Anti‐Seizure Medications Alter Functional and Effective Connectivity as Measured With Intracranial Electroencephalography

BACKGROUND: Anti-seizure medications (ASMs) control seizures through distinct neuronal mechanisms. While suppressing seizures, putative spill-over effects can lead to multiple cognitive side effects, some of which can be explained by the modulation of brain connectivity. This modulation can be studied with functional and effective connectivity. While functional connectivity provides information on statistical dependencies between two brain regions, effective connectivity as measured with cortico-cortical evoked potentials (CCEPs) through single pulse electrical stimulation (SPES) can elucidate the underlying mechanisms of the effective connectivity with high temporal and spatial resolution. CCEPs are often performed while patients are on ASMs; however, the effect of ASMs on CCEPs is largely unknown given limited opportunity to experimentally study this in the human brain. We hypothesized that ASMs would alter the connectivity within the seizure network more often than in healthy brain areas. OBJECTIVE: We aimed to understand the effects of ASMs on functional and effective connectivity. METHODS: We recruited seven patients undergoing invasive monitoring with depth electrodes for medically refractory epilepsy, five of whom underwent SPES with 5 mA, 150 µsec per phase square wave pulses at 1 Hz frequency while the patients were on ASMs as part of ongoing research projects (ASM-ON). Since patients did not have their typical seizures, a second SPES session was performed by the clinical team while off ASMs to reduce their seizure threshold (ASM-OFF). We recorded a total of 565 bipolar channels and electrically stimulated a total of 17 seizure onset zone (SOZ), 6 early propagation zone (EPZ), 36 irritative zone (IZ), and 64 non-involved zone (NIZ) electrode pairs across five patients. We compared the amplitude and latency of early and late voltage deflections (N1, N2) and root mean square values of CCEPs between two sessions. In a partly overlapping cohort of five patients, we also recorded 1-h long rest sessions (ASM-ON and ASM-OFF) to show the changes in functional connectivity and graph theoretical measures obtained from the slow fluctuations in broadband high frequency activity. RESULTS: ASMs preferentially modulated excitability within the epileptic network, with the highest rates of significant CCEP amplitude changes observed for seizure network (SOZ→SOZ: 15.9% N1, 15.5% N2; EPZ→EPZ: 15.6% N1, 11.1% N2; IZ→IZ: 11.6% N1, 12.4% N2) compared to NIZ→NIZ (5.7% N1, 7.4% N2), and with amplitude effects consistently exceeding latency effects across all tissue class combinations (p < 0.05, group significance at q < 0.05). We also found that the RMS based connectivity and functional connectivity were altered more often outside of the seizure network. CONCLUSIONS: Overall, we found that ASMs altered the effective connectivity within the seizure network more than within non-involved regions, whereas functional connectivity was altered more often within non-involved regions. This study is the first study revealing with high spatiotemporal resolution that ASMs can alter brain effective and functional connectivity in multiple different ways.