Authors :
Presenting Author: Elliot Neal, MD – University of South Florida
Samantha Schimmel, BS – University of South Florida
Emma Dunn, BS – University of South Florida
Keaton Piper, MD – University of South Florida
Adam Alayli, BS – University of South Florida
Zeegan George, MD – University of South Florida
Gavin Lockard, BS – University of South Florida
Molly Monsour, BS – University of South Florida
Fernando Vale, MD – University of South Florida
Yarema Bezchlibnyk, MD PhD – University of South Florida
Rationale:
Resting state functional MRI (rsfMRI) maps brain networks to detect functional connectivity (FC). This aids in surgical planning for resective epilepsy surgery by localizing patient specific epileptogenic zones and predicting postoperative seizure freedom (SF). rsfMRI may be used with electroencephalogram (EEG) to localize regions with positively correlated patterns of activation (positive epileptogenic network, PEN) and regions with negatively correlated patterns of activation (negative epileptogenic network, NEN). Postoperative functional connectivity (FC) and its association with SF are poorly understood. Here, we use rsfMRI to examine network connectivity and postoperative SF in patients with temporal lobe epilepsy (TLE).
Methods:
Nine patients with TLE were included, all had rsfMRI at the pre-op, post-op, and one year post-op time periods. Of those patients, six were seizure free and three were not. The PEN and NEN were generated non-invasively in all patients as previously described using a computational model based on scalp EEG and rsfMRI. ROIs included were those which when sufficiently disconnected by surgery were associated with SF: precuneus, caudal anterior cingulate cortex (cACC), rostral anterior cingulate cortex (rACC), isthmus of anterior cingulate cortex (iACC), precentral gyrus (PC), and posterior cingulate cortex (PCC). Student T test was used for comparison of continuous variables with an alpha of 0.05 determining significant differences.
Results:
The PEN and NEN were successfully modeled at each time point and connectivity differences in the relevant ROIs were measured. As previously reported, the disconnection immediately after surgery at each ROI was significantly higher in the SF group compared to the NSF group, determined by an increase in NEN FC and decrease in PEN FC. At the one-year time point, there were three groups. First, the precuneus, iACC, and rACC all demonstrated a regression to the mean at 1 year; there was no significant difference in FC between the SF and NSF groups. Second, the PCC and cACC FC remained significantly more disconnected in the SF compared the NSF group, but not changed from the immediate post op period. Third, the precentral gyrus ROI showed a divergence; the FC difference was significantly increased at 1 year compared to the post op time period.
Conclusions:
In this study, we used rsfMRI and EEG to highlight the different patterns of network reorganization in SF and NSF patients. At 1 year, three group patterns emerged. The FC of the precuneus, iACC, and rACC no longer differed between SF and NSF patients, the immediate FC differences between SF and NSF in the PCC and cACC persisted, and the FC of the precentral gyrus diverged more between SF and NSF patients. This study is one of the first to describe the evolution of epileptogenic network after surgery and show that the network evolves differently when in the presence of uncontrolled seizures. Future studies will investigate the mechanism of the network rearrangement and the relationship to neuropsychological function over time.
Funding: None