JLE

Epileptic Disorders

MENU

PET imaging in epilepsy: basal ganglia and thalamic involvement Volume 4, supplement 3, Supplement 3, December 2002

Figures

See all figures

Several authors have suggested that the basal ganglia are involved in the control of epileptic seizures [1], but the precise role of the subcortical structures is still debated. Clinical, and electrophysiological studies [2], as well as experimental data have demonstrated basal ganglia abnormalities in generalized and partial seizures (see Deransart and Depaulis; Veliskova et al.; Vercueil and Hirsch in this issue). Furthermore, functional imaging data have suggested that the basal ganglia are involved in partial seizures. Ictal hyperperfusion of the striatum was shown using single photon emission computed tomography (SPECT) in patients with temporal lobe epilepsy [3]. Several studies have also reported a striatal interictal ipsilateral hypometabolism using positron emission tomography (PET) and fluorodeoxyglucose (FDG) as well as neurotransmission changes in patients with partial epilepsy [4, 5]. This paper will describe imaging studies of the basal ganglia in patients with epilepsy, including anatomical data, and focusing especially on metabolic, GABAergic and dopaminergic studies.

Morphometric analysis of the basal ganglia

Few studies have been conducted on the anatomy of the basal ganglia in epilepsy. Several authors have reported subtle changes in the volume of some subcortical areas, such as the thalamus [6, 7]. The relationships between these changes and the epilepsy itself have seldom been investigated. A recent study using volumetric measurements of subcortical nuclei using magnetic resonance imaging (MRI) was performed in a group of 27 patients with mesial temporal lobe epilepsy (MTLE) versus 14 controls [8]. The authors reported a significant reduction of the volume of the thalamus, the caudate nucleus and the putamen, bilaterally. No significant correlation between volume measurement of the subcortical structures and age of the patients, age at onset, duration of epilepsy, total seizure frequency and frequency of generalized seizures, was found. In patients with febrile convulsions, there was a bilateral decrease of the volume of the putamen and an ipsilateral decrease of thalamic volume, whereas in patients without febrile convulsions the volume of the caudate nuclei was smaller on both sides [8]. This study suggested that structural changes occur in the basal ganglia of patients with partial epilepsy.

Cerebral metabolism in partial epilepsies

Using PET, several authors have reported subcortical functional changes in patients with partial epilepsy. Thalamic changes were the most frequently reported subcortical abnormalities [4, 5, 9]. Striatal interictal ipsilateral hypometabolism was also observed using FDG-PET in patients with mesial temporal lobe epilepsy (figure 1) and in extra- temporal lobe epilepsies (figure 2) [4, 5]. Indeed, in patients with frontal lobe epilepsy, large cortical hypometabolism and severe subcortical hypometabolism that involved the caudate nucleus, the striatum and the thalamus has been described [10].

We recently conducted a PET study in 30 patients with mesial temporal lobe epilepsy to evaluate the correlation between ictal symptoms and interictal functional changes in subcortical structures [11]. Contralateral dystonia was previously described as a good lateralizing sign of mesiotemporal lobe seizures [12, 13]. The hypothesis was that the dystonia observed during temporal lobe seizures could be related to the propagation of the seizures to subcortical structures. In our PET study, a significant association was found between cerebral metabolism and ictal dystonia in patients. Patients with ictal dystonia had more severe hypometabolism in the striatal region ipsilateral to the seizure focus. A similar association was found with ipsilateral orbito-frontal metabolism but no other cerebral regions [11]. These findings are consistent with a previous ictal SPECT study showing that patients with temporal lobe epilepsy and ictal dystonia had hyperperfusion in the epileptogenic temporal lobe and the ipsilateral basal ganglia, whereas patients without ictal dystonia had only temporal lobe hyperperfusion [3]. The combined ictal and interictal data suggest that the striatum and thalamus take part in the network of partial seizures of temporal lobe origin (figure 3), perhaps reflecting propagation pathways.

Cerebral metabolism in generalized epilepsies

Very few studies have been performed in generalized epilepsy and it is generally believed that patients with primary generalized epilepsy have normal neuroimaging studies. Using FDG-PET, Theodore et al. reported no significant interictal changes in patients with absence seizures and tonic-clonic generalized seizures [14]. However, Swartz et al. showed some degree of regional decreases in relative glucose uptake in patients with juvenile myoclonic epilepsy. During a visual, working memory task, the anticipated physiological increase of metabolic activity was not observed in the dorsolateral prefrontal cortex, the premotor cortex or the basal frontal cortex [15].

