that are not consonant with the person’s education, social and cultural background.’ (Sadock and Sadock, 2000). As with hallucination, this term has an interesting history (Berrios and Dening, 1996) and its exact meaning and usage have evolved continuously, reflecting trends in psychology. Delusional themes commonly include: guilt, worthlessness, ill-health, persecution, reference, grandeur, love, jealousy, poverty, infestation, and religion. A range of beliefs are also recognised that lie somewhere between the delusional and non-delusional, these ‘overvalued ideas’ are often best considered as sustained and unreasonable pre-occupations, the unlikely validity of which the holder has little insight into.

same limitation almost certainly applies to the production of complex psychological symptoms. Since the scalp EEG is often normal despite widespread activity on SEEG, we regard this investigation as an unreliable indicator of ongoing ictal activity and it is not considered any further. This series of papers provides no definitive answers to questions about mechanisms of hallucinatory or psychotic states, nor do we that all psychiatric states have the same physiological basis, but we intend our review to provide a framework for future investigation into the potential physiological causes of psychosis and more clearly to define the overlap that is undoubtedly present between epilepsy and psychosis.

Definitions The thesis within neurology that delusions, illusions and hallucinations in epilepsy are a symptom of localised or network-based neuronal spike or spike-wave activity points to a fundamental distinction between the approaches of neurology and psychiatry. An exemplar of this problem can be viewed in the classic description given by Williams (1959) of an hallucination: ‘by popular usage an hallucination, which is a percept without a stimulus, may be organic or psychotic. An organic hallucination occurs when through brain disease the patient has a percept without stimulus, into the nature of which he usually has insight. Local disturbance of the brain has evoked a perceptual response. . . But percepts may arise within the body—–be proprioceptive as well as exteroceptive, and there can be no fundamental difference between sensations felt in a limb or through the eye as a result of a local epileptic discharge. . . For the physiological changes in the cerebrum responsible for both are similar. They can all be considered organic hallucinations, caused by the epileptic discharge. . . The feelings called fear, depression or pleasure arising in the attack [also] have no local reference, in other regards they are similar, and for physiological purposes can be considered to be organic hallucinations.’

From the psychiatric perspective, the definition of hallucination is broadly similar (Table 1), but the usage is quite different. Whilst most psychiatrists would consider the patient experiencing complex auditory hallucinations or persecutory delusions to be ‘psychotic’, probably none would extend this term to include those experiencing ictal affects such as fear, depression, pleasure, deja vu, rage or indeed an epigastric rising sensation. The electrophysiological basis of these phenomena are not considered due to the historical emphasis of descriptive psychopathologists on providing a system of empirically derived, phenomenologically based categories, without reference to causation. The meticulous but empirical categorization of hallucinatory symptoms in psychiatry has served the discipline well, but does not advance an understanding of the cerebral mechanisms. We hope that the neurological approach, embedded in physiology, will illuminate at least some of these interesting and challenging clinical hallucinatory phenomena. Distinction between these diverse categories in epilepsy is not possible at the electrophysiological level. Repeated seizures or stimulation of a single area, even within the same patient can produce different psychic responses (Baldwin, 1960; Weingarten et al., 1976; Horowitz et al., 1968) whilst stimulation of widely distinct areas within the same individual can produce remarkably similar phenomena (Horowitz et al., 1968; Penfield and Perot, 1963; Fish et al., 1993). Yet psychiatry must remain reliant on the distinction between these phenomena in order to produce reliable diagnoses and to advise on treatment and prognosis as well as to manage risk. This semiological approach need not detract from an appreciation of the electrographic basis, thus, in the rest of this paper, we use the term hallucination broadly, to refer to all these phenomena.

Experiential mental phenomena evoked by brain stimulation The first major study of auras (a simple partial seizure occuring within seconds before a complex partial or secondarily

