Brian H Silverstein1, Eishi Asano2, Ayaka Sugiura3, Masaki Sonoda3, Min-Hee Lee4, Jeong-Won Jeong5. 1. Translational Neuroscience Program, Wayne State University, Detroit, MI, USA. 2. Translational Neuroscience Program, Wayne State University, Detroit, MI, USA; Dept. of Pediatrics, Wayne State University, Children's Hospital of Michigan, Detroit, MI, USA; Dept. of Neurology, Wayne State University, Children's Hospital of Michigan, Detroit, MI, USA. 3. Dept. of Pediatrics, Wayne State University, Children's Hospital of Michigan, Detroit, MI, USA. 4. Dept. of Pediatrics, Wayne State University, Children's Hospital of Michigan, Detroit, MI, USA; Translational Imaging Laboratory, Wayne State University, Detroit, MI, USA. 5. Translational Neuroscience Program, Wayne State University, Detroit, MI, USA; Dept. of Pediatrics, Wayne State University, Children's Hospital of Michigan, Detroit, MI, USA; Dept. of Neurology, Wayne State University, Children's Hospital of Michigan, Detroit, MI, USA; Translational Imaging Laboratory, Wayne State University, Detroit, MI, USA. Electronic address: jjeong@med.wayne.edu.
Abstract
INTRODUCTION: Cortico-cortical evoked potentials (CCEPs) are utilized to identify effective networks in the human brain. Following single-pulse electrical stimulation of cortical electrodes, evoked responses are recorded from distant cortical areas. A negative deflection (N1) which occurs 10-50 ms post-stimulus is considered to be a marker for direct cortico-cortical connectivity. However, with CCEPs alone it is not possible to observe the white matter pathways that conduct the signal or accurately predict N1 amplitude and latency at downstream recoding sites. Here, we develop a new approach, termed "dynamic tractography," which integrates CCEP data with diffusion-weighted imaging (DWI) data collected from the same patients. This innovative method allows greater insights into cortico-cortical networks than provided by each method alone and may improve the understanding of large-scale networks that support cognitive functions. The dynamic tractography model produces several fundamental hypotheses which we investigate: 1) DWI-based pathlength predicts N1 latency; 2) DWI-based pathlength negatively predicts N1 voltage; and 3) fractional anisotropy (FA) along the white matter path predicts N1 propagation velocity. METHODS: Twenty-three neurosurgical patients with drug-resistant epilepsy underwent both extraoperative CCEP recordings and preoperative DWI scans. Subdural grids of 3 mm diameter electrodes were used for stimulation and recording, with 98-128 eligible electrodes per patient. CCEPs were elicited by trains of 1 Hz stimuli with an intensity of 5 mA and recorded at a sample rate of 1 kHz. N1 peak and latency were defined as the maximum of a negative deflection within 10-50 ms post-stimulus with a z-score > 5 relative to baseline. Electrodes and DWI were coregistered to construct electrode connectomes for white matter quantification. RESULTS: Clinical variables (age, sex, number of anti-epileptic drugs, handedness, and stimulated hemisphere) did not correlate with the key outcome measures (N1 peak amplitude, latency, velocity, or DWI pathlength). All subjects and electrodes were therefore pooled into a group-level analysis to determine overall patterns. As hypothesized, DWI path length positively predicted N1 latency (R2 = 0.81, β = 1.51, p = 4.76e-16) and negatively predicted N1 voltage (R2 = 0.79, β = -0.094, p = 9.30e-15), while FA predicted N1 propagation velocity (R2 = 0.35, β = 48.0, p = 0.001). CONCLUSION: We have demonstrated that the strength and timing of the CCEP N1 is dependent on the properties of the underlying white matter network. Integrated CCEP and DWI visualization allows robust localization of intact axonal pathways which effectively interconnect eloquent cortex.
