Investigation of Networks Underlying Hyperkinetic Seizures utilizing Ictal SPECT

Objective: Study neural networks involved in hyperkinetic seizures (HKS) using ictal SPECT.

Methods: We retrospectively identified 18 HKS patients evaluated at Cleveland Clinic between 2005 - 2015 with video-EEG monitoring and ictal SPECT. Semiology was confirmed by the consensus of two epileptologists' independent review and classified as type I, II or III HKS. SPECT data was analyzed by two independent physicians using z score of 1.5. Ictal hyperperfusion patterns for each group were analyzed visually and with SPM (Statistical Parametric Mapping). Spatial normalization to MNI space for each patient's data was performed, followed by flipping of data from patients with left-sided ictal onset to the right side. Finally, an average z score map for each group was calculated.

Results: Visual analysis as well as SPM identified different patterns of ictal hyperperfusion in the three subtypes of HKS: type I seizures showed hyperperfusion in a more anteriorly located network involving the anterior insula, orbitofrontal cortex, cingulate and anterior perisylvian region and rostral midbrain. Type II seizures were associated with hyperperfusion in a more caudally located network involving the orbitofrontal cortex, cingulate (middle and posterior), basal ganglia, thalami and cerebellum. Type III seizures showed a mixed pattern of SPECT hyperperfusion involving the temporal pole and anterior perisylvian region.

Conclusions: The three different semiological subtypes of HKS are each associated with distinct patterns of hyperperfusion providing further insight into the neural networks involved. This knowledge may inform placement of invasive EEG electrodes in patients with HKS semiology undergoing presurgical evaluation.