Delusions, illusions and hallucinations in epilepsy generalized tonic-clonic seizure) was that of Gowers (1901) who reported psychical auras in 4.6% of over 2000 cases. By 1938 Penfield had discovered that such mental phenomena could also be reproduced by electrical stimulation of the temporal lobe in epileptic patients during surgical procedures performed for intractable epilepsy (Penfield, 1938). His particular interest was in the evocation of memory and he subsequently made a distinction between what he called ‘experiential’ and ‘interpretive’ mental phenomena. These have received remarkably little attention from subsequent researchers and are therefore briefly reviewed here. Experiential phenomena were said to represent mental events from the patient’s personal past; they may be especially vivid and combine elements of perception, memory and affect in a unified subjective experience (Gloor et al., 1982). Interpretive phenomena had to do with the present circumstances of the patient and included illusions and emotions. Gloor (1990) summarises the key features of ‘experiential responses’ as follows: (a) there may be a vivid or intrusive recall of a past event; (b) there is a feeling of familiarity or reminiscence (déjà vu, deja vécu); (c) the characteristic sensation of dreaminess; (d) the patient is said to be always aware of the incongruity and illusory nature of the experience; (e) affective states such as fear, sadness, guilt, anger or sexual excitement are common; (f) these responses typically lack certain features such as forward motion in time (with the exception possibly of musical hallucinations) and scenes do not evolve; and (g) auditory hallucinations are said to be almost entirely without semantic content (i.e., they lack coherent meaning). Penfield and Jasper (1954) subdivided psychical seizures into four groups: (1) illusions, (2) emotions, (3) hallucinations and (4) forced thinking. For Penfield, who was essentially a localisationist, hallucinations were in fact memory images with the particular memory evoked depending on which engram happened to be closest to the site of stimulating electrode. A key element of Penfield’s later thinking was the idea that experiential responses occur virtually only in seizures arising from the lateral temporal isocortex, where the stimulus was strong enough to provoke an after-discharge (in stimulation-induced seizures an ‘after-discharge’ is the term used to describe the continuation of neuronal activity which outlasts the stimulation train. This may indicate spread to other structures; Penfield and Perot, 1963). The localisation to the lateral cortex has been shown to be incorrect, and we go into further detail here, to emphasise the fundamental point that much of the complex hallucinatory experience in epilepsy cannot be well localised — this ‘phrenological’ conceptualisation is not backed up by the electrophysiological data — although as we will see below the more elementary the phenomenon the more localised it tends to be. Halgren et al. (1978) reported the results of 3495 stimulations of the medial temporal lobe of 36 patients with psychomotor epilepsy. Just 267 (7.6%) elicited a mental response which included hallucinations of complete scenes, déjà vu, anxiety and visceral sensations. Like Penfield they found that the presence of an after-discharge was necessary but not sufficient for an experiential response to occur, whilst also suggesting that mental phenomena evoked by medial temporal lobe stimulation were idiosyncratic and related to the personality of the patient. Others have taken

165 Table 2 Experiential illusions and hallucinations observed with stereotaxic exploration of the temporal lobes. Experience Visual illusions Elementary visual hallucinations (phosphenes) Complex visual hallucinations Elementary auditory hallucinations Complex auditory hallucinations Olfactory hallucinations Familiarity (déjà vu) Unfamiliarity (jamais vu) Forced thinking Fear Anger Irritation Emotional distress (depression, guilt etc.) Thirst Feeling of bodily distortion

No. of observations

No. of patients

9 15a

3 3

18 0

5 0

3

2

2 23 0 10 >49 1b >3 6c

1 4 0 2 7 1 1 3

10 2

2 1

Adapted from Gloor et al. (1982). a All induced by electrical stimulation. b Angry mood and facial expression (no aggression). c In one instance may have been caused by strong nausea.

up this point and attempted to integrate personality factors by arguing that hallucinations are symbolically related to ongoing psychodynamic processes (Mahl et al., 1964; Rayport and Ferguson, 1974; Ferguson et al., 1969). These factors were not addressed in Penfield’s earlier studies, nor were attempts made to check the veracity of patients evoked ‘memories’ (Trimble, 1991). The anatomical basis was revised further by Gloor et al. (1982) who imputed a key role for limbic structures rather than temporal neocortex. In his study he attempted to reproduce fragments of experiential seizure phenomena by electrical stimulation of intra-cerebral depth electrodes in 35 patients with medically intractable epilepsy. Psychosensory symptoms were elicited in 18 (52%) and had usually occurred as part of the patients previous seizure experience. Stimulation of limbic structures produced experiential phenomena far more frequently than stimulation of temporal neocortex or white matter with the amygdala producing the most responses (n = 44), followed by hippocampus (n = 26) and parahippocampal gyrus (n = 12). Even complex visual hallucinations were reported with limbic stimulation alone, whilst stimulation of temporal neocortex and white matter resulted in only five elementary visual hallucinations which were most likely caused by stimulation of optic radiation fibres. In only two instances was the temporal neocortex alone involved in the production of a response. The seizure phenomena elicited are outlined above in Table 2. More recent work (Gloor, 1990; Fish et al., 1993) continues to cast doubt on Penfield’s original emphasis on the role of the temporal neocortex, and indeed on the concept that the symptoms have a well-localised basis. Fish et

