Quantifying a frequency modulation response biomarker in responsive neurostimulation

Abstract

Objective: Responsive Neurostimulation (RNS) is an effective treatment for controlling seizures in patients with drug-resistant focal epilepsy who are not suitable candidates for resection surgery. A lack of tools for detecting and characterizing potential response biomarkers, however, contributes to a limited understanding of mechanisms by which RNS improves seizure control. We developed a method to quantify ictal frequency modulation, previously identified as a biomarker of clinical responsiveness to RNS.Approach: Frequency modulation is characterized by shifts in power across spectral bands during ictal events, over several months of neurostimulation. This effect was quantified by partitioning each seizure pattern into segments with distinct spectral content and measuring the extent change from the baseline distribution of spectral content using the squared Earthmover's distance. Main results:We analyzed intracranial electroencephalography data from 13 patients who received RNS therapy, six of whom exhibited frequency modulation on expert evaluation. Subjects in the frequency modulation group had, on average, significantly larger and more sustained changes in their Earthmover's distances (mean = 13.97×10 -3 ± 1.197×10 -3 ). In contrast, those subjects without expert-identified frequency modulation exhibited statistically insignificant or negligible distances (mean = 4.994×10 -3 ± 0.732×10 -3 ).Significance: This method is the first step towards a quantitative, feedback-driven system for systematically optimizing RNS stimulation parameters, with an ultimate goal of truly personalized closed-loop therapy for epilepsy.

Deep net detection and onset prediction of electrographic seizure patterns in responsive neurostimulation

“…iESPnet is a multitasking artificial network that jointly makes iESP detection and onset time estimation. Modulation patterns in the power spectra domain have been observed to evolve at the beginning of an electrographic seizure in iEEG recordings from patients responding to the RNS therapy 11,38 . Thus, the assessment of iESP onset time by automatic methodologies is a necessary step toward automated identification of electrographic seizure pattern modulation biomarkers.…”

“…Lead implantation sites varied between patients and were grouped as Neocortex, Hippocampus, Neocortex‐Hippocampus, or Cortical Malformation (eTable 1). Each recording was a 60–180‐s segment of four‐channel iEEG, online band‐pass filtered at 4–125 Hz, and sampled at 250 Hz 11 . The iESP annotation process replicated the clinical workflow of specialized epilepsy centers.…”

“…Each recording was a 60-180-s segment of four-channel iEEG, online bandpass filtered at 4-125 Hz, and sampled at 250 Hz. 11 The iESP annotation process replicated the clinical workflow of specialized epilepsy centers. iESPs and their onsets were initially visually identified and annotated by an experienced epilepsy surgery neurophysiologist (V.K.)…”

“…Modulation patterns in the power spectra domain have been observed to evolve at the beginning of an electrographic seizure in iEEG recordings from patients responding to the RNS therapy. 11,38 Thus, the assessment of iESP onset time by automatic methodologies is a necessary step toward automated identification of electrographic seizure pattern modulation biomarkers. Precise seizure onset estimation can contribute toward the accurate identification of these biomarkers, and our approach resulted in the lowest onset time error estimation currently found in the literature (Table 1).…”

Circadian changes in aperiodic activity are correlated with seizure reduction in patients with mesial temporal lobe epilepsy treated with responsive neurostimulation

“…The RNS System (NeuroPace, Inc.) is uniquely positioned to utilize an electrographic biomarker from the long‐term intracranial recordings stored by the device. Thus far, a few potential biomarkers of the clinical response to RNS have been identified: a decrease in interictal spike frequency, 10 ictal frequency modulation, 11,12 and frequency‐dependent reorganization of interictal functional connectivity 13 . In addition, preimplant biomarkers may prognosticate eventual response; for example, both intracranial network synchronizability from stereoelectroencephalography recordings 14 and global functional connectivity from magnetoencephalography 15 have been associated with clinical outcomes.…”