Dr Terence O'BrienTerence J. O'Brien, MBBS, MD
Automated systems and devices may have the ability to provide a wearable, out-of-hospital seizure diagnostic monitoring method for patients with epilepsy, suggest findings from a recent study.

Data showed that there was no significant difference between classification results, whether the diagnosis was from the automated method or standard inpatient video electroencephalography (EEG) monitoring. Particular to the classification results, McNemar's test produced an exact P‐value of 0.25.

The automated system detected epileptic seizures (ES) and psychogenic non‐epileptic seizures (PNES) with a sensitivity of 72.7% and specificity of 100%. The positive (PPV) and negative predictive (NPV) values for PNES classification were 81.3% and 100%, respectively.

The investigators, led by Terence J. O'Brien, MBBS, MD, head, Departments of Neuroscience and Medicine, and deputy head of school, Central Clinical School, Monash University; program director, Alfred Brain; and  director of neurology, and deputy director of research, Alfred Health, noted that “these results demonstrate the potential utility of this automated, ambulatory, non‐invasive wearable device in differentiating between ES and PNES.”

Currently, 2 devices with this capability—Empatica’s Embrace smartwatch for seizure tracking and epilepsy management, and Brain Sentinel’s SPEAC system, electromyography (EMG) device for monitoring convulsive seizures—are approved by the FDA. The device utilized in this recent trial was an Apple iPod Touch (4th generation), with a built-in micro-electro-mechanical system (MEMS) accelerometer.

The investigators acknowledged that “the clinical utility of the approach presented in this report may be further enhanced with additional sensors as with devices such as Empatica and Brain Sentinel.”

In total, the study included 11 patients during inpatient video EEG, who experienced 24 convulsive seizures consisting of ≥20 seconds of sustained activity, narrowed down to 13 PNES from 5 patients and 11 ES from 6 patients. There were more than 661 hours, with an average of 2.4 false alarms per day (67 total). Data were based on a 20-second threshold for the automated system.

The results of the study also showed that of the 11 epileptic events detected, 8 were correctly classified as ES. The positive and negative likelihood ratios of classifying PNES were respectively 3.67 and 0. “As there were no patients in this study with both epilepsy and PNES, the likelihood ratios were not calculated for such cases and are thus not applicable to them,” O'Brien and colleagues wrote.

When the investigators adjusted the time threshold to 5 seconds, the algorithm correctly classified 12 of the 19 ES events (63.2% sensitivity) and 28 of the 33 PNES events (84.8% sensitivity). Compared to the 20-second threshold, the ES detection rate dropped by 30.4% while the PNES detection rate remained at 100%. The false alarm rate increased to 5.44 per day (150 false alarms in 661.58 hours).

“In the past decade, such wearable systems have garnered a lot of attention as they present a potential alternative for continuous non‐invasive monitoring in [a] patient-home setting. However, the utility and requirements of such systems from [a] clinical perspective are rarely considered,” O’Brien and colleagues detailed. “The findings of this study suggest that automated detection and differentiation of convulsive epileptic and non‐epileptic seizures can be done using accelerometer‐based wrist‐worn wearable systems.”

The investigators did admit some limitations, such as the sample size. They also noted that there are some points to be considered during the development of such seizure algorithms, including the importance of high differentiation specificity for convulsive ES events, as misclassification can result in misdiagnosis, and the minimization of the trade‐off between performance and minimum seizure duration, as some patients have briefer events.
REFERENCE
Naganur VD, Kusmakar S, Chen Z, Palaniswami MS, Kwan P, O’Brien TJ. The utility of an automated and ambulatory device for detecting and differentiating epileptic and psychogenic non‐epileptic seizures. Published online May 3, 2019. doi: 10.1002/epi4.12327.