166 al. (1993) studied the clinical responses obtained by electrical stimulation of the temporal and frontal lobes in 75 patients undergoing pre-surgical evaluation using chronic intra-cerebral EEG recording. Responses were subdivided into: (1) complex perceptual illusions and hallucinations; (2) mnemonic phenomena (flashbacks of personal memories and déjà vu) and (3) affective responses. Experiential phenomena (without an after-discharge or one limited to the stimulation site) were elicited in 20 patients. Auditory hallucinations occurred in two (in both cases combined with visual hallucinations), visual hallucinations and illusions in eight, fear also in eight and affective responses other than fear in four. In common with previous findings stimulation of the amygdala produced the most responses, followed less often by the hippocampus and only rarely from the temporal neocortex. The same phenomenon could be elicited by stimulation of widely varying areas even within the same patient, and the habitual auras could in some cases be elicited by stimulation of different areas. The authors also conclude that unless limbic structures are activated experiential phenomena (including emotional responses, illusions of familiarity and both complex auditory and visual hallucinations) do not occur. They speculate that limbic activation by means of a seizure discharge may add an affective dimension to perceptual data processed by the temporal neocortex which ‘may be required for endowing them with emotional immediacy.’ Vignall et al. (2007) provided an elegant SEEG study of the ‘dreamy state’ as originally described by Jackson. 40 stimulations were carried out in 16 subjects and 15 seizures in 5 subjects. A total of 15 sensations of déjà vecu, 35 visual hallucinations and 5 feelings of strangeness were recorded. 45% of the dreamy states were evoked by stimulation of the amygdale, 37.5% of the hippocampus and 17.5% of the parahippocampal gyrus. There was no involvement of the temporal neocortex in any of the dreamy states recorded in spontaneous seizures or by stimulation and indeed the authors considered that spread to the lateral cortex inhibited the state. It was also thought that the déjà vecu and visual hallucinations were part of a clinical continuum consisting of a memory relived by the patient, and that the study demonstrated the existence of large neural networks underpinning memory recall which could be activated via the hippocampus, amygdala and rhinal cortex. Gloor (1990) evolved what he called the ‘matrix theory’ to try to explain these findings. He considered that the experiential phenomena in temporal lobe epilepsy were not the result of loss of inhibitory control, as Hughlings Jackson (Jackson, 1958) and Halgren et al. (1978) believed, but were the result of positive activation of limbic structures. He hypothesised that: ‘evocation of an experience depends on the formation of a specifically patterned matrix made up of excited and inhibited neurons dispersed within widely distributed neuronal populations in large areas of isocortex and of the limbic system. Such a matrix which encodes its perceptual, mnemonic and affective information can be thought of as ‘representing’ an experience and to be specific for it. What carries the specific information is not the activity of any single cell within this population, but the specific pattern of connectivity woven between the neurons which creates a distributed matrix of excitation and inhibition and is specific for the experience it ‘represents’. . . [He proposed that] an epileptic discharge within the temporal lobe at the

B. Elliott et al. onset of a seizure, when it has not yet become too diffuse or too intense, may be able to recreate a specific matrix that may be similar or identical to that normally encoding a natural experience. Repeated discharge through mechanisms of synaptic plasticity known to be affected by epileptic discharge may have strengthened the interconnectivity of the neurons constituting such a matrix.’ This hypothesis would require the requisite structures to be reciprocally interconnected and this would appear to be the case (e.g., Van Hoesen, 1982). If these phenomena do arise from activation of matrices in distributed neuronal networks, then they could presumably be elicited from different areas of the temporal lobe including temporal isocortex, thus allowing reconciliation with Penfield’s earlier observations. Similarly, it would not be surprising that stimulation of the limbic ‘receiving end’ could elicit responses that reflect functions of the areas from which these connections originate. It is also clear that these ‘experiential’ states differ markedly from the complex psychotic symptoms seen in the functional psychoses or in the psychoses of epilepsy, although it is not clear to what extent this sort of specific psychopathology was examined for. However, the key themes that do emerge are that complex hallucinations and a range of abnormal affective states do occur in response to stimulation of predominantly limbic structures with relatively limited localisational specificity, at least compared to primary neocortex.

SEEG correlates of hallucinatory symptoms recorded during seizures in patients with epilepsy Two vital technical developments underpin much of the work cited in this review. These are the ability to record EEG activity from deep structures (particularly deep limbic structures) via SEEG and the ability to make long-term recordings via EEG telemetry. These developments permit the study of ictal and interictal phenomena. The data reduction and storage needed, require advanced computing and electrode technologies, as well as atlases for electrode placement. The first studies were conducted in the 1950s and the technologies are now used on a worldwide basis. These data provide the most convincing evidence concerning the anatomical bases of hallucinatory phenomena; here we review some examples from what is a large literature. (a) Affective and related psychological symptoms The studies of Penfield, Gloor and others are mentioned above. Other groups have elaborated upon this work. Wieser and colleagues have provided a number of illuminating and very detailed SEEG studies in patients being evaluated for epilepsy surgery. At the core of this work lies a detailed review of 213 seizures in 29 patients (Wieser, 1983), using arrays of recording electrodes typically placed in the hippocampus, parahippocampal gyrus, amygdala and cingulate gyrus. A wide variety of psychological symptoms have been recorded during periods of ongoing epileptic seizure activity (as recorded by these deep electrodes), particularly if this activity was prolonged, therefore amounting

Delusions, illusions and hallucinations in epilepsy

167

Figure 1 Long lasting clonic discharge in the left periamygdalar region during a rage attack (Wieser, 1983).

to limbic status epilepticus. Five examples are provided below: An intracranial EEG recorded during a rage attack is shown in Fig. 1 and another recorded during a laughing fit can be seen in Fig. 2, both showing left periamygdala discharges. SEEG recordings in a further patient who demonstrated markedly fluctuating symptoms (comprising fugue states, visual distortions, visual hallucinations, anxiety and visceral sensations) correlated with deep, predominantly right temporal (lingual gyrus, gyrus of Heschl, hippocampus and amygdala) discharges (Wieser et al., 1985). Wieser (1980) reports a fourth case of a 22-year-old female in whom fifty-three seizures were recorded of which forty-nine were accompanied by clinical signs in various forms: (1) right sided hippocampal discharge was associated with feelings of anxiety and visceral sensations which were more pronounced if the anterior temporal lobe and in particular the amygdala

