|Abstract: ||Epilepsy is a chronic neurological disorder that is characterized by recurrent manifestations of epileptic seizures. This disorder is controlled, but not cured, by antiepileptic drugs in about 70% of patients. Therefore, there is still a need for new therapies. Ideally, preventing epilepsy to manifest in patients at risk would be a major achievement. Therefore, we need to know how a healthy brain transforms into an epileptic brain, a process that is called epileptogenesis. The objective of this thesis was to study functional network changes during temporal lobe epileptogenesis. To reach this goal we applied the amygdala kindling model in the macaque monkey (Macaca mulatta) and performed structural MRI and four types of functional nuclear imaging techniques during kindling epileptogenesis. We imaged (1) ictal brain perfusion using 99mTc-ECD SPECT; (2) interictal brain metabolism using [18F]-FDG µPET; (3) interictal cannabinoid type 1 receptor expression using [18F]-MK-9470 µPET and (4) interictal GABAA receptor expression using [11C]-FMZ µPET. When the monkeys reached stable stage IV seizures, i.e. when they were fully kindled, we studied effective connectivity of the kindled amygdala using electrical microstimulation during fMRI (EM-fMRI). Finally, we performed a short pilot study of deep brain stimulation (DBS) in two imaging-defined targets in order to perturb seizure activity.|
In the first research chapter (chapter 3), we described the progression of the amygdala kindling model in terms of seizure duration, seizure semiology and perfusion changes. Similar to Wada and colleagues, who were the first to systematically describe the electroclinical changes in the rhesus monkey amygdala kindling model, we found a very slow progression in afterdischarge duration (ADD). We applied daily electrical stimulation in order to elicit a seizure for 477 days in monkey K. and 515 days in monkey S. During this period, ADD ranged from on average 18s in the first clinical stage to 52s in the last clinical stage. Four different clinical stages in the rhesus monkey during amygdala kindling were defined by Wada and colleagues. We followed this categorization of seizures based on their semiology. Stage I was defined as interruption of ongoing behavior; stage II seizures were characterized by oral automatisms; stage III by unilateral dystonic and clonic movements and stage IV as secondarily generalized seizures with bilateral dystonic and clonic movements. In each of these four stages, we imaged the cerebral perfusion during a seizure using ictal SPECT imaging. Therefore, we injected the radiotracer 99mTc-ECD immediately before evoking a seizure. The uptake of the radiotracer reflects then the perfusion distribution during the first part of the seizure. By subtracting the perfusion distribution in the baseline state, we could measure the seizure specific increases and decreases in perfusion. From the early stage I on, we observed a distributed pattern of regions of hyper- and hypoperfusion in cortical and subcortical structures. Combining the perfusion changes from the four stages, we defined for each monkey a ‘common network’, i.e. the perfusion changes that are present in every seizure stage. Over time, we saw that most perfusion changes arose from this common network.
In chapter 4, we described the preliminary results of our PET experiments. Besides ictal perfusion, we were interested in investigating interictal changes in brain metabolism, CB1R and GABAA receptor expression in our monkey kindling model. So far, we can only draw conclusions about the last stage of kindling. Despite a large intersubject variability in our baseline group (n=8), we observed significant changes in cerebral metabolism and CB1R expression. For monkey K., changes were seen in the stimulated amygdala (hypometabolism), the contralateral amygdala (increased CB1R expression), the ipsilateral hippocampus (increased CB1R expression), the bilateral putamen (hypometabolism), the ipsilateral insula (hypometabolism), the frontal lobe (hypometabolism contralaterally, and decreased CB1R expression bilaterally) and the contralateral occipital cortex (hypermetabolism and increased CB1R expression). Monkey S. expressed interictal changes during stage IV in the ipsilateral insula (hypometabolism), the occipital cortex (hypometabolism contralaterally, decreased CB1R expression bilaterally), the contralateral hippocampus (hypermetabolism, increased CB1R expression), contralateral putamen (increased CB1R expression) and the auditory part of the temporal cortex (increased CB1R expression).
When the monkeys were fully kindled, we investigated to which degree the observed ictal perfusion changes reflected the connectivity of the seizure focus. Therefore, we applied EM-fMRI to study effective connectivity of the kindled site (chapter 5). We observed in both monkeys widespread fMRI activations and deactivations in bilateral cortical and subcortical structures. On the regional level, many changes in fMRI signal were overlapping with the ictal perfusion changes of stage IV. On the voxel level, we observed that in general only a part of an anatomically defined region presented changes in both connectivity and perfusion. Additionally, EM-fMRI revealed widespread areas that were multisynaptically connected to the seizure focus in stage IV kindled monkeys.
In the last research chapter, we applied deep brain stimulation (DBS) in the ipsilateral mediodorsal nucleus of the thalamus and the contralateral dentate nucleus of the cerebellum in order to interfere with seizure activity, and ultimately to abolish seizure activity as a potential therapy for temporal lobe seizures. We used a range of different DBS parameters, applied at different timings with respect to the kindling pulse. We observed some modulation of the ADD, but seizure activity was never completely interrupted. Also, clinical expression of the seizures remained similar to baseline stage IV conditions.