JLE

Epileptic Disorders

MENU

Seizures and the basal ganglia: a review of the clinical data Volume 4, supplement 3, Supplement 3, December 2002

Figures

See all figures

The striatum is considered to be the major receptive and integrative component of the basal ganglia and receives dense projections from the entire cerebral cortex and the amygdala. Subsequently, this structure projects indirectly to thalamic and brainstem structures (see Slaght et al. in this issue). It is therefore quite possible that epileptic discharges originating everywhere in the cortex affect basal ganglia. The basal ganglia can either be involved passively, as a propagation pathway or in executing some aspects of the ictal process. Alternatively, the basal ganglia may actively interact with the epileptic phenomena as a homeostatic factor to control epileptic activity. Since the 80's, numerous animal studies have provided data that supports the latter view (see Deransart and Depaulis; Veliskova et al. in this issue). The aim of the present paper is to reappraise the clinical and electrophysiological evidence that supports the involvement, either passive (propagation) or active (control), of the basal ganglia in human epileptic seizures.

Historical perspectives

While Jackson's theories regarding the cortical contribution to epileptogenesis are well known, his descriptions of Òmiddle and lower levelÓ convulsions have been, surprisingly, largely neglected. Middle-level fits were associated with motor or rolandic cortex and striatum, and most likely correspond to what we today regard as focal motor seizures. Jackson strongly believed Òthe striatum to be the part discharged in convulsions beginning unilaterallyÓ and that the convulsions contained Òprocesses representingmovementsÓ [2]. The highest level was localized to the frontal, or what he called ÒprefrontalÓ regions, and considered these seizures to be Òepilepsy properÓ. Jackson also proposed the concept of a lower brain-stem level (cord, medulla, and pons...) for the origin of some seizures [1].

In 1954, Penfield and Jasper [3] reconsidered the concept of middle and lower level convulsions, and proposed the concept of the centrencephalic system. ÒThe centrencephalic system may be defined as that neuron system, centeringin the higher brainstem, which has been up to the present, or may be in the future, shown to have equal functional relationships with the two cerebral hemisphereÓ. According to these authors, the various cortical areas receive projections of the diencephalon, and their separate functions can only be carried out when a diencephalic center and a cortical area are acting together as a unit.

In 1972, Gastaut and Broughton [4], introduced the concept of corticonuclear sectors as a fundamental pathophysiological property of the propagation patterns of partial seizures. As they stated Òthe relatively diffuse discharge tends to propagate along projection pathways to other distant but functionally connected normal neuronal groups, as well as spreading to neighboring populations of cellsÓ. This explains how the discharge rapidly evolves to include an entire system with its cortical and subcortical connections and, eventually, others adjacent to it at either the cortical or subcortical poles of the system.

Nevertheless, to date, there are few data to demonstrate the involvement of basal ganglia in seizures propagation. In 1949, Hayne et al. [5] reported their experience of electrical activity of subcortical areas in epilepsy. Their data showed that cortex and subcortical regions in epileptic patients display comparable normal and abnormal activities although various subcortical areas may show entirely independent, abnormal activity. However, isolated seizure discharges from the cortex are common, whereas they are rare in subcortical structures. The fact that they do occur is important, because they may account for therapeutic failure in some cases where a cortical focus is removed.

Ajmone Marsan and Abraham [6] reported their experience with the use of chronically implanted electrodes in seizure disorders. There were respectively 6, 4, 2, 1 patients with electrode implantations within the putamen, globus pallidus, caudate nucleus, and subthalamus. Unfortunately, the authors did not mention the results on the interictal or ictal EEG activities in these subcortical structures. More recently, Rektor et al. [7] investigated the presence of ictal discharges within the putamen by recording with depth electrodes, the epileptic activities in patients with temporal lobe epilepsy. No epileptic interictal or ictal discharges were noticed in the putamen. When the activity remained localized to the seizure-onset zone, the activity of the basal ganglia did not change. The spread of epileptic activity to other cortical structures was associated with slowing of basal ganglia EEG activity. Epileptiform activity has been recorded directly from the subthalamic nucleus in association with scalp recordings ([8-10] see also Chabardés et al. in this issue). These data are in line with the acknowledged presence of a pathway from the cortex to the subthalamus in humans, as described in animal studies [10] (see Slaght et al. in this issue).

