Abstracts

Rationale: The neural mechanisms of altered brain connectivity in idiopathic generalized epilepsy (IGE) remain poorly understood. A promising neuroimaging-based animal model of IGE is provided by photosensitive baboons, whose seizures can be activated by intermittent light stimulation (ILS). In this study, we compared cerebral blood flow (CBF) responses during ILS in photosensitive, epileptic (PS) and healthy control (CTL) baboons using H215O-PET. Methods: Eight photosensitive (PS) and six control (CTL) baboons matched for age, gender and weight, were classified on the basis of scalp EEG findings performed during the neuroimaging studies. Continuous intravenous ketamine (5-13 mg/kg) was used for sedation and the baboons were monitored for the level of sedation and epileptic discharges using scalp EEG. Five H215O-PET studies were acquired alternating between resting and activation (ILS at 25 Hz) scans. PET images were acquired in 3D mode (CTI/Siemens HR+ scanner, 63 contiguous slices, 2.4 mm thick) and co-registered with MRI (3T Siemens Trio, T1-weighted 3D Turboflash sequence, TE/TR/TI=3.04/2100/785 msec, flip angle=13 degrees). We examined the effective connectivity associated with visual stimulation in both groups using structural equation modeling (SEM), a novel, iterative model-generating procedure. Results: Because of the differences in group size and connectivity, eleven regions of interest (ROIs) were identified in the PS baboons (Figure 2) compared to eight in the CTL baboons (Figure 1). SEM demonstrated similarities in neural connectivity of the primary visual pathways (green) between PS and CTL animals. However, second order pathways (red) diverged between the two models, with increased interhemispheric connectivity demonstrated in the PS animals. Furthermore, while the connectivity maps in the CTL animals appeared stable, third (blue) and fourth (yellow) order connections were apparent in the PS baboons, many of which were reciprocal. Even after four iterations, latent connectivity to regions outside of the PS model remained high. Conclusions: Overall connectivity of visual pathways during ILS was different between PS and CTL baboons at the level of second order pathways. The diminished stability of the PS model compared to the CTL group may be attributed to the influence of latent connection not included in this model, as well as to the presence of reciprocating interhemispheric pathways, which may represent bidirectional feedback loops associated with the generation of generalized epileptic discharges. Evaluation of further baboons will help to expand the PS model to include further ROIs commonly connected to the epileptic networks, such as the thalamus, insula or SMA.