Authors :
Presenting Author: Helen Brinyark, BS – University of Alabama at Birmingham
Rebekah Chatfield, BS – The University of Alabama at Birmingham
Caila Coyne, MEng – University of Alabama at Birmingham
Dalia Ramirez, n/a – The University of Alabama at Birmingham
Arie Nakhmani, PhD – University of Alabama at Birmingham
Benjamin Cox, MD – University of Alabama at Birmingham
Rachel Smith, PhD, MS, BS – University of Alabama at Birmingham
Rationale:
Localization of the seizure onset zone (SOZ) is critical to surgically treat drug-resistant epilepsy. Methods of SOZ localization include analyzing the brain’s response to single-pulse electrical stimulation (SPES), where high-amplitude evoked responses indicate regions of pathological cortical hyperexcitability. Seizures evoked during SPES have prompted the question as to how the application of electrical stimulation to induce seizures could be used to circumvent the challenges that come from waiting passively for seizures to localize the SOZ, as this process can take weeks with some patients never experiencing a seizure during monitoring. The neural resonance hypothesis proposed by Smith et al. (2022) uses dynamical network models built from SPES-evoked responses to construct Sigma plots, which relate the magnitude of the response of the whole brain to a range of input stimulation frequencies. Peaks in the Sigma plot indicate global resonant frequencies at which electrical stimulation may be more likely to induce an epileptic seizure.
Methods:
In 10 patients who consented to undergo stimulation to induce seizures (SIS) at UAB, we stimulated at Sigma resonant frequencies and at standard frequencies (5, 10, 15, 20 Hz) and elicited seizures at both resonant and non-resonant frequencies (Figure 1). In 3 patients who experienced stimulation-induced seizures at a non-resonant frequency, we recalculated our models to construct Bode plots, which relate the local response magnitude of the brain to a range of input stimulation frequencies, for each electrode. In a single patient, we prospectively stimulated at SOZ-Bode resonant frequencies, e.g., local resonant frequencies that were common between Bode plots for SOZ electrodes, in addition to Sigma resonant frequencies and the standard frequencies.
Results:
We found that for 16 of 18 seizures and 18 of 26 after discharges induced at a “non-resonant” frequency, the stimulation frequency was a resonant frequency in a Bode plot belonging to an SOZ electrode (Figure 2). In our prospective evaluation, we more effectively elicited epileptic events when stimulating at SOZ Bode frequencies (8/17 stimulations) than when stimulating at Sigma frequencies (5/17 stimulations). Our retrospective analysis for this patient revealed that the event-inducing stimulation frequency for 11 of 16 induced after discharges and 5 of 8 induced seizures matched a resonant frequency in an SOZ Bode plot.
Conclusions:
These findings suggest that stimulation of the epileptic network at a resonant frequency relevant to the SOZ may better elicit seizures than stimulating at a global resonant frequency. Thus, neural resonance provides a way to guide SIS in targeted SOZ subnetworks, which could lead to improved epileptogenic network localization and overall surgical outcomes.
Funding:
This work was funded by the UAB School of Engineering, AES Junior Investigator Award 1042632, and the CURE Epilepsy Taking Flight Award 1061181.