Figure 2 Laughing fit accompanied by rhythmic discharges in left periamygdalar region (electrode 1) slowing down to 0.5 s complexes. (Wieser, 1983).

were the site of seizure discharge, (2) more elaborate symptoms were associated with spread of the discharge to lateral temporal cortex so modifying activity in the lingual gyrus when visual experiences were prominent and to the right gyrus of Heschl during musical hallucinations. The final example is an episode captured on SEEG which Wieser calls a ‘visual delusion’. During a spontaneous epileptic discharge from the temporo—occipital cortex the patient believed for a short time that her medicine cabinet was in fact her neighbours stove (Fig. 3; Wieser, 1983). (b) Auditory Ictal auditory hallucinations have a particularly specific localisation to discharges in or near Heschl’s

Figure 3 ‘Visual delusion’ during a temporo-occipital discharge (most prominent in 4/8—9, also visible in 5/1—2). The insert at the bottom shows enlarged portions of the record between the horizontal bars. On the right is a brain map with the electrode positioning for this patient. (Wieser, 1983).

168

B. Elliott et al.

Table 3 Distribution of the main subjective symptoms at seizure onset according to electrophysiologic subtype (Maillard et al. (2004)). n = number in each subgroup; (%) percentage of those in subgroup experiencing each hallucinatory type. Ictal features Auditory hallucination or illusion Visual hallucination or illusion Sensory hallucination or illusion (visual, auditory or vestibular) Gustatory hallucination Fear Viscerosensory symptoms a

Medial n = 24

Medial—lateral n = 18

Lateral n = 13

Degree of significance

1 (4.2) 2 (8.3) 3 (12.5)

2 (11.1) 3 (16.7) 6 (33.3)

6 (46.2) 4 (30.8) 11 (84.6)

P = 0.005a P = 0.21 P < 0.0001a

2 (8.3) 9 (37.5) 19 (79.2)

1 (5.6) 4 (22.2) 9 (50)

0 0 3 (23.1)

P = 0.78 P = 0.026a P = 0.004a

Significant.

gyrus and the auditory association areas (Wieser, 1980, 1983). Maillard et al. (2004) using stereoelectroencephalography (SEEG), analysed 187 spontaneous seizures recorded from 55 patients with medically intractable TLE. Patients were classified into medial (M; n = 24), lateral (L; n = 13) and medio-lateral (ML; n = 18) groups on the basis of the electrophysiological findings. Viscerosensory symptoms, fear and the dreamy state were significantly more frequent in the M and ML groups. Whereas only auditory hallucinations and illusions significantly differentiated the L from the M or ML groups (46.2% in L vs. 4.2% and 11.1% in M and ML respectively). The results are summarised in Table 3 and are consistent with the localising value of these symptoms namely Heschl’s gyrus for primary auditory hallucinations, more extended and lateral parts of the superior temporal gyrus for complex auditory hallucinations and basal—temporal gyri and temporooccipital junction for complex visual hallucinations (Bancaud and Talairach, 1993). Auditory auras appear to have a similar localisation, and possibly even lateralisation. Clarke et al. (2003) describe the phenomenon of ‘ear-plugging’ in localisation-related epilepsy. In their series, three children who demonstrated unilateral or bilateral ear plugging (i.e., placing their hands over an ear in what is probably an attempt to block out an auditory hallucination) at the onset of partial seizures were investigated with scalp video electroencephalography (VEEG), magnetoencephalography (MEG) and MRI. All three plugged their ears during auditory hallucinations which were localised to the superior temporal gyrus, and appeared to lateralise the site of seizure onset to the contralateral temporal lobe. Mohamed et al. (2006) later investigated auditory hallucinations in 6 children using MEG. Three patients had elementary hallucinations, one had a complex hallucination and two had both complex and elementary hallucinations. Sounds heard included stampeding elephants, unbearable and buzzing sounds as well as rushing water. Two patients complained of amplification of sound whilst another at age eleven heard friend’s voices talking about him. All 6 patients demonstrated clustered MEG spike sources in the superior temporal gyrus; two had scattered spikes in the superior temporal gyrus as well as clustered MEG spike sources in the left inferior and middle frontal gyri or parieto-occipital region.

