Authors :
Presenting Author: Renee Proost, MD – University Hospital Leuven, KULeuven
Jaiver Macea, MD – KULeuven; Wim Van Paesschen, MD, PhD – University Hospital Leuven, KULeuven; Lieven Lagae, MD, PhD – University Hospital Leuven, KULeuven; Katrien Jansen, MD, PhD – University Hospital Leuven, KULeuven
Rationale:
This study investigates the performance of a multimodal wearable device for the detection of tonic seizures in a pediatric childhood epilepsy cohort, with a focus on Lennox-Gastaut patients.
Methods:
The Plug ‘n Patch system (Byteflies), a multimodal wearable device using the Sensor Dot and replaceable electrode adhesives, was used to detect tonic seizures parallel to prolonged video-electroencephalogram (EEG). Multiple biosignals were detected: behind-the-ear EEG, surface electromyogram (EMG), electrocardiogram (ECG) and accelerometer (ACC)/gyroscope (GYR). Seizures were blindly annotated by a neurologist. Seizure characteristics were described and performance was assessed by sensitivity, positive predictive value (PPV), F1 score and false alarm rate (FAR) per hour. Performance was compared to seizure diaries kept by the caretaker.
Results:
In 13 patients, 99 tonic seizures were detected. Epilepsy syndrome was defined as Lennox-Gastaut in seven out of 13 patients (54%), six patients had other forms of (developmental) epileptic encephalopathies or drug resistant epilepsy. All but one patient had an intellectual disability. We’ve observed that tonic seizures presented with a stereotypical sequence of changes in biosignals. There was usually a high amplitude wave on EEG followed by paroxysmal fast activity (PFA), which was frequently contaminated with muscle artifact on the adhesive signals. At the same time or milliseconds after, ACC and GYR signals showed a sudden movement and EMG showed muscle activation, which could be constant or increasing in amplitude. The deflection of the ACC signal was steep and short, followed by a cessation of further movement which allowed to discriminate from movement artifacts. Ictal tachycardia was present in 56% of available ECG signals. Overall sensitivity of the blind annotations was 41%, with a PPV of 9%, an F1 score of 14% and a median FAR per hour of 0.75. Performance increased to an F1 score of 66% for nightly seizures lasting at least 10 seconds (sensitivity 66%, PPV 66%) and 71% for nightly seizures lasting at least 20 seconds (sensitivity 62%, PPV 82%). For these seizures there were no false alarms in 10 out of 13 patients. Some technical issues may have influenced the results: in some patients there was loosening of the ECG or EMG adhesives and sometimes signals were lost for an amount of time and later regained. Sensitivity of seizure diaries reached a maximum of 52% for prolonged (≥ 20 seconds) nightly seizures, even though caretakers slept in the same room.
Conclusions:
We showed that it is feasible to use a multimodal wearable device with multiple adhesive sites in children with epilepsy and intellectual disability. Offline manual detection of tonic seizures outperformed seizure diaries for prolonged nightly seizures, although sensitivity did not reach satisfactory values for clinical practice due to signal quality issues. The recognition of seizure specific signatures using multiple modalities can help in development of automatic seizure detection algorithms.
Funding:
The study was supported by a Personalized Medicine Interdisciplinair Coöperatief Onderzoek (ICON, HBC.2019.2521) project of Vlaams Agentschap Innoveren en Ondernemen (VLAIO).