P18-T Surface EMG signals indicate seizure severity: Correlation with postictal generalized EEG suppression in convulsive seizures

Abstract

Background Postictal generalized suppression (PGES) is a surrogate marker for seizure severity and the risk of sudden unexpected death in epilepsy patients (SUDEP). However, measuring PGES requires EEG recordings, which are not feasible on long term in an outpatient setting. There is a need for determining seizure severity using modalities implemented into wearable devices. Material and methods We studied 42 patients with 79 major convulsive seizures. Duration of the seizures and their phases (tonic-onset, tonic-maintenance and clonic) was measured automatically using surface EMG signals from the deltoid muscles. We compared them with PGES duration determined according to previously published criteria. We noted the semiological types of generalized seizures, presence of oral tonicity, occurrence of seizure during sleep or wakefulness. Results Automatic EMG detection of the deltoid muscle identified the generalized tonic-clonic seizures (GTCS) with an accuracy of 86.49%. PGES duration correlated with oral tonicity ( ρ = 0.034), tonic-maintenance phase (p = 0.009), clonic phase (p = 0.018) and total seizure (p = 0.021) durations, sleep enhancing these relationships. Seizures with oral tonicity had higher means of durations of PGES (p = 0.028) and clonic phase (p = 0.005). PGES ⩾ 20 s was associated with seizures type 1 ( ρ = 0.040), maintenance phase ⩾ 10 s ( ρ = 0.001) and clonic phase ⩾ 33s ( ρ = 0.021). We devised an algorithm that identified PGES ⩾ 20 s with a sensitivity of 100% using maintenance and clonic phase durations. Conclusion PGES, oral tonicity and seizures type 1 are closely related. Automatic, EMG-based measurement of seizure phases can accurately identify GTCS. Measurement of the maintenance and clonic phases can help predict dangerous PGES duration ⩾ 20 s.