Abstracts

Rationale: Hypermotor seizures (HMS) are defined as excessive motor activity that is usually directed and complex in nature. Rheims et al 1 differentiated two sub-types of HMS: type I if the axis of the movement is anterior/posterior and type II if the axis is horizontal/rotational. The exact localization and networks involved in generating HMS is poorly understood. We aim to evaluate the cortical and subcortical activation in HMS using ictal SPECT (Single-photon emission computed tomography) according to semiological subtypes. Methods: We retrospectively studied all patients who were evaluated at Cleveland Clinic between 2005 and 2015 with video EEG who had HMS and ictal SPECT. Semiology was confirmed by two independent epileptologists. SPECT findings were reviewed by two independent physicians using Z score of 1.5 which was validated in a previous study 2. We identified 3 types of HMS, the aforementioned type I and II and a new type we classified as type III which is a mixture of types I and II. We excluded those who had generalization prior to SPECT injection. EEG, MRI, PET, surgical outcomes (if surgery was done) and pathology data were collected. In order to compare the different SPECT perfusion patterns in each group, we first normalized each individual patient to the MNI space then flipped the patients with left-sided ictal onset to the right side. We then used SPM (Statistical Parametric Mapping) to calculate an average Z score map for each group and compare hyperperfused areas on the final average maps for the groups. Results: 18 patients met inclusion criteria. Age range was 7-56 years. Seven patients (39%) had abnormal MRI. EEG was non-localizable in 12 (66%). 10 patients (55%) patients underwent surgery; 7 patients with > 22 months follow-up were completely seizure free or had worthwhile reduction in seizures. Eight patients had Type I, 6 patients with type II and 4 patients with type III HMS. SPECT injections were delivered within 10.7 seconds ( 7.2 seconds). Based on visual analysis, we found that seizures of each type involved different structures: type I seizures were generated by a more anteriorly located network which encompasses anterior insula, orbitofrontal cortex, cingulate and anterior perisylvian region and rostral midbrain. Type II seizures were generated by a more caudally located network involving the orbitofrontal cortex, cingulate (middle and posterior), basal ganglia, thalami and cerebellum. Type III seizure showed a mixed pattern of SPECT activation. When the different patterns of activation where collectively processed for each subtype using SPM, we found that type I involved the anterior insula and the anterior cingulate, where type II involved the insula, cerebellum and the thalami. Type III involved the temporal pole, and the anterior perisylvian region. Conclusions: We have identified distinct hyperperfusion patterns in the 3 different subtypes of HMS. This may provide further insight into the anatomical correlation with semiology and guide future invasive implantation targeting the appropriate network. Funding: The study was not funded by any grant