It is also interesting to observe that hallucinations of language or speech can arise from both dominant and non-dominant epileptic foci. (c) Olfactory Although every medical student associates TLE with ‘olfactory hallucinations’, they are in fact relatively infrequent in epilepsy. Chen et al. (2003) examined case-notes of 217 patients who underwent temporal lobectomy for medically intractable TLE. Just twelve (5.5%) reported olfactory auras, with lesions affecting mesial temporal structures in eleven. Diagnosis on histology included gliosis (n = 7, 58.3%), neoplasm (n = 4, 33.3%) and a single arteriovenous malformation (8.3%). This supports the view that olfactory auras are rare in TLE as well as the finding by Acharya et al. (1998) that tumour is the most likely aetiology. The only SEEG study we were able to identify was that quoted by Wieser (1983) in combination with gustatory hallucinations—–see below. (d) Gustatory Gustatory hallucinations are also considered rare in epilepsy. Gowers (1909) found only one case in a series of 1102 patients. Gibbs et al. (1948) and Currie et al. (1971) reported prevalence rates of just 1.3% and 3% respectively and the most recent series by Hausser-Hauw and Bancaud (1987) reported a prevalence rate of 4%. In the latter study gustatory hallucinations were investigated using SEEG and found to occur in 10 patients (50%) during temporal lobe seizures, in six (30%) during parietal seizures and in four (20%) during parietotemporal seizures. Isolated brief gustatory hallucinations could be elicited from stimulation of: the right rolandic operculum, parietal operculum, amygdala, hippocampus, medial temporal gyrus and the anterior part of the right temporal gyrus. Seizures with gustatory manifestations could be induced by electrical stimulation of the right hippocampus and amygdala and left hippocampus and indeed Wieser (1983) gives an example of left hippocampal status recorded on SEEG resulting in both olfactory and gustatory hallucinations. (e) Somatic Baldeweg et al. (1998) using SEEG demonstrated an association between somatic hallucinations and activity in the post-central gyrus, parietal operculum, insula and inferior parietal lobule. The insula is a key relay station between frontotemporal cortical areas and limbic regions. Penfield was the first investigator to underline

Delusions, illusions and hallucinations in epilepsy the similarity between symptoms observed during temporal lobe seizures and those evoked by insular cortex stimulation (Penfield and Jasper, 1954). His early work noted that focal insula lobe seizures produced abdominal sensations and gastro-intestinal movement, findings substantiated by later electrical stimulation experiments under local anaesthesia (Penfield and Rasmussen, 1950; Penfield and Kristiansen, 1951; Penfield and Jasper, 1954). In a later study, Penfield and Faulk (1955) stimulated the insula cortex of 6 patients undergoing craniotomy for focal epilepsy. In four, stimulation produced effects on the stomach which varied from inhibited gastric motility, to the production of violent activity with a marked increase in tone. The issue of whether abdominal and epigastric sensations were represented in this region was left an ‘open question’, with many of the sensory responses thought be secondary to motor changes in the gastro-intestinal tract. Despite these early findings, the role of the insula in TLE has remained more or less unexplored, largely due to its relative inaccessibility to depth EEG recording. Recent improvements in SEEG technology such as smaller electrodes and greater accuracy of localisation using MRI, now permit chronic insular cortex recordings to be taken. Isnard and Mauguiere (2005) provide a description of the clinical features of insular lobe seizures in 50 patients. These began as simple partial seizures occurring in full consciousness, followed by a sensation of laryngeal constriction, parasthesia, dysarthric speech and/or elementary auditory hallucinations. Other authors however, have shown that the symptomatology of insula seizures may reflect widespread propagation or distribution and argue that the symptomatology may not be very anatomically specific, with Ryvlin et al. (2006), for example, describing nocturnal hypermotor seizures suggestive of frontal lobe epilepsy arising in the insular cortex, and a similar case was reported by Nguyen et al. (2008) in which insular seizures were diagnosed by SEEG and insular resection resulted in seizure freedom. (f) Visual hallucinations and illusions Ictal elementary visual phenomena are common in occipital seizures with estimated prevalence rates ranging from 8% (Marques-Assis et al., 1971) to 72% (Salanova et al., 1992). In a review by Taylor et al. (2003), elementary hallucinations were divided into positive (simply shaped flashes of colour or light, phosphenes) and negative (scotoma, hemianopia, amaurosis) manifestations. Simple illusions may also occur where objects may appear to change in size (macropsia and micropsia), shape (metamorphopsia), or lose colour (achromatopsia). Panayiotopoulos (1999) provides a qualitative analysis of ictal visual symptoms in 9 patients with idiopathic occipital epilepsy with visual hallucinations (IOEVH). He found that elementary ictal visual hallucinations comprised mostly multiple bright coloured spots, circles, or balls lasting for between 5 and 30 s, although in one patient up to 10 min. Their onset was usually unilateral, appearing in the temporal hemifields and then moving horizontally to the contralateral side. Whilst they may multiply, flash and change in size, occipital paroxysms were only present on the scalp EEG in three.

