CorNET: Deep Learning Framework for PPG-Based Heart Rate Estimation and Biometric Identification in Ambulant Environment.

Advancements in wireless sensor network technologies have enabled the proliferation of miniaturized body-worn sensors, capable of long-term pervasive biomedical signal monitoring. Remote cardiovascular monitoring has been one of the beneficiaries of this development, resulting in non-invasive, photoplethysmography (PPG) sensors being used in ambulatory settings. Wrist-worn PPG, although a popular alternative to electrocardiogram, suffers from motion artifacts inherent in daily life. Hence, in this paper, we present a novel deep learning framework (CorNET) to efficiently estimate heart rate (HR) information and perform biometric identification (BId) using only a wrist-worn, single-channel PPG signal collected in ambulant environment. We have formulated a completely personalized data-driven approach, using a four-layer deep neural network. Two convolution neural network layers are used in conjunction with two long short-term memory layers, followed by a dense output layer for modeling the temporal sequence inherent within the pulsatile signal representative of cardiac activity. The final dense layer is customized with respect to the application, functioning as: regression layer-having a single neuron to predict HR; classification layer-two neurons that identify a subject among a group. The proposed network was evaluated on the TROIKA dataset having 22 PPG records collected during various physical activities. We achieve a mean absolute error of 1.47 ± 3.37 beats per minute for HR estimation and an average accuracy of 96% for BId on 20 subjects. CorNET was further evaluated successfully in an ambulant use-case scenario with custom sensors for two subjects.

The impact of head movements on EEG and contact impedance: an adaptive filtering solution for motion artifact reduction.

Designing and developing a comfortable and convenient EEG system for daily usage that can provide reliable and robust EEG signal, encompasses a number of challenges. Among them, the most ambitious is the reduction of artifacts due to body movements. This paper studies the effect of head movement artifacts on the EEG signal and on the dry electrode-tissue impedance (ETI), monitored continuously using the imec's wireless EEG headset. We have shown that motion artifacts have huge impact on the EEG spectral content in the frequency range lower than 20 Hz. Coherence and spectral analysis revealed that ETI is not capable of describing disturbances at very low frequencies (below 2 Hz). Therefore, we devised a motion artifact reduction (MAR) method that uses a combination of a band-pass filtering and multi-channel adaptive filtering (AF), suitable for real-time MAR. This method was capable of substantially reducing artifacts produced by head movements.

High density wireless EEG prototype: Design and evaluation against reference equipment.

A high density wireless electroencephalographic (EEG) platform has been designed. It is able to record up to 64 EEG channels with electrode to tissue impedance (ETI) monitoring. The analog front-end is based on two kinds of low power ASICs implementing the active electrodes and the amplifier. A power efficient compression algorithm enables the use of continuous wireless transmission of data through Bluetooth for real-time monitoring with an overall power consumption of about 350 mW. EEG acquisitions on five subjects (one healthy subject and four patients suffering from epilepsy) have been recorded in parallel with a reference system commonly used in clinical practice and data of the wireless prototype and reference system have been processed with an automatic tool for seizure detection and localization. The false alarm rates (0.1-0.5 events per hour) are comparable between the two system and wireless prototype also detected the seizure correctly and allowed its localization.