Authors :
Presenting Author: Ke Xie, MSc – McGill University
Jessica Royer, PsyD – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Raul Rodriguez-Cruces, PhD – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Hans Auer, BSc – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Alexander Ngo, BSc – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Ella Sahlas, BSc – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Judy Chen, BSc – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Donna Cabalo, MSc – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Shahin Tavakol, MSc – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada; Birgit Frauscher, MD, PhD – Analytical Neurophysiology Laboratory, Montreal Neurological Institute and Hospital, McGill University, Montreal, QC, Canada; Andrea Bernasconi, MD – Neuroimaging of Epilepsy Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute and Hospital, McGill University, Montreal, QC, Canada; Neda Bernasconi, MD, PhD – Neuroimaging of Epilepsy Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute and Hospital, McGill University, Montreal, QC, Canada; Luis Concha, MD, PhD – Brain Connectivity Laboratory, Institute of Neurobiology, Universidad Nacional Autónoma de Mexico (UNAM), Mexico; Boris Bernhardt, PhD – Multimodal Imaging and Connectome Analysis Laboratory, McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
Rationale:
Excitation-inhibition (E:I) imbalance is theorized as a core pathophysiological mechanism of temporal lobe epilepsy (TLE). So far, this theory has mainly been supported by experimental work in non-human animals. Here, we set out to non-invasively elucidate the pattern of cortical E:I imbalance in TLE relative to controls, and further assess its impact on large-scale brain dynamics.
Methods:
The work involved two datasets: (1) the discovery dataset comprising 35 drug-resistant TLE patients and 40 age- and sex-matched healthy controls, (2) the replication dataset consisting of 25 drug-resistant TLE patients and 25 matched healthy controls. Node-wide Hurst exponent index (H-index), a proxy for the overall E:I ratio within that node, was estimated using a univariate maximum likelihood method and discrete wavelet transform by modeling the resting-state fMRI time series as multivariate fractionally integrated processes. We examined associations with structurally governed network dynamics via a network control theory (NCT) framework. In brief, NCT simulates functional dynamics from structural connectomes, and allows for quantifying average controllability at a nodal level (
Fig 2a). Surface-based linear models separately compared the H-index and average controllability in TLE relative to controls, while controlling for age and sex. To identify the relationship between H-index and average controllability changes, we calculated subject-specific mean H-index and average controllability within the same clusters and correlated them.
Results:
In controls, we observed a higher H-index (
i.e., lower E:I ratio) in parieto-occipital cortices and a lower H-index (
i.e., higher E:I ratio) in frontal and temporo-limbic cortices (
Fig 1a). TLE patients exhibited a significant reduction in whole-brain H-index compared to controls (Cohen’s
d = -0.78,
p < 0.001), reflecting an elevated E:I ratio. Surface-based analysis revealed marked decreases in H-index in bilateral temporal and frontal cortices in TLE (FDR < 0.05,
Fig 1b). Stratifying the topography into functional communities revealed pronounced effects in transmodal networks such as the default mode, frontoparietal, and attention networks (FDR < 0.05,
Fig 1c). Findings were replicated in an external dataset (whole-brain/significant clusters:
d = -0.64/-0.66,
p < 0.05;
Fig 1d). Considering brain dynamics, brain areas with high average controllability included the centro-parietal middle and superior frontal cortices in controls (
Fig 2b). Regional analysis revealed extensive alterations in average controllability in TLE, with increases in the prefrontal cortex (control+), and decreases in temporal and parieto-occipital cortices (control-,
Fig 2b). Notably, individual mean average controllability was closely correlated with the mean H-index (control+/-:
r = -0.35/0.28,
p = 0.001/0.008;
Fig 2c).
Conclusions:
Our findings demonstrate extensive E:I imbalance in TLE patients, and show a close association with altered network dynamics. These findings may help to pave the way for understanding how macroscale structural and functional imbalances may contribute to atypical temporo-limbic excitability in TLE.
Funding: Ke Xie (CSC: 202006070175)