Raising healthcare costs and new demographic, societal and health trends are pressing for a change in healthcare systems. Lifestyles are shifting towards busier time schedules with little motivation and time left for health management. Healthcare needs are transitioning from episodic to chronic. Population is aging. This calls for radical changes in how care will be provided, targeting preventive care, effective provision of continuous treatment, personalized and connected health.
Body area networks (BANs) are seen as a key enabler for this change, see fig. 1. Such a BAN provides medical, lifestyle, wellness, assisted living, sports and entertainment functions for the user. It comprises a series of miniature sensor nodes, implanted or located at the body surface. Each node has its own energy supply, consisting of storage and energy harvesting devices. Each node has enough intelligence to carry out its task. Furthermore, each node is able to communicate with other sensor nodes or with a gateway node worn on the body. The gateway node communicates with the outside world using a standard telecommunication infrastructure such as a wireless local area or cellular phone network. On the other extremity of the network, experts then provide services to the individual wearing the BAN. Intelligent or expert systems further include data fusion algorithms for the aggregation of body sensor data into metrics quantifying an individual's health status, his physical, cognitive and emotional state. Next generation of BANs will include feedback loops for health, performance or stress management.
Figure 1: Technology vision for 2015: people will be carrying their body area network providing health, lifestyle and entertainment functions to the user.
Early deployment of technology in different application cases are translated into critical technology obstacles that need to be solved in order to enable a widespread deployment of BANs. Imec is combining its research achievements and breakthroughs in integrated prototypes of wireless health and wellness monitors, which can be evaluated in real application environments. Some of the latest prototypes are highlighted below.
Smart ECG necklace
Imec's electrocardiography (ECG) necklace targets cardiac activity monitoring in every-day life situations. It measures a bipolar ECG signal between two Ag/AgCl electrodes attached to the body, and connected to the necklace using standard leadware. Low-power and high performance ECG monitoring is achieved through the use of a proprietary single channel application-specific integrated circuit (ASIC) for biopotential read-out. The system features a commercial low-power microcontroller and low-power radio, providing embedded processing capabilities and wireless communication within 10m range. Alternatively, data can be stored on a secure digital (SD) card for applications in which wireless connectivity is not required. The ECG necklace achieves 1 week autonomy on a 165mAh Li-ion battery, while continuously streaming ECG data, see fig. 2 and fig. 3a+b.
Figure 2: Smart ECG necklace for ambulatory cardiac monitoring; the system embeds a beat detection algorithm for robust and accurate beat detection.
Figure 3a: True ambulatory monitoring using imec's ECG necklace.
Figure 3b: True ambulatory monitoring using imec's ECG necklace.[cap(SR1106F4)] Wireless 8-channel EEG monitoring prototype integrated in a headband; standard din32 connectors can be used for connection to the system.
An optimized beat detection algorithm is implemented in the necklace. Based on continuous wavelet transform, the algorithm has been optimized for high accuracy on resource-constrained embedded systems. The optimized algorithm achieves a sensitivity of 99.65% and positive predictivity of 99.79% on the MIT/BIH arrhythmia database, and maintains its performances under levels of physical activity corresponding to daily life activities. In particular, positive predictivity of the optimized algorithm was shown to increase by 14% at low signal-to-noise ratio (SNR) (around 1dB) when compared to the initial algorithm. The optimized algorithm also shows very good accuracy in time, with 99.9% of the beats detected with perfect accuracy. The robust and accurate beat detection algorithm is particularly suited for applications requiring reliable and precise cardiac rhythm or heart rate variability analysis.
The main advantages of the smart ECG necklace are its low power consumption, its embedded beat detection algorithm and the performance of its integrated bio-potential ASIC. Up-coming pilot studies will evaluate the technology for use in various applications including arrhythmia detection, stress monitoring and epilepsy monitoring.
Wearable EEG headset
The wireless 8-channel electronencephalography (EEG) system enables ambulatory EEG monitoring. The system measures 8 EEG signals in a referential configuration, with the reference usually placed at the mastoid. The system relies on an ultra-low-power ASIC for the acquisition of the EEG signal, characterized by a high common mode rejection ratio (CMRR) (120dB) and low noise (60nV/sqrt(Hz)). The ASIC features an on-chip low-power ADC (11 bits), calibration and electrode impedance measurement nodes, and consumes only 200μW. In addition, the integrated EEG monitoring system includes a low-power microcontroller and radio providing local processing and wireless communication functionalities. The overall system consumes less than 10mW when sampling and streaming the data continuously at 1KHz. The system is integrated in a box of total size 50x30x10mm3, easily packaged in an elastic headband, see fig. 4.
An important burden in remote and ambulatory EEG monitoring is the need to set up all the electrodes for an accurate recording. Many applications would benefit from an EEG headset that can easily be set up and worn by the user. In 2009, Holst Centre and imec have introduced a first prototype of EEG headset, attempting to provide an easy-to-set-up device for measuring EEG, see fig. 5. The headset consists of 10 electrodes in total, placed according to the International 10-20 standard. Commercial EL120 reusable electrodes are used, for their special contact posts designed for use through hair and fur. These electrodes are Ag/AgCl coated contact resistive electrodes. These electrodes are embedded into specially designed electrode housing, connecting the electrode to the headset, providing mechanical tilt, and ensuring signal transmission.
Figure 5: Electrode set-up is an important burden to ambulatory and remote EEG monitoring: imec's wireless EEG headset as a first attempt to easy-to-set-up EEG monitoring.
The EEG headset has been shown to successfully record alpha waves in people at rest or slow motion, providing that the headset nicely fits the shape of the subject's head. Technology evaluation also points to important remaining challenges for achieving ambulatory EEG recording. Future work will focus on compensating motion and physiological artifacts in EEG recordings, improving dry electrodes and corresponding analog read-out circuitry, and developing ultra-low-power processor for EEG processing.
Listening to your brain wave?
Staalhemel is a work of art by Christoph De Boeck (°1972, Belgium) and a joint initiative between Holst Centre, imec and the art centre