Abstracts

Rationale: Prediction of seizure onset would aid treatment development in patients with generalized epilepsy. Recent evidence suggests that seizures may be more probable in certain states as measured with dynamic functional connectivity. The objective of this study, therefore, was to investigate the brain-state dynamics of patients with generalized genetic epilepsy and spike-wave discharges using functional MRI and Leading Eigenvector Dynamics Analysis (LEiDA) [1].

Methods: We conducted a functional neuroimaging study in 34 patients with generalized epilepsy (20 children and 14 adults) and 39 healthy controls (20 children and 19 adults). All participants underwent a resting-state functional magnetic resonance imaging (fMRI) study paired with simultaneous electroencephalography (EEG). The EEG was used to identify the occurrences of generalized spike-waves (GSW). We parcellated the brain into 90 cerebral regions according to the Automated Anatomical Labelling atlas, extracted the time series fMRI blood-oxygen-level-dependent (BOLD) signal per region and devised a whole-brain functional connectome (Figure 1). Recurrent functional connectivity configurations were computed using dimensionality reduction and clustering, adopting the previously developed LEiDA approach [1] (Figure 1). Seven recurring brain states were selected to provide an optimal clustering solution. The occurrence probabilities, durations and transition probabilities of each state were compared between the patients and controls.

Results: We found patients were more likely to reside within a state of global coherence, in which there was BOLD phase alignment across all 90 regions, compared with healthy controls (59±3 vs. 39±5%, p=0.0004) (Figure 2). This global coherence state also had a significantly higher probability of occurring during periods where GSWs were present in comparison to the overall occurrence probability across all scanning sessions (70±8 vs. 48±3%, p=0.007).

Compared to healthy control participants, patients were less likely to reside in three states corresponding to spatially specific networks (labeled ventral attention, frontoparietal and visual), showing a reduced occurrence probability and duration (Figure 2).

Conclusions: Our study suggests that patients with generalized epilepsy are more frequently found to be in a functional brain state characterized by global coherence, while they visit ventral attention, frontoparietal, and visual states less. Taken together, the dynamic repertoire of functional connectivity states in patients is reduced compared to controls. Further work will investigate if these functional connectivity biomarkers can inform predictive models of seizure-onset.

References:
1. Cabral, J., et al., Cognitive performance in healthy older adults relates to spontaneous switching between states of functional connectivity during rest. Sci Rep, 2017. 7(1): p. 5135.

Funding: Please list any funding that was received in support of this abstract.: The study was funded by Medical Research Council (MR/K013998/1); Action Medical Research (P4646). Wellcome/EPSRC Centre for Medical Engineering [WT 203148/Z/16/Z].