Ictal basal ganglia involvement: clinical observations

Unilateral upper limb dystonic posture contralateral to the ictal temporal lobe discharge represents the clinical feature usually attributed to the spread of ictal discharge to the basal ganglia [11]. We would suggest herein that two other clinical patterns of epileptic seizure may involve basal ganglia: (i) rotatory seizures and (ii) paroxysmal dyskinesia- like seizures. The main hypothesis underlying these seizures is that the spread of the ictal discharge should disorganize normal functions within the basal ganglia and thus lead to the impairment of the ÒprimaryÓ ictal behavior.

Unilateral dystonic posturing during temporal lobe seizures

During the clinical course of a typical temporal lobe seizure, an asymmetrical dystonic posture may affect the upper limb, contralaterally to the epileptic focus [12, 13]. This has been attributed to the spread of the discharge to the ipsilateral basal ganglia, mainly the sensorimotor part of the putamen, as supported by ictal single photon emission computed tomography (SPECT) data, showing ipsilateral hyperperfusion involving the lenticular nucleus [14] (see also Semah in this issue). Interictally, putamen hypometabolism shown by a fluorodeoxyglucose positron emission tomography (FDG-PET) study in such patients, confirms a more profound alteration within the basal ganglia [15]. It is noteworthy that persistence of ictal contralateral dystonia posture was reported in a hemispherectomized patient (sparing the basal ganglia and a small amount of ventrobasal prefrontal cortex) experiencing recurrent seizures after surgery [16]. This observation suggests that descending pathways from the basal ganglia to the brainstem may be preferentially involved in the production of dystonia, since the ÒascendantÓ pathway via thalamocortical afferents was removed. Further evidence supporting the Òdescendant theoryÓ is the report of a patient showing congenital mirror movements and temporal lobe epilepsy, whose dystonia was not mirrored during seizures [17]. If the dystonia was conveyed through basal ganglia to the thalamus and back to the motor cortex, it should have been mirrored during the seizure according to the presumed cortico-spinal tract involvement in such phenomena. Its absence may imply that another tract, e.g. directly from the pallidum to the brainstem and spinal cord is involved. This proposal is in line with the presumed separate physiopathology of ÒposturalÓ and ÒactionÓ dystonia, attributed respectively to the involvement of descendant and ascendant pathways from the pallidum. These views are represented in figures 1a and 1b, showing how an epileptic activity involving the temporal lobe could activate motor programs subserved by cortico-subcortical circuits. The spread of the discharge to the ipsilateral basal ganglia is suggested to be responsible for the contralateral ictal dystonic posturing (figure 1b). It is interesting to mention that a similar involvement of basal ganglia has been suggested by SPECT studies during frontal seizures associated with dystonic movements [18]. These data suggest the existence of a common, final pathway accounting for the production of dystonic movements during localization-related seizures and carried out by basal ganglia.

Rotatory seizures

Rotarory seizure is defined as one or more sudden rotations of the entire body around its vertical axis for at least 180 o, during an epileptic seizure. There are more than one hundred case reports in the literature concerning these ictal rotations [19-34]. Several patients with rotatory seizures have location-related seizures from very diverse origins (occipital, temporal, frontal, frontotemporal seizures). In strong contrast to versive seizures during which head and eye turning is often the first observed manifestation of focal seizure activity, turning behavior in rotatory seizures is more integrated with the ictal behaviour occurring in the context of localization-related seizures. This observation, together with the large distribution of cortical areas potentially implicated in this type of seizure, strongly suggest a final common pathway.

Interestingly, some patients with rotatory seizures did have subcortical lesions. A sixteen-year-old boy, with Wilson's disease and severe brain lesions of the bilateral frontal lobe, lenticular nuclei and thalami, was described with rotatory seizures (clockwise turns, two or three times around his axis) [23]. In another report, a female patient with a thalamic lesion exhibited paroxystic 360 o turning behaviour, followed by speech difficulties, and generalized tonic-clonic seizure [25]. In this latter case, the circling was contralateral to the lesion. A female patient with tuberous sclerosis was also described with rotatory seizures [22] (case detailed in [35]). She had a subependymal nodule in the thalamo-striatal sulcus, adjacent to the left caudate nucleus, and during the seizures, she turned to the right. Interestingly, the rotatory movements were alleviated by pimozide (a dopamine antagonist) and aggravated by small doses of levodopa, suggesting involvement of dopamine transmission [35]. More recently, in another case report of rotational seizure in the context of tuberous sclerosis, no anomaly within or in the proximity of the basal ganglia was noted, suggesting that another mechanism could be involved [33]. In a patient with a frontotemporal epilepsy, the impulse to rotate was accompanied by hyperperfusion in the lenticular nucleus, contralateral to the side of the rotation [34]. In summary, these data suggest that basal ganglia dysfunction may be partly responsible for the production of the rotatory behaviour.