169 Palmini et al. (1993) investigated 8 patients, in whom it was unclear whether the seizure originated in the occipital or temporal lobes, with depth electrodes. One suffered sudden loss of vision as the aura, another reported colours of discrete objects blurring together, and a third described blurred vision affecting the entire visual field. All three of these had an occipital focus on SEEG, and in all cases the seizure propagated to the ipsilateral mesial temporal lobe and neocortex. Right occipital lobe supra- and infra-calcarine structures were involved in 93% of ictal onset in the first patient. Seizures began mesially in right infra-calcarine region in the second, and in the third, 80% originated on the mesial surface of the left supra- and infra-calcarine regions. In the fourth patient who reported loss of vision and a sense of freezing in the left eye, 75% of the seizures at onset occurred simultaneously in the right mesial and lateral occipital lobe as well as in ipsilateral temporal neocortex and hippocampus. Intra-cerebral recordings in proven occipital lobe epilepsy show rapid spread to posterior temporal, parietal and frontal regions (Taylor et al., 2003; Bancaud, 1969; Takeda et al., 1970). Such cases emphasise the observation that seizures in primary cortex may be quite localised, in contrast to the widely distributed seizures of association or limbic cortex, and propagation through the corpus-callosum to the opposite occipital lobe in occipital lobe epilepsy is often a late finding in adults (Williamson and Spencer, 1986). Involvement of the occipito-temporal cortex renders the hallucinations more complex and colourful. Ictal complex visual hallucinations tend to last a few seconds to minutes with the patient retaining insight into the unreality of the experience (Taylor et al., 2003). There is a reported high association with parieto-occipital lesions (Lance and Smee, 1989) but lesions of the temporal lobe have also been identified (David et al., 1945). Palinopsia is a rather characteristic epileptic feature, in which images persist or reduplicate, and has been localised to the right posterior cerebral region. In the rare phenomena of autoscopia, subjects perceive mirror images of themselves of normal size, shape and density, they may be in situations from their past or performing complex tasks. This may arise from seizures affecting the occipital-temporal junction zone (Sveinbjornsdottir and Duncan, 1993; Dewhurst and Pearson, 1955; Ionasescu, 1960). Complex visual hallucinations have a much more diffuse anatomical basis than simple hallucinations. In the study of Panayiotopoulos (1999) of idiopathic occipital epilepsy for example, only one patient experienced a complex visual hallucination, whereas Palmini et al. (1993) report that their patient with coloured flashes affecting particularly the left upper field and associated with unmotivated fear, nausea and formed visual and auditory hallucinations had a seizure onset in the right temporal lobe in all cases. Several other examples have already been cited above including the findings of Gloor (1982) of complex visual hallucinations resulting from limbic stimulation alone, and the findings of Fish et al. (1993) that unless limbic structures are activated complex visual hallucinations do not occur. It would seem that as with more complex auditory hallucinations, a

170

B. Elliott et al. more diffuse network of activation is required for their aetiogenesis and the involvement of limbic structures is key.

Discussion Here we have adopted the neurological definition of ‘hallucination’ which includes affects such as fear, depression, anger and déjà vu within its rubric as well as including the more traditional forms of hallucination described by psychiatrists. What does emerge from the above review is that, in epilepsy, any distinction between the various abnormalities of affect, thought (delusions) and perception (illusions and hallucinations) is not mirrored by differences at the anatomical and electrophysiological level. Electrical stimulation of the same area within the limbic system can produce symptoms belonging to each or all of these categories, whereas stimulation of widely segregated areas can produce a remarkably similar mental response. This lack of strict anatomical specificity within the limbic system for the various types of hallucinatory phenomenon induced by electrical stimulation or electrographic seizure activity is striking, and can be explained only by invoking theories of widely distributed neuronal networks. There is however, greater anatomical specificity when one considers elementary hallucinatory phenomena induced by stimulation of primary sensory areas (visual, auditory, other sensory and somatic), this is again mirrored by the often well-localised phenomenology of seizures arising from these areas (viz. auditory hallucinations to Heschl’s gyrus and visual illusions to the temporal and occipital neocortex). Having said this, visual symptomatology is in itself a poor anatomical discriminator, especially if complex, and can occur in seizures apparently with origin in occipital, temporal (mesial and neocortical), other limbic and parietal cortex. A wide variety of psychological symptoms (including hallucinations in most sensory modalities) have now been recorded during periods of ongoing epileptic seizure activity using intracranial EEG techniques. As Gloor suggested, involvement of limbic structures once again appears to be key in complex hallucinatory symptomatology, and it is perhaps not surprising that stimulation of the limbic ‘receiving end’ may elicit a range of responses that reflect the function of areas from which these cortico-limbic connections originate. This finding is of particular interest when we come to consider the electrophysiological correlates of complex psychotic states in the second paper, where the implication of these observations for understanding mechanisms underpinning psychiatric states is explored. Whilst it is clear that the short lived hallucinations described in this paper bear little resemblance to those encountered in the functional psychoses, limbic seizure activity, particularly when prolonged, (i.e., amounting to complex partial status epilepticus), can induce complex hallucinatory states which are remarkably similar to those observed in disorders such as schizophrenia and it is to these that we turn our attention in the second paper.

