Abstracts

Rationale: Electroencephalography (EEG) has long been the gold standard for capturing and characterizing seizures, but the paroxysmal nature of these events presents various barriers long recognized by the clinician. Routine EEG is limited by the rarity of events captured during brief recordings, and more prolonged ambulatory EEG is often logistically impractical and technically challenging for monitoring more than a few days. Here we describe a novel discreet, non-invasive, easy-to-use, in-ear EEG device that enables long-term, remote EEG monitoring of the brain. The goal of the study was to identify the source localization of neural activity recorded by in-ear EEG compared to ground truth intracranial EEG (iEEG) recordings obtained via stereo-EEG (sEEG) depth electrodes and to evaluate the sensitivity for detecting seizures of different sites of onset and patterns of propagation. We hypothesized that the in-ear EEG would be most sensitive for detecting temporal lobe seizures given the proximity of the earbuds to the inferior and lateral temporal lobes.

Methods: We have recorded 530 hours of simultaneous iEEG and in-ear EEG in eight patients with medically-refractory epilepsy undergoing presurgical evaluation with sEEG in an epilepsy monitoring unit. We co-registered post-operative MRIs with pre-operative MRIs to generate three-dimensional cortical models of individual patients and to identify the precise locations and anatomic area of iEEG electrodes. Four independent neurologists visually inspected the aligned iEEG and in-ear EEG for comparative analysis.

Results: Results showed that 16 out of the 18 seizures detected with iEEG were also visible on the in-ear EEG, demonstrating an overall sensitivity of 89% (average across all patients=97 %, min=83 %, max=100%). The ability to detect seizures on the in-ear EEG indeed depended on the anatomical location of the seizure onset zone (SOZ) and propagation zone (PZ). All detected seizures had either a SOZ and/or a PZ in the lateral temporal lobe. The two seizures that were not detected on the in-ear EEG had a SOZ in the mesial temporal lobe and frontal lobe, respectively, and remained very focal. In addition to providing a high signal fidelity to clinicians, the device was also found to be comfortable and well-adopted by patients, with an average wear time of 88 hours, and a comfort rating of 8.3 out of 10 (1 = unbearable, 10 = very comfortable).

Conclusions: These preliminary results demonstrate the feasibility of using an in-ear wearable device to record long-term EEG and capture temporal lobe epilepsy. An unobtrusive yet reliable wearable device opens the door to routinely collect additional, long-term, remote electrographic evidence that may help clinicians make or clarify an epilepsy diagnosis, guide therapy to those diagnosed, and perhaps for the first time, provide objective seizure quantification over time. We now plan to expand our inclusion criteria to extra-temporal and primary generalized epilepsies, as well as to utilize machine learning algorithms to automate seizure and interictal epileptiform discharge quantification.

Funding: Please list any funding that was received in support of this abstract.: -