Latencies from intracranial seizure onset to ictal tachycardia: A comparison to surface EEG patterns and other clinical signs

Abstract

SUMMARYObjective: Information on the relative timing of electroencephalography (EEG) seizure onset, ictal tachycardia (ITC), and first other clinical seizure manifestations is crucial for an understanding of the potential benefit of ITC-detection based closed-loop intervention systems for epilepsy treatment. This study analyzes the temporal relation of ITC, other clinical signs, and seizure onset in relation to intracranial and surface EEG. Methods: Seventy-eight seizures with ITC from 13 patients undergoing invasive EEG recordings with simultaneous recordings of electrocardiography (ECG), intracranial EEG (iEEG) and surface EEG, and video recordings to determine clinical onset were analyzed. Latencies for ITC were calculated for thresholds of 100 bpm and for a 20% heart rate increase above baseline obtained 60 s prior to seizure onset on iEEG. Patient-based, seizure-based, and seizure origin-based analyses were performed. Results: Mean latencies between seizure onset in invasive EEG and the following onset of ITC in the seizure-and patient-based analysis for both thresholds varied between 21.6 and 23.7 s, showing that ITC is an ictal rather a preictal phenomenon. In 10 of 13 patients and in 56 of 78 seizures, at least one of the thresholds for ITC was crossed before any other clinical sign. In the majority of cases, ITC also preceded ictal onset as determined in surface EEG. Latencies to ictal tachycardia were shorter for hippocampal than for extrahippocampal temporal seizure onset. ITC occurred earlier in right than in left temporal seizures. Significance: iEEG preceded other seizure manifestations, but ictal tachycardia was an early sign particularly in mesial temporal and in right temporal seizure onset and often preceded not only other clinical manifestations but also first visible patterns in surface EEG. Detection of ictal tachycardia was thus the best noninvasively assessed marker for closed-loop interventions in this multimodally assessed patient group.

Postictal immobility and generalized EEG suppression are associated with the severity of respiratory dysfunction

“…In patients at high risk of SUDEP, interventions aimed at reversing impaired arousal and PI may reduce SUDEP risk by noninvasive stimulation of the patient in the postictal period . Such interventions could be automated using closed‐loop systems and employing secondary markers for seizures or alternatively with direct stimulation of arousal circuits with responsive neurostimulation …”

Early experiences with tachycardia‐triggered vagus nerve stimulation using the AspireSR stimulator

“…In cases reported here, IT was detected after seizure onset on surface EEG. However, IT can potentially be detected prior to the EEG and the clinical seizure onset (Hirsch et al, 2015). Tachycardia-based detection may thus not only offer advantages in terms of ease of recordings and of computational analyses, as compared to EEG-based seizure detection, but may offer, for certain patients, advantages in the timing of a closed-loop intervention.…”

“…Ictal tachycardia (IT), a physiological change that the patient may not be aware of, can be the first detectable clinical manifestation or occur later during the course of the seizure (Parisi et al ., ; Nilsen et al ., ; Hirsch et al ., ). The prevalence of ictal tachycardia has been reported to be between 33% and 100% (Garcia et al ., ; Zijlmans et al ., ; Işik et al ., ).…”

“…Here, we report early experiences with ictal tachycardia-based automated stimulation of the vagus nerve to treat drugresistant focal epilepsy. Ictal tachycardia (IT), a physiological change that the patient may not be aware of, can be the first detectable clinical manifestation or occur later during the course of the seizure (Parisi et al, 2005;Nilsen et al, 2010;Hirsch et al, 2015). The prevalence of ictal tachycardia has been reported to be between 33% and 100% (Garcia et al, 2001;Zijlmans et al, 2002;Işik et al, 2012).…”

One-shot Learning for iEEG Seizure Detection Using End-to-end Binary Operations: Local Binary Patterns with Hyperdimensional Computing

“…For the majority of the patients (10 out of 16 in Table I), our algorithm quickly learns from one or two seizures, and achieves perfect (100%) specificity and sensitivity with k-fold cross-validation, where k is the total number of seizures minus the number of trained seizures. Our algorithm shows 18.2 s average delay in detection, which is well suited for several important applications considering that iEEG seizure onset often precedes clinical onset by more than 20 s [16]. Furthermore, a 20 s delay from seizure onset detection may in many cases be fast enough to quench seizure activity or to prevent further spreading and secondary generalization, which is the hallmark of disabling or even life-threatening seizures.…”