References Acharya, V., Acharya, J., Luders, H., 1998. Olfactory epileptic auras. Neurology 51 (1), 56—61. Baldeweg, T., Spence, S., Hirsch, S.R., 1998. Gamma-band electroencephalographic oscillations in a patient with somatic hallucinations. Lancet 352, 620—621. Baldwin, M., 1960. In: Ramey, E.R., O’Doherty, D.S. (Eds.), Electrical Stimulation of the Mesial Temporal Region. Electrical Studies on the Unanaesthetised Brain. Hoeber, New York, pp. 156—159. Bancaud, J., Talairach, J., 1993. Crises du lobe temporal. Sanofi, Paris. Bancaud, J., 1969. Epileptic crises of occipital origin (stereoelectroencepholographic study). Rev. Otoneuroophthalmol. 41, 299—314 (French). Berrios, G.E., Dening, T.R., 1996. Pseudohallucinations: a conceptual history. Psychol. Med. 26 (4), 753—763. Chen, C., Shih, Y.H., Yen, D.J., Lirng, J.F., Guo, Y.C., Yu, H.Y., Yiu, C.H., 2003. Olfactory auras in patients with temporal lobe epilepsy. Epilepsia 44 (2), 257—260. Clarke, D.F., Otsubo, H., Weiss, S.K., Chitoku, S., Chuang, S.H., Logan, W.J., Smith, M.L., Elliot, I., Pang, E.W., Rutka, J.T., Snead, O.C., 2003. The significance of ear plugging in localisation-related epilepsy. Epilepsia 44 (12), 1562—1567. Currie, S., Heathfield, K.W.G., Henson, R.A., Scott, D.F., 1971. Clinical course and prognosis of temporal lobe epilepsy: a survey of 666 patients. Brain 94, 173—190. David, A.S., 2004. The cognitive neuropsychiatry of auditory verbal hallucinations: an overview. Cognit. Neuropsychiatry 9 (1/2), 107—123. David, R., Couloujon, R., Hecaen, H., 1945. Equivalents comitiaux de type hallucinatoire divers iniques symptoms d’une tumeur temporale gauche. Ann. Med. Psychol. 2, 93—96. Dewhurst, K., Pearson, J., 1955. Visual hallucinations of the self in organic disease. J. Neurol. Neurosurg. Psychiatry 18, 53—57. Ferguson, S.M., Rayport, M., Gardner, R., Kass, W., Weiner, H., Reiser, M.F., 1969. Similarities in mental content of psychotic states, spontaneous seizures, dreams, and responses to electrical brain stimulation in patients with temporal lobe epilepsy. Psychosom. Med. 31, 479—498. Fish, D.R., Gloor, P., Quesney, F.L., Olivier, A., 1993. Clinical responses to electrical brain stimulation of the temporal and frontal lobes in patients with epilepsy: pathophysiological implications. Brain 16, 397—414. Gibbs, E.L., Gibbs, F.A., Fuster, B., 1948. Psychomotor epilepsy. Arch. Neurol. Psychiatry 60, 331—339. Gloor, P., Olivier, A., Quesney, L.F., Andermann, F., Horowitz, S., 1982. The role of the limbic system in experiential phenomena of temporal lobe epilepsy. Ann. Neurol. 12 (2), 129—144. Gloor, P., 1990. Experiential phenomena of temporal lobe epilepsy: facts and hypotheses. Brain 113, 1673—1694. Gowers, W.R., 1901. Epilepsy and Other Chronic Convulsive Diseases. Churchill, London. Gowers, W.R., 1909. The Hughlings—Jackson lecture on special sense discharges from organic disease. Brain 32, 303—326. Halgren, E., Walter, R.D., Cherlow, D.G., Crandall, P.H., 1978. Mental phenomena evoked by electrical stimulation of the human hippocampal formation and amygdala. Brain 101, 83—117. Hausser-Hauw, C., Bancaud, J., 1987. Gustatory hallucinations in epileptic seizures: electrophysiological, clinical and anatomical correlates. Brain 110, 339—359. Horowitz, M.J., Adams, J.E., Rutkins, B.B., 1968. Visual imagery on brain stimulation. Arch. Gen. Psychiatry 19, 469—486. Ionasescu, V., 1960. Paroxysmal disorders of the body image in temporal lobe epilepsy. Acta Psychiatr. Scand. 35, 171—181. Isnard, J., Mauguiere, F., 2005. The insula in partial epilepsy. Rev. Neurol. (Paris) 161 (1), 17—26.

