Large scale brain network level excitation and inhibition imbalance in patients with epilepsy
Poster No:
446
Submission Type:
Abstract Submission
Authors:
Hui Chen1, Min Wang1, Yin Wang1, Zhoukang Wu1, Liangjiecheng Huang1, Mengyuan Liu1, Zhiqiang Zhang2, Xiaosong He1
Institutions:
1University of Science and Technology of China, Hefei, Anhui, 2Jinling Hospital, the First School of Clinical Medicine, Southern Medical University, Nanjing, Jiangsu
First Author:
Co-Author(s):
Zhiqiang Zhang
Jinling Hospital, the First School of Clinical Medicine, Southern Medical University
Nanjing, Jiangsu
Jinling Hospital, the First School of Clinical Medicine, Southern Medical University
Nanjing, Jiangsu
Introduction:
Epilepsy is a neurological disorder characterized by an enduring predisposition to generate epileptic seizures, which are abnormal neural discharges arising from disrupted excitation and inhibition (E:I) balance of the brain. Evidence from both surface and intracranial EEG support the notion that epilepsy is associated with E:I imbalance in epileptogenic regions [1]. Nonetheless, due to the known limits of both EEG techniques, whether and how large scale brain network level E:I balance are disrupted in patients with epilepsy remains largely unknown. A previous study has provided evidence that the E:I ratio can also be effectively evaluated through the resting-state BOLD signals, as the inverse of the time series' Hurst exponent (H) [2]. Such measure of E:I ratio is also associated with myelination and shapes the structure-function coupling of the brain, supporting its biological relevance [3]. Accordingly, here we employ resting-state fMRI (rsfMRI) to evaluate the E:I ratio of large scale brain networks among different types of epilepsy, encompassing both focal epilepsy such as temporal lobe epilepsy (TLE) and idiopathic generalized epilepsy (IGE).
Methods:
We enrolled 209 patients with epilepsy and 100 demographically matched healthy controls (HC) in this study. Prior to any surgery, all subjects underwent rsfMRI and T1 scans using a 3T MRI scanner. The patient group was subsequently devided into 3 subgroups: (1) patients with TLE with left-sided hippocampal sclerosis (LHS, 80 in total); (2) patients with TLE with right sided HS (RHS, 80 in total); and (3) patients wtih IGE (49 in total). RsfMRI data were preprocessed using fMRIPrep [4] and XCP-D [5], including slice-timing correction, head-motion correction, segmentation, coregistration, normalization, despiking, nuisance regression, interpolation, temporal filtering (0.01-0.08 Hz), and smoothing. We used the 200 parcel version of 7 network Schaefer atlas [6] and the 32 parcel version of the Melbourne subcortex atlas [7] to extract BOLD time-series for each subject. Each time series was modeled as multivariate fractionally integrated processes, and the corresponding Hurst exponent was estimated via the univariate maximum likelihood method and a discrete wavelet transform [2]. We then summarized the Hurst exponent by each functional network by hemisphere. We also summarized the subcortical regions into a subcortical network for each hemisphere, except for amygdala and hippocampus which were grouped into the limbic network [8]. Statistical inferences were made with permutation-based independent t-tests with FWE correction for multiple comparisions [9]. We applied 1000000 permutations for each test.
Results:
We found significant but yet selective differences between patients with epilepsy and HC. In specific, patients with LHS presented bilateral reduction of Hurst exponent in the limbic network (left: t=-3.15, p=0.022; right: t=-3.65, p=0.004) (Fig 1A, Fig 2A), while patients with RHS exhibits unilateral reduction of Hurst exponent in the right limbic network (t=-3.47, p=0.008), right control network (t=-3.25 , p=0.017), and right default mode network (t=-4.04, p=0.001) (Fig 1B, Fig 2B). Interestingly, after correction for multiple comparison, we did not observe any significant difference between patients with IGE and HC (|t|<1.62, p>0.661).
Conclusions:
At large scale brain network level, we found significant E:I imbalance in patients with focal epilepsy, but not in patients with generalized epilepsy. Such difference may be attributed to the more prominent structural abnormalites (i.e., HS) in the former cohort. Regardless, the revealed E:I imbalance largerly overlaps with known epileptogenic networks. There results may further our understanding of the neural mechanisms underlying the diagnosis and prognosis of epilepsy.
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism) 1
Novel Imaging Acquisition Methods:
BOLD fMRI 2
Keywords:
Epilepsy
FUNCTIONAL MRI
MRI
1|2Indicates the priority used for review
·Boxplot of Hurst exponent (H) in patients with epilepsy and healthy controls (HC).
·Comparison of Hurst exponents between patients and healthy controls.