Abstracts

Rationale: For patients with epilepsy, impaired responses during seizures severely reduce quality of life but are difficult to measure in a consistent, objective manner. Focal deficits can help localize the region of seizure onset, and global deficits in level of consciousness are important for determining seizure severity and in making decisions about driving safety. Previously, our studies validated a standardized testing battery called the responsiveness in epilepsy scale (RES), which prospectively evaluated patient cognition levels during seizures. The testing procedure relied on trained hospital personnel to monitor patients at bedside and to perform the RES testing during seizures. Because testing in this manner was naturally error-prone and subject to missing many seizures entirely, our revised aim was to automate the RES by means of playing a sequence of RES video clips triggered by computer-based seizure detections.Methods: As opposed to the standard RES, the computerized version called Robo-RES was created to capture more seizure events, to eliminate testing errors, and to reduce latency between RES questioning and seizure onset without sacrificing the comprehensive set of cognitive and sensorimotor functions of interest. This was accomplished by first recording videos of RES questioning then dynamically combining them into a PsychoPy experiment task. The goal was to allow patients to begin RES testing immediately after seizure onset by viewing videos on an all-in-one wall-mounted PC instead of waiting for a human tester to begin asking questions. The Robo-RES testing was conjointly initiated by a patient pushbutton event or by Persyst-12 automatic seizure detection, where the seizure detection was monitored bythe clinical video/EEG acquisition computer.Results: We tested the reliability of the system using simulated seizure events obtained through pre-recorded video/EEG data from 14 seizures in 4 patients. Automatic behavioral testing was initiated by the Robo-RES system for 100% (14/14) of seizures with a mean time lag of 2.1 s and no false-positive test initiations. The clinical recording system continued to function normally during simultaneous operation of the user interface and behavioral testing initiation.Conclusions: Fully automated behavioral testing during seizures is feasible using the Robo/RES system. Computer-controlled behavioral testing during the ictal and postictal periods can potentially greatly enhance the clinical information obtained during video/EEG monitoring. This approach may improve the evaluation of seizure severity, localization, and lateralization and further guide therapeutic interventions for people with epilepsy.