As already quoted, rotatory seizures have been also reported in patients with 3 Hz spike-and-wave discharges [21] and other forms of idiopathic generalized epilepsy [19, 20, 27, 28, 31]. It has been suggested that an imbalance in the excitability of the hemispheres could be relevant in generalized epilepsy patients with rotatory seizures [28]. It can be speculated that the imbalance in the excitability of the hemispheres may reflect the imbalance of subcortico-cortical reciprocal influence, as it represents a well-known mechanism of regulation of cortical excitability (see below). This imbalance may also be responsible for the circling behaviour seen in non-epileptic patients with putamen lesions [36] or post-encephalitic bilateral temporal lobe lesions [37]. In summary, there is substantial evidence suggesting an implication of the basal ganglia in the production of continuous or paroxysmal rotatory behaviour. Figure 2 illustrates the proposal for the imbalance theory of the basal ganglia involvement responsible for the execution of the lateralized cursive behavior. It has been suggested that dopaminergic stimulation in the non-epileptic rodent bearing unilateral basal ganglia lesions, could act as a model for this paroxysmal rotatory behaviour (first suggested in reference [24], for a detailed presentation of the model see references [38, 39]). Accordingly, in patients with epilepsy, the ictal impulsion to move should promote the rotational behaviour, as a result of the basal ganglia imbalance (figure 2).

Paroxysmal dyskinesia-like seizures

Paroxysmal dyskinesias are brief episodes of involuntary movements, mainly of the choreodystonic type, without impairment of consciousness or ictal EEG changes [40]. Lance proposed three subtypes: paroxysmal kinesigenic choreoathetosis (PKC), familial paroxysmal dystonia (FPD) and exercise-induced paroxysmal dystonia (EIPD) [41]. Later, nocturnal paroxysmal dyskinesia was described as a fourth group. While this last category of paroxysmal dyskinesia was first believed to represent either a sleep disorder or a basal ganglia paroxysmal dysfunction, there is accumulating evidence to suggest that it is related to nocturnal frontal epilepsy [42, 43, 47]. Moreover, even diurnal episodes with paroxysmal choreodystonic movements have been attributed to frontal epilepsy, although no scalp EEG ictal changes occurred. Lombroso suggested that electrical discharges picked up from the medial frontal lobe with depth electrodes, could spread to the caudate nucleus where it could be responsible for the genesis of paroxysmal dyskinesia [43]. An imbalance in dopaminergic reactivity has been suggested in another case report [44].

Paroxysmal dyskinesias, either of the kinesiogenic or the non-kinesiogenic-type are borderline entities, thought to be linked to acute basal ganglia dysfunctions [40]. They may be related to familial channelopathies and, in some instance, a relationship to overt epilepsy has been shown. In such families, patients may have infantile convulsions, later followed by paroxysmal dyskinesia [45, 46]. This association may suggest a common pathophysiology.

In summary, as frontomesial seizures may closely mimic paroxysmal dyskinesia, a strong interrelationship between these premotor cortical areas and the basal ganglia can be suggested. Accordingly, the Òprimer moverÓ can either be the premotor cortical area (in frontomesial epilepsy) or the basal ganglia (in paroxysmal dyskinesia), leading to a common semiology.

Figure 3 represents a tentative illustration suggesting how basal ganglia dysfunction may contribute to the phenomenology of paroxysmal dyskinesias.