INTRODUCTION: Cortico-cortical evoked potentials (CCEPs) are utilized to identify effective networks in the human brain. Following single-pulse electrical stimulation of cortical electrodes, evoked responses are recorded from distant cortical areas. A negative deflection (N1) which occurs 10-50 ms post-stimulus is considered to be a marker for direct cortico-cortical connectivity. However, with CCEPs alone it is not possible to observe the white matter pathways that conduct the signal or accurately predict N1 amplitude and latency at downstream recoding sites. Here, we develop a new approach, termed "dynamic tractography," which integrates CCEP data with diffusion-weighted imaging (DWI) data collected from the same patients. This innovative method allows greater insights into cortico-cortical networks than provided by each method alone and may improve the understanding of large-scale networks that support cognitive functions. The dynamic tractography model produces several fundamental hypotheses which we investigate: 1) DWI-based pathlength predicts N1 latency; 2) DWI-based pathlength negatively predicts N1 voltage; and 3) fractional anisotropy (FA) along the white matter path predicts N1 propagation velocity. METHODS: Twenty-three neurosurgical patients with drug-resistant epilepsy underwent both extraoperative CCEP recordings and preoperative DWI scans. Subdural grids of 3 mm diameter electrodes were used for stimulation and recording, with 98-128 eligible electrodes per patient. CCEPs were elicited by trains of 1 Hz stimuli with an intensity of 5 mA and recorded at a sample rate of 1 kHz. N1 peak and latency were defined as the maximum of a negative deflection within 10-50 ms post-stimulus with a z-score > 5 relative to baseline. Electrodes and DWI were coregistered to construct electrode connectomes for white matter quantification. RESULTS: Clinical variables (age, sex, number of anti-epileptic drugs, handedness, and stimulated hemisphere) did not correlate with the key outcome measures (N1 peak amplitude, latency, velocity, or DWI pathlength). All subjects and electrodes were therefore pooled into a group-level analysis to determine overall patterns. As hypothesized, DWI path length positively predicted N1 latency (R2 = 0.81, β = 1.51, p = 4.76e-16) and negatively predicted N1 voltage (R2 = 0.79, β = -0.094, p = 9.30e-15), while FA predicted N1 propagation velocity (R2 = 0.35, β = 48.0, p = 0.001). CONCLUSION: We have demonstrated that the strength and timing of the CCEP N1 is dependent on the properties of the underlying white matter network. Integrated CCEP and DWI visualization allows robust localization of intact axonal pathways which effectively interconnect eloquent cortex.
Authors: Yijing Wu; Andrew L Alexander; John O Fleming; Ian D Duncan; Aaron S Field Journal: J Comput Assist Tomogr Date: 2006 Mar-Apr Impact factor: 1.826
Authors: Thomas J Whitford; Marek Kubicki; Shahab Ghorashi; Jason S Schneiderman; Kathryn J Hawley; Robert W McCarley; Martha E Shenton; Kevin M Spencer Journal: Neuroimage Date: 2010-10-25 Impact factor: 6.556
Authors: Eleonora Bartoli; William Bosking; Yvonne Chen; Ye Li; Sameer A Sheth; Michael S Beauchamp; Daniel Yoshor; Brett L Foster Journal: Curr Biol Date: 2019-10-03 Impact factor: 10.834
Authors: Jurgen Hebbink; Dorien van Blooijs; Geertjan Huiskamp; Frans S S Leijten; Stephan A van Gils; Hil G E Meijer Journal: Brain Topogr Date: 2018-12-06 Impact factor: 3.020
Authors: F M Zauli; M Del Vecchio; S Russo; V Mariani; V Pelliccia; P d'Orio; I Sartori; P Avanzini; F Caruana Journal: Philos Trans R Soc Lond B Biol Sci Date: 2022-09-21 Impact factor: 6.671
Authors: Ezequiel Gleichgerrcht; Adam S Greenblatt; Tanja S Kellermann; Nathan Rowland; W Alexander Vandergrift; Jonathan Edwards; Kathryn A Davis; Leonardo Bonilha Journal: Epilepsy Res Date: 2021-02-05 Impact factor: 3.045