Early Seizure Detection with an Energy-Efficient Convolutional Neural Network on an Implantable Microcontroller

Abstract

Implantable, closed-loop devices for automated early detection and stimulation of epileptic seizures are promising treatment options for patients with severe epilepsy that cannot be treated with traditional means. Most approaches for early seizure detection in the literature are, however, not optimized for implementation on ultra-low power microcontrollers required for long-term implantation. In this paper we present a convolutional neural network for the early detection of seizures from intracranial EEG signals, designed specifically for this purpose. In addition, we investigate approximations to comply with hardware limits while preserving accuracy. We compare our approach to three previously proposed convolutional neural networks and a feature-based SVM classifier with respect to detection accuracy, latency and computational needs. Evaluation is based on a comprehensive database with long-term EEG recordings. The proposed method outperforms the other detectors with a median sensitivity of 0.96, false detection rate of 10.1 per hour and median detection delay of 3.7 seconds, while being the only approach suited to be realized on a low power microcontroller due to its parsimonious use of computational and memory resources.

Machine learning from wristband sensor data for wearable, noninvasive seizure forecasting

“…[3][4][5][6][7] To this end, most studies have utilized either electrocorticography (ECoG) or scalp electroencephalography (EEG) as well as, to a lesser extent, electrocardiography (ECG), and have demonstrated that robust differentiation between preictal and interictal periods as well as early seizure detection is possible with a better-than-chance performance. [8][9][10][11][12][13] Furthermore, seizure forecasting has traditionally performed best when algorithms were trained or optimized at the individual patient level, which often required some sort of training or adjustment phase 8,14 prior to deployment.…”

A Recent Investigation on Detection and Classification of Epileptic Seizure Techniques Using EEG Signal

“…It also uses spectrogram short-term im-ages and achieved 100% accuracy. Other studies that used spectrograms include [153,158] and [176]. A 3D kernel of Wei et al [169] combined 2D images of an individual-channel EEG time series to obtain a 3D image.…”

Directions of Deep Brain Stimulation for Epilepsy and Parkinson’s Disease

“…Currently, most CNN algorithms are higher complexity and executed using CPU or GPUs. A network architecture called SeizureNet on a low-power processing microcontroller unit to predict seizure ( Hügle et al, 2018 ). There is another low-power CNN processor (TrueNorth developed by IBM) has been used for seizure detection ( Merolla et al, 2014 ).…”