Toward an Effective Brain Connectome Using Cortico-cortical Evoked Potentials
Abstract number :
1.177
Submission category :
3. Neurophysiology / 3E. Brain Stimulation
Year :
2022
Submission ID :
2203983
Source :
www.aesnet.org
Presentation date :
12/3/2022 12:00:00 PM
Published date :
Nov 22, 2022, 05:22 AM
Authors :
Hossein Shahabi, PhD – University of Southern California; Kenneth Taylor, PhD – Charles Shor Epilepsy Center, Neurological Institute, Cleveland Clinic, Cleveland, OH, USA; Anand Joshi, PhD – Signal and Image Processing Institute, University of Southern California, Los Angeles, CA, USA; Dileep Nair, MD – Charles Shor Epilepsy Center, Neurological Institute, Cleveland Clinic, Cleveland, OH, USA; Richard Leahy, PhD – Signal and Image Processing Institute, University of Southern California, Los Angeles, CA, USA
Rationale: Cortico-cortical evoked potentials (CCEPs) are induced by single-pulse electrical stimulation (SPES). CCEPs can map brain connectivity and localize the epileptogenic zone. Previous studies had shown that outward connectivity, as measured by CCEPs amplitude, is larger relative to a normal baseline when ictal onset zones are stimulated. However, full interpretation of CCEPs requires extensive knowledge regarding the physiology of the stimulated site, epilepsy type, and pathology of the epileptogenic zone (EZ), which is unavailable for many patients. Additionally, electrode implantations are heterogeneous and only cover a portion of the brain. We aim to address these challenges by constructing an effective connectivity atlas (connectome) using population data. By establishing normative brain connectivity values across the population from CCEPs data outside the epileptic network, we can then identify epilepsy-affected connections in individual patients.
Methods: We analyzed CCEPs in 38 patients who underwent presurgical evaluation using Stereoelectroencephalography (SEEG). For each stimulation site, we computed outward connectivity as the root-mean-square (RMS) value of averaged CCEP responses 15 to 300 ms after the stimulation. To reduce inter-subject variability, connectivity matrices were normalized to an individual baseline for each patient. Next, we mapped adjacency matrices from electrode space to the USCBrain atlas with 208 cortical and subcortical regions of interest (ROIs). We developed a nonlinear optimization algorithm to estimate and distinguish effective connectivity in non-epileptic regions of brain (‘baseline’ connectome) from connections in the epileptic network in each patient. We calculated baseline connectome values by averaging all relevant connections for each connectivity pair identified as non-epileptic. Figure 1 depicts the proposed framework.
Results: We observed a denser network for the left hemisphere, indicating more patients were initially identified with left hemisphere epilepsy. The right hippocampus was identified as the region with the largest outward connectivity in the baseline connectome, consistent with previous studies. The left superior temporal gyrus, left paracentral lobule, right hippocampus, right inferior temporal gyrus, and right amygdala were among the top five ROIs with the highest chance that their connections were identified as abnormal. This observation was in alignment with the fact that in a significant number of patients in our study, those regions were chosen as presumptive EZ and were resected. Finally, the anatomical locations of identified epilepsy-affected connections were in line with the resected area for six out of seven patients with seizure freedom after surgery.
Conclusions: We have proposed a data-driven approach to labeling connections between anatomical regions based purely on their CCEPs measurements. The resulting connectome may serve as an additional tool to help guide clinicians in identification of the EZ for improved outcomes.
Funding: This research was supported in part by the National Institutes of Health under awards R01NS089212, U01EB023820, and R01EB026299.
Neurophysiology