Both Amplitude and Sleep Stage Can Influence the Association of High Frequency Oscillations with the Epileptogenic Zone
Abstract number :
1.19
Submission category :
3. Neurophysiology / 3G. Computational Analysis & Modeling of EEG
Year :
2021
Submission ID :
1826297
Source :
www.aesnet.org
Presentation date :
12/4/2021 12:00:00 PM
Published date :
Nov 22, 2021, 06:53 AM
Authors :
Kevin Tyner, PhD Candidate - University of Nebraska Medical Center; Timothy Fink - Gonzaga University; William Stacey - University of Michigan; Stephen Gliske - University of Nebraska Medical Center
Rationale: High Frequency Oscillations (HFOs) are a promising biomarker to help guide resective surgery in patients with refractory epilepsy, although both healthy and diseased brain can produce HFOs. Our objective was to identify qualitative features of HFO morphology which can disambiguate pathological from physiological HFOs, accounting for the confounding factor of state of vigilance.
Methods: We modified an existing computational model of state of vigilance to produce pathological and physiological HFOs in NREM, REM, and awake conditions. Based on the model, we hypothesized which features and states of vigilance would be most distinguishing. We then tested these hypotheses using logistic regression with an approximately three million HFOs collected from 31 patients over all states of vigilance.
Results: Computational modeling showed that HFO amplitude disambiguated normal from abnormal HFOs at all states of vigilance better than other tested features, with the highest AUROC value (0.81) during the awake state. In our clinical data, HFO amplitude was strongly associated with the seizure onset zone (SOZ), even when adjusting for state of vigilance and interaction terms (odds ratio 1.93). We also found that, after controlling for amplitude and interactions, HFO rate during awake state had a slightly higher association with SOZ. Here, p-values are not indicated as all events were highly significant.
Conclusions: The cortical model made accurate predictions about the state of vigilance, features of HFOs, and their association with SOZ. The data suggest that HFO amplitude can be used to discriminate pathological and physiological HFOs, regardless of sleep stage. The slight advantage of wake over the other sleep stages further suggest that sleep scoring may not be necessary for reliable and robust HFO interpretation.
Funding: Please list any funding that was received in support of this abstract.: National Institutes of Health (K01-ES026839 and R01-NS094399), and the Doris Duke Foundation (grant number 2015096).
Neurophysiology