Basal ganglia pathology and its influence on epilepsy

Since the early medical literature, cases have been reported showing an inverse relationship between epilepsy and parkinsonism, with seizures vanishing as the parkinsonian state developed [48-53]. Moreover, some epidemiological studies suggest that epilepsy is observed at a remarkably low rate in patients with idiopathic Parkinson's disease (IPD) compared to an age-matched normal population [55-57]. Patients with Parkinson's disease seem to be relatively protected from seizures. It has recently been reported that in Òpatients with epilepsy who develop Parkinson's disease, seizures become easier to controlÓ [54]. These authors suggested that, ÒThis may be due to the fact that the globus pallidus and the substantia nigra pars reticulata flood the brainstem with inhibitory stimuli and tend to turn off the structures that promote seizuresÓ [54]. As basal ganglia influence cortical activity by synchronizing large-scale activities through intracortical inhibition, it has been hypothesized that dopamine deprivation leads to a failure of this form of intracortical inhibition [58]. This deficit, observed in patients with Parkinson's disease, can be restored by apomorphine, a dopamine agonist, as shown by transcranial magnetic studies [59]. Thus, it is also possible that the drug treatments for IPD, mainly based on dopaminergic drugs, may suppress seizures. There is evidence from patients with untreated epilepsy that intracortical inhibition is enhanced in idiopathic generalized epilepsy, thus facilitating the synchronization of abnormal activities [60, 61]. According to these data, obtained separately in IPD patients and patients with idiopathic generalized epilepsy, it is conceivable that both disorders constitute two opposite poles of a spectrum involving intracortical inhibition, ranging from insufficient (IPD) to excessive (idiopathic generalized epilepsy).

In this setting, studies addressing the effect of dopamine, the main neurotransmitter within the basal ganglia, on epilepsy and seizures are of some relevance. Levodopa has been studied in patients with myoclonic epilepsy [62] or photosensitivity [63]. In five patients with primary generalized epilepsy exhibiting photosensitivity, subcutaneously injected apomorphine transiently blocked the response [63]. However, low doses of subcutaneously injected apomorphine did not change the rate of spontaneous spike-and-wave discharges [63]. A few reports of patients with Parkinson's disease exhibiting photosensitivity have also been published, where dopamine or a dopamine agonist suppressed the photosensitivity [52, 64-66].

The effect of an acute basal ganglia lesion on the course of epilepsy is poorly documented. An 8year-old girl with left hemisphere, multifocal cortical dysplasia and severe intractable epilepsy causing daily seizures and mental retardation, developed an acute encephalopathy including choreoathetotic movements, ataxia and confusion [67]. During the 6 month period of recovery, she did not have any epileptic attack even though the antiepileptic treatment remained the same. The brain MRI revealed a bilateral, T2, high intensity signal involving the heads of the caudate and putamen nuclei [67]. After 6 months of follow- up, the seizures reappeared. The authors suggested that the propagation pathway of epilepsy through basal ganglia must have been transiently involved in this patient. Although heterogeneous, the amount of data presented here are strongly suggestive of an interaction between basal ganglia pathology, either chronic or acute, and the occurrence of seizures in patients with epilepsy. The modulation of the cortical excitability, by the basal ganglia, especially the intracortical inhibition responsible for synchronicity [58], remains to be examined.

Interictal basal ganglia and neuropsychiatry

If basal ganglia are involved in the control of seizures, as suggested by animal studies, and in particular in the maintenance of interictal states, one should expect clinical signs reflecting the interictal functional recruitment of these circuits. Obviously, patients with epilepsy never exhibit parkinsonian signs interictally, but they could present behavior and thought disorders potentially related to a basal ganglia dysfunction. Imaging studies have highlighted subcortical metabolism changes in patients with temporal lobe epilepsy [15, 68] (see also Semah in this issue). While some authors have suggested that subcortical hypometabolism ipsilateral to the epileptic focus should reinforce the epileptogenic potential in such patients [68], accumulating evidence supports a strong participation of basal ganglia in some aspects of clinical psychiatry. In particular, as suggested by a recent review paper [69], basal ganglia dysfunction may contribute to depression, psychotic behavior, obsessive-compulsive behavior, a full range of manifestation known to occur during interictal states. Whether these manifestations could be related to interictal basal ganglia dysfunction remains speculative.

CONCLUSION

A wide range of ictal and interictal manifestations may reflect the involvement of basal ganglia at any point in the history of an epileptic disorder. Regarding the ictal clinical semiology, movement disorders associated with epileptic electrographic discharges could indicate a participation of basal ganglia. Dysfunction of basal ganglia interictally may modify the expression of seizures, which could, in turn, be responsible for some aspects of the interictal phenomenology, through a basal ganglia dysfunction. Understanding the role of the basal ganglia in epilepsy may lead to innovative therapeutic approaches.