Epileptogenic Zone Fingerprint in RNS ECoG Recordings
Abstract number :
3.206
Submission category :
3. Neurophysiology / 3C. Other Clinical EEG
Year :
2025
Submission ID :
450
Source :
www.aesnet.org
Presentation date :
12/8/2025 12:00:00 AM
Published date :
Authors :
Presenting Author: Mohammad Alisali, MD – University of Texas Health Science Center at Houston
Vladimir Vashin, BS – UTHealth Houston
Mabel Marilyn Dinane Devapriame, MD – University of Texas Health Science Center at Houston
John Mosher, PhD – University of Texas Health Science Center at Houston
Johnson Hampson, MBBE – UTHealth Houston
Abdulrahman Alwaki, MD – Department of Neurology, McGovern Medical School, The University of Texas Health Science Center at Houston, Houston, TX, USA
Yash Vakilna, MS – University of Texas Health Science Center at Houston
Jay Gavvala, MD – UTHealth Houston
Rationale: Grinenko et al. (2018) identified a characteristic fingerprint of the epileptogenic zone in time-frequency plots of SEEG data from patients who underwent successful epilepsy surgery. This fingerprint includes: (1) sharp transients or spikes preceding (2) multiband fast activity, occurring concurrently with (3) suppression of lower frequencies. Responsive neurostimulation (RNS) records intracranial EEG and delivers stimulation in response to ictal patterns. A fundamental assumption of RNS therapy is the accurate placement of leads in the epileptogenic zone. However, the presence of a seizure fingerprint in RNS ECoG data has not been previously investigated.
Methods: We screened the PDMS database for seizures with low-voltage fast activity (LVFA) onset occurring within the first three months post-implantation, to avoid stimulation artifacts. Twenty patients met these criteria and had available raw RNS data. For analysis, we selected two seizures per patient (except for two patients, where only one suitable seizure was identified). To remove the DC component, we computed the first derivative of each ECoG time series. Time-frequency decomposition was performed using a complex Morlet wavelet transform, implemented via the morlet_transform function in the Brainstorm software suite (Tadel et al., 2011). Two reviewers independently examined the time-frequency plots and scored the presence of each fingerprint component based on the criteria described by Grinenko et al. Final ratings were determined by consensus between the initial reviewers, with a third reviewer resolving any disagreements.
Results: A total of 38 seizures from 20 patients were analyzed, including two seizures per patient in 18 patients and one in the remaining two. Twenty-six seizures (68%) exhibited all three components of the seizure fingerprint (Figure 1). Among the 18 patients with two seizures analyzed, five showed the fingerprint in one seizure but not the other; the remainder showed consistent findings (i.e., fingerprint present in both or absent in both).
Conclusions: We demonstrate, for the first time, the presence of a seizure fingerprint in RNS ECoG recordings. Further investigation is warranted to determine whether the presence of this fingerprint correlates with clinical outcomes.
Funding: None
Neurophysiology