Rationale:
Children with benign epilepsy with centrotemporal spikes (BECTS) have a mild seizure disorder and moderate cognitive difficulties. Many posit that interictal epileptiform discharges (IEDs) arising from one or both motor cortices affect cognition by disrupting brain connectivity. A single paper describes wide-spread, bilateral increases in connectivity during epochs with IEDs compared to IED-free epochs and concludes that BECTS resembles a generalized epilepsy.
1
We sought to reproduce this finding using a different connectivity methodology – the weighted Phase Lag Index (wPLI) – which more stringently controls for volume conduction of the EEG signal.
Method:
We identified 35 children with a clinical history and EEG consistent with BECTS: 15 with left-hemispheric and 20 with right-hemispheric IEDs. EEGs were manually annotated with a goal of marking at least 20 IEDs in sleep and wakefulness. Chosen IEDs were at least 1 second away from neighboring IEDs and from other artifacts (i.e. eye blinks). EEG analysis was performed using the MNE-Python toolboxes.2,3 Two epoch types were defined: (1) pre-IED (700 to 200ms before the IED) and (2) IED (0 to 500ms during the IED). EEG was transformed using the Filter-Hilbert method and wPLI of the low beta frequency band (12-20Hz) was computed for the pre-IED and IED epochs. Specifically, we assessed connectivity between the central electrodes (C4 for right IEDs, C3 for left IEDs) and all other electrodes. The beta frequency band was chosen a priori as stable connectivity estimates can be achieved in shorter epochs, thus permitting the pre-IED and IED epochs to be temporally close. wPLI ranges between 0 (no connectivity) and 1 (perfect connectivity). Average pre-IED and IED connectivity were calculated for each patient and change in connectivity was assessed at the group level using paired t-tests. Significance was set at 0.025 to account for the two IED groups and after Bonferroni correction for multiple comparisons, changes with a p< 0.001 were considered significant.
Results:
The average age of patients at time of EEG was 8.5+/-2.5yrs and 30% were taking seizure medications. Please see Table 1 for connectivity results. During IED epochs, beta wPLI increased between the central electrodes and all other electrode with one exception: connectivity between C4-F4 and between C3-F3 decreased, though this was not significant. In children with right-hemispheric IEDs, connectivity between C4 and left hemispheric leads (Fp1, F3, F7, T5 and O1) increased significantly. In parallel, children with left hemispheric IEDs showed higher connectivity between C3 and right hemispheric (Fp2, F4, F8, T4, T6, P4, O2) and vertex leads (Fz, Cz, Pz).
Conclusion:
After controlling for volume conduction, IEDs in BECTS lead to diffusely increased connectivity. Laterality of IEDs matters, with focal IEDs leading to larger increases in connectivity with the contralateral vs. within the ipsilateral hemisphere.
Funding:
:F. Baumer is supported by a K23 Career Development Award (NINDS K23 NS116110). F. Baumer is also grateful for a generous gift made by the O'Farrell-Principe Family to the Pediatric Epilepsy Research Fund which has supported B. Goad's efforts. C. Lee-Messer is supported by the Wu Tsai Neurotranslate grant.
FIGURES
Figure 1