Neurotransmitters studies

PET studies of various neurotransmission systems in epileptic patients have provided informations about the involvement of the basal ganglia in epilepsy. Most of these studies have evaluated the GABAergic and dopaminergic systems.

GABAergic system

Flumazenil (FMZ), a specific benzodiazepine (BZ) antagonist, binds reversibly to the benzodiazepine-GABAA receptors (BZR). FMZ labeled with carbon 11 was developed as a PET-tracer nearly 20 years ago by Mazière et al. and is widely used for imaging the benzodiazepine-GABAA receptor [16]. There is still controversy as to whether FMZ-PET abnormalities reflect changes in neuronal density. Burdette et al. reported a close correlation between neuronal density and BZR density, measured using autoradiography, in the hippocampus [17]. However, Koepp et al. found no correlation between neuronal density, MRIbased hippocampal volume and PET measurement of BZR density [18]. Ryvlin et al. also reported data on the variability of FMZ-PET results in two patients who underwent 2 FMZ-PET examinations at 3 month intervals, suggesting that mechanisms other than neuronal density could be responsible for FMZ-PET binding changes [19]. Thus it seems that FMZ binding may not only reflect neuronal loss but also benzodiazepine receptor changes.

Because FMZ-PET studies mainly focus on the clinical usefulness of FMZ as compared to FDG-PET during the presurgical evaluation of medically refractory patients with partial epilepsy, very few data are available regarding the basal ganglia [20]. Most of these studies have shown that FMZ-PET reveals more restricted areas of functional changes than does FDG-PET. Using group analysis of patients with temporal lobe epilepsy (TLE), two recent PET studies failed to detect flumazenil binding changes in the basal ganglia, suggesting that BZR binding is not altered in the basal ganglia of patients with TLE [21, 22].

In untreated generalized epilepsies such as childhood and juvenile absence epilepsy, two studies did not find any change in the thalamus, the basal ganglia, the cerebellum or the occipital and parietal cortex, using FMZ-PET [23, 24]. Others have reported a moderate decrease of FMZ binding in the thalamus of patients with generalized tonic- clonic seizures [25].

Dopaminergic system

The role of dopamine in epilepsy was suggested several decades ago when dopamine receptor antagonists were introduced. These compounds were found to help schizophrenia but to increase seizures. Lamprecht and others saw epilepsy as a reduction in dopaminergic function [26], but others have reported elevated urinary, plasma, or cerebral spinal fluid levels of dopamine in epileptic patients (for review see Starr [27]). Experimental studies have suggested that the monoamines might regulate the initiation and spread of seizure activity [27]. It has been suggested that paroxysmal activity in the cortex leads to an increased activity of the glutamatergic fibers, thereby increasing tonic release of dopamine and downregulating dopamine receptors [28]. It thus seems reasonable to think that the basal ganglia are involved in seizure propagation and that dopamine may be important in this process, as well as in determining the seizure threshold (see Deransart and Depaulis in this issue). We recently conducted a study on the involvement of the dopaminergic system in epilepsy [29]. Fourteen patients with epileptic seizures associated with ring chromosome 20 (rCh20) mosaicism were studied. The clinical seizures of these patients consisted of Òabsences statusÓ, seizures with pseudo-focal onset or hypertonic seizures often with bilateral synchronous EEG abnormalities. The most distinguishing feature is the long duration of the seizures, often lasting more than 15 minutes, suggesting a dysfunction of the endogenous systems involved in the control of seizures. In these patients, [18F]-fluorodopa positron emission tomography (PET) was performed during continuous EEG recording. [18F]fluoro- L-DOPA uptake was significantly decreased in the putamen and in the caudate nucleus of patients as compared to 10 control subjects (figure 4). Although the significance of these findings or the cause or consequence of seizures remains unknown, these data suggest, along with experimental data (see Deransart and Depaulis in this issue), that the dopaminergic system plays a role in epileptic seizures. This role remains to be elucidated and both experimental and clinical studies are currently being performed to answer these questions.

CONCLUSION

In conclusion, functional imaging studies strongly suggest that the basal ganglia are involved in epileptic seizures. PET studies should be performed in patients with epilepsy to investigate the basal ganglia changes in other major brain neurotransmitters. Further studies are also needed to evaluate the significance of these abnormalities in the different syndromes of generalized and partial epilepsies.

Acknowledgements: The authors thank the staff of the PET center for their assistance in carrying out the PET studies. We would also thank F. Chassoux for help in the discussion of these data. Original research reported by the author was supported in part by a grant from GSK, France.