Delusions, illusions and hallucinations in epilepsy Jackson, J.H., 1879. Lectures on the diagnosis of epilepsy, Lecture III. Medical Times and Gazette 1, 141—143 (reprinted in: Taylor, J. (Ed.), Selected Writings of John Hughlings Jackson, vol. 1, On Epilepsy and Epileptiform Convulsions. New York: Basic Books; 1958. pp. 295—307). Lance, J.W., Smee, R.J., 1989. Partial seizures with visual disturbance treated by radiotherapy of cavernous hemangioma. Ann. Neurol. 26, 782—785. Mahl, G.F., Rothenberg, A., Delgado, J.M.R., Hamlin, H., 1964. Psychological response in the human to intracerebral electrical stimulation. Psychosom. Med. 26, 337—368. Maillard, L., Vignal, J.P., Gavaret, M., Guye, M., Biraben, A., McGonigal, A., Chauvel, P., Bartolomei, F., 2004. Semiologic and electrophysiologic correlations in temporal lobe seizure subtypes. Epilepsia 45 (12), 1590—1599. Marques-Assis, L., Livramento, J.A., Cury, M., 1971. Estudo clinico de 68 casos de epilepsia occipital. Arq. Neuropsiquiatr. 29, 49—54. Mohamed, I.S., Otsubo, H., Pang, E., Chuang, S.H., Rutka, J.T., Dirks, P., Weiss, S.K., Snead, O.C., 2006. Magnetoencephalographic spike sources associated with auditory auras in paediatric localisation-related epilepsy. J. Neurol. Neurosurg. Psychiatry 77, 1256—1261. Nguyen, D.K., Nguyen, D.B., Malak, R., Leroux, J.M., Carmant, L., Saint-Hilaire, J.M., Giard, N., Cossette, P., Bouthillier, A., 2008. Revisiting the role of the insula in refractory partial epilepsy. Epilepsia (Aug) (Epub ahead of print). Palmini, A., Andermann, F., Dubeau, F., Gloor, P., Olivier, A., Quesney, L.F., Salanova, V., 1993. Occipitotemporal epilepsies: evaluation of selected patients requiring depth electrodes studies and rationale for surgical approaches. Epilepsia 34 (1), 84—96. Panayiotopoulos, C.P., 1999. Elementary visual hallucinations, blindness, and headache in idiopathic occipital epilepsy: differentiation from migraine. J. Neurol. Neurosurg. Psychiatry 66, 536—540. Penfield, W., Faulk Jr., M.E., 1955. The insula: further observations on its function. Brain 78 (4), 30—470. Penfield, W., Jasper, W.W., 1954. Epilepsy and the Functional Anatomy of the Human Brain. Little Brown, Boston. Penfield, W., Kristiansen, K., 1951. Epileptic Seizure Patterns. Springfield Ill. Penfield, W., Perot, P., 1963. The brains record of auditory and visual experience: a final summary and discussion. Brain 86, 595—696. Penfield W, Rasmussen T., 1950. The Cerebral Cortex of Man. Macmillan, New York. Penfield, W., 1938. The cerebral cortex in man. I. The cerebral cortex and consciousness. Arch. Neurol. Psychiatry 40, 417—442.

171 Rayport, M., Ferguson, S.M., 1974. Qualitative modification of sensory responses to amygdaloid stimulation in man by interview content and context. Electroencephalogr. Clin. Neurophysiol. 34, 714. Ryvlin, P., Rheims, S., Risse, G., 2006. Nocturnal frontal lobe epilepsy. Epilepsia 47 (Suppl. 2), 83—86. Salanova, V., Andermann, F., Oliver, A., Rasmussen, T., Quesney, L.F., 1992. Occipital lobe epilepsy: electro-clinical manifestations in 29 patients treated surgically. Neurology 41 (Suppl. 1), 365. Sadock, B.J., Sadock, V.A., 2000. Kaplan & Sadock’s Comprehensive Textbook of Psychiatry, Vol. 1., 7th Ed. Lippincott Williams & Wilkins, p. 800. Sveinbjornsdottir, S., Duncan, J.S., 1993. Parietal and occipital lobe epilepsy: a review. Epilepsia 34 (3), 493—521. Takeda, A., Bancaud, J., Talairach, J., Bonis, A., Bordas-Ferrer, M., 1970. Concerning epileptic attacks of occipital origin. Electroencephalogr. Clin. Neurophysiol. 28, 647—648. Taylor, I., Scheffer, I.E., Berkovic, S.F., 2003. Occipital epilepsies: identification of specific and newly recognised syndromes. Brain 126, 753—769. Trimble, M.R., 1991. The Psychoses of Epilepsy. Raven Press, New York. Van Hoesen, G.W., 1982. The parahippocampal gyrus: new observations regarding its cortical connections in the monkey. Trends Neurosci. 5, 345—350. Vignall, J.P., Maillard, L., McGonigal, A., Chauvel, P., 2007. The dreamy state: hallucinations of autobiographic memory evoked by temporal lobe stimulations and seizures. Brain 130 (Pt 1), 88—99. Weingarten, S., Cherlow, D.G., Halgren, E., 1976. The relationship of hallucinations to the depth structures of the temporal lobe. In: Sweet, W.H., Obrador, S., Martin-Rodriguez, J.G. (Eds.), Neurosurgical Treatment in Psychiatry, Pain and Epilepsy. University Arc Press, Baltimore. Wieser, H.G., Hailemariam, S., Regard, M., Landis, T., 1985. Unilateral limbic epileptic status activity: stereo EEG, behavioural and cognitive data. Epilepsia 26 (1), 19—29. Wieser, H.G., 1980. Temporal lobe or psychomotor status epilepticus. A case report. Electroencephalogr. Clin. Neurophysiol. 48, 558—572. Wieser, H.G., 1983. Depth recorded limbic seizures and psychopathology. Neurosci. Biobehav. Rev. 7, 427—440. Williams, D., 1959. The structure of emotions reflected in epileptic experiences. Brain 79 (1), 29—57. Williamson, P.D., Spencer, S.S., 1986. Clinical and EEG features of complex partial seizures of extratemporal origin. Epilepsia 27 (Suppl. 2), 46—63.