Although the differential diagnosis of ICD-10 and DSM-IV dissociative disorders explicitly requires the exclusion of symptoms with an identifiable neurological basis, many of the phenomena associated with epilepsy, particularly temporal lobe epilepsy, have been regarded as dissociative in nature (Devinsky et al., 1989; Good,
1993). Indeed, ICD-10 includes a specific category for dissociative disorders due to a general medical condition, which encompasses many of the symptoms exhibited by individuals with epilepsy. The absence of such a category from DSM-IV, however, reflects doubt concerning the value of attaching the dissociative label to these phenomena. In my view, such doubt is well justified in many cases.
The notion that many epileptic phenomena can be regarded as dissociative is based, to a considerable extent, on the frequent occurrence of amnesia in epilepsy (Good, 1993). Individuals with complex partial or generalized seizures typically display profound amnesia for events occurring during the ictus. Moreover, certain people experience a postictal fugue state characterized by apparently purposeful behaviour for which they are subsequently amnesic, much like dissociative fugue. Despite their prima facie resemblance to dissociative phenomena, however, these events should not be regarded as episodes of dissociation (Good, 1993). Dissociative amnesia is characterized by an inability to retrieve information that has been learnt and is present in memory despite its inaccessibility (American Psychiatric Association, 1994). Amnesia for ictal events, in contrast, reflects a disruption in normal information processing, resulting from uncontrolled neural activity, that prevents the encoding of new material during the ictus. The ictal amnesia is not a product of a retrieval failure, therefore, but simply the absence of memories to retrieve. It is for this reason that this form of amnesia is irreversible, unlike most cases of dissociative amnesia (American Psychiatric Association,
Postictal fugue should not be regarded as a dissociative episode for similar reasons. Unlike dissociative fugue, postictal fugue is characterized by a disruption in consciousness associated with significant confusion and an abnormal EEG (Thomas and Trimble, 1997). The apparently purposeful behaviour displayed in postictal fugue is not dissociated from ongoing cognitive activity as it is in dissociative fugue; rather, it occurs in the relative absence of such activity. The inability to reverse the amnesia associated with postictal fugue serves as an illustration of this fact. The behavioural automatisms often observed in the context of complex partial seizures and regarded as a dissociative phenomenon by some (Good, 1993) are amenable to a similar interpretation. Like the behaviours exhibited during postictal fugue, ictal automatisms only occur in the context of a disruption in consciousness and disturbed behavioural control; genuinely dissociated behaviours are noteworthy because they occur despite an otherwise intact ability to control action2. In both cases, it is likely that these behaviours result from the uncontrolled activation of circumscribed motor programs by epileptic discharges in neural sites associated with behavioural control. The fact that such automatisms are particularly characteristic of seizures originating in the frontal lobes lends support to this view.
Many of the phenomena associated with partial seizures originating in the temporal lobes, such as hallucinations, sensory and cognitive auras, déjà vu and déjà veçu should also be distinguished from true episodes of dissociation. Although hallucinations and auras have a phenomenology that departs from external reality, these phenomena involve the paradoxical integration of information within conscious awareness. As such, they may be more appropriately regarded as phenomena of association rather than dissociation (Kihlstrom, 1994). The phenomenology of epileptic hallucinations and auras probably originates in the activation of representational structures in the temporal lobes, either directly by seizure activity in representational networks, or indirectly through seizure-related stimulation of limbic structures such as the amygdala and anterior cingulate (Bancaud et al., 1994). Experiences of epileptic déjà vu and déjà veçu are also more associative than dissociative and may involve a similar neurophysiological process. For example, stimulation of the amygdala by seizure-related discharges could imbue current perceptions and cognitions with an unwarranted emotional colouring that may be experienced as a sense of having encountered the situation before (Bancaud et al., 1994; Sierra and Berrios, 1998).
Certain phenomena associated with temporal lobe epilepsy can, however, be regarded as genuine examples of dissociation according to the scheme described by Cardena (1994). Depersonalization and derealization commonly occur in the
2 'Dissociated' behaviours should be distinguished from normal automatic behaviours because the latter are generally in accord with system goals and can therefore be considered voluntary (Brown, 1999; see also Cardena, 1994).
context of temporal lobe epilepsy and involve an alteration in consciousness characterized by dissociation from the self and/or the environment. According to Sierra and Berrios (1998), depersonalization and derealization are the products of a vestigial defence mechanism evolved to provide the optimum processing conditions for adaptive behaviour in the face of threat. By this view, extreme anxiety triggers an inhibitory response from the left prefrontal cortex that dampens output from the sympathetic nervous system, through inhibition of the amygdala and anterior cingulate. In turn, the right prefrontal cortex is activated by ascending arousal systems controlled by uninhibited amygdala circuits, generating further inhibition of the cingulate. As a result, the individual experiences a sense of vigilant alertness devoid of any emotional or cognitive content, a state that is ideally adapted for the control of action in the face of extreme and potentially debilitating danger. If this response is triggered in the absence of threat, however, the resulting sense of deper-sonalization and derealization can itself be highly unpleasant and incapacitating. Given the validity of this account, depersonalization and derealization in the context of temporal lobe epilepsy may be the result of seizure activity in the amygdala that prevents the emotional tagging of perceptual and cognitive information prior to its entry into conscious awareness3. It may be that this process is the product of transient disconnection between the amygdala and sensory areas resulting directly from seizure activity. Alternatively, it may reflect an indirect defensive response in the face of anxiety elicited by seizure-based stimulation of the amygdala. Intuitively, one suspects that the former is the more plausible possibility, although the latter cannot be ruled out a priori. Following this account of deper-sonalization and derealization, these phenomena can be regarded as dissociative in sense (ii) of the term; whether they should, in the context of epilepsy, be regarded as the result of a dissociative defence mechanism remains an empirical issue.
The fact that few epileptic phenomena can be regarded as dissociative in any strict sense reflects widespread confusion over what actually constitutes dissociation. Although widely endorsed, the idea that any breakdown in memory, consciousness, identity, perception or behavioural control is dissociative overextends the term and diminishes its descriptive validity (Cardeña, 1994). Amnesia cannot be considered genuinely dissociative unless it involves an inability to retrieve intact information that should, under normal circumstances, be available for recall (Kihlstrom, 1994). Amnesia resulting from a failure to encode information, including that which occurs in the context of epilepsy, does not fall within this category. Loss of behavioural control can only be considered dissociative if it is within the context of an otherwise intact ability to control action. Seizure-related motor phe-
3 Such a process could also be responsible for the paradoxical sense of unfamiliarity that characterizes jamais vu, a phenomenon commonly observed in epilepsy.
nomena, including complex automatisms, are not dissociative because they occur only in the context of reduced behavioural control in general. Current psychiatric taxonomies do not make these distinctions clearly enough, relying instead on a purely descriptive approach that precludes precise classification based on the mechanisms underlying different phenomena.
Was this article helpful?