Multicenter clinical assessment of improved wearable multimodal convulsive seizure detectors

Abstract

Running title: Wearable multimodal motor seizure detectors Onorati et al. 2 Summary Objective:New devices are needed for monitoring seizures, especially those associated with sudden unexpected death in epilepsy (SUDEP). They must be unobtrusive, automated and provide false alarm rates bearable in everyday life. This study quantifies the performance of new multimodal wrist-worn convulsive seizure detectors. Methods:Hand-annotated video-electroencephalography seizure events were collected from 69 patients at 6 clinical sites. Three different wristbands were used to record electrodermal activity (EDA) and accelerometer (ACM) signals, obtaining 5,928 hours of data, including 55 convulsive epileptic seizures (6 focal tonic-clonic seizures and 49 focal-to-bilateral-tonicclonic seizures) from 22 patients. Recordings were analyzed off-line to train and test two new machine learning classifiers and a published EDA and ACM-based classifier. Moreover, wristband data were analyzed to estimate seizure-motion duration and autonomic responses. Results:The two novel classifiers consistently outperformed the previous detector. The most efficient (Classifier III) yielded sensitivity of 94.55%, and false alarm rate (FAR) of 0.2 events/day.No nocturnal seizures were missed. Most patients had less than 1 false alarm every 4 days with FAR below their seizure frequency. When increasing the sensitivity to 100% (no missed seizures) the FAR is up to 13 times lower than the previous detector. Furthermore, all detections occurred before the seizure ended, providing reasonable latency (median: 29.3 s, range: 14.8-151 s). Automatically estimated seizure durations were correlated with true durations, enabling reliable annotations. Finally, EDA measurements confirmed the presence of post-ictal autonomic dysfunction, exhibiting a significant rise in 73% of the convulsive seizures. Onorati et al. 3 Significance:The proposed multimodal wrist-worn convulsive seizure detectors provide seizure counts that are more accurate than previous automated detectors and typical patient self-reports, while maintaining a tolerable FAR for ambulatory monitoring. Furthermore, the multimodal system provides an objective description of motor behaviour and autonomic dysfunction, aimed at enriching seizure characterization, with potential utility for SUDEP warning.

Amygdala‐stimulation‐induced apnea is attention and nasal‐breathing dependent

“…This study demonstrates a role for the amygdala in voluntary respiratory control and allows for further study of this pathway in dysfunctional breathing states - not only in those with epilepsy but also infants with developmental abnormalities and in neurodegenerative disorders. It is our hope that understanding how we can prevent amygdala-stimulation-induced apnea will pave the way for research into potential avenues to explore for the prevention of SUDEP whether by more detailed examination into these neuronal pathways and alterations induced by seizures or by preventing SUDEP by more acute interventions such as wearable convulsive seizure detectors to aide in arousing patients as an early intervention 50. …”

Machine learning applications in epilepsy

“…One recent multicenter study of 69 patients combined wrist-worn accelerometer data with measurements of electrodermal activity (a reflection of sweat gland activation by the sympathetic nervous system), achieving a sensitivity of 94.6% and false-positive rate of 0.20 per day (AUC 94%), with a median detection latency of 29.3 seconds using an SVM classifier. 40 Milošević et al also applied SVM to data from extremity-worn accelerometers combined with surface electromyography (sEMG) recordings from both biceps, achieving a combined sensitivity of 90.9%, clinical latency of 10.5 seconds, and false-positive rate of 0.45 alarms per 12-hour recording in seven pediatric patients. Notably, the authors found 40%-45% of these false alarms to be of clinical interest, including for falls, EEG lead removal, or concerning behaviors.…”

Strategies for non‐EEG seizure detection and timing for alerting and interventions with tonic–clonic seizures

“…In addition, this study showed a near linear correlation between the surge in EDA and the duration of postictal generalized EEG suppression (PGES) following GTCS. The greater intensity of autonomic disturbances after convulsive seizures may partially explain the linkage between dysautonomia, PGES, and SUDEP . Combining EDA sensors with other biosignals, including ACM and PPG (heart rate), in wearable devices is a novel and appealing approach for ambulatory seizure detection.…”

“…The greater intensity of autonomic disturbances after convulsive seizures may partially explain the linkage between dysautonomia, PGES, and SUDEP. 21 Combining EDA sensors with other biosignals, including ACM and PPG (heart rate), in wearable devices is a novel and appealing approach for ambulatory seizure detection. Further research will be needed to develop algorithms capable of discriminating between sympathetic arousal caused by epileptic seizures and those common to daily activities, stress, anxiety, and other neuropsychiatric disorders (eg, psychogenic nonepileptic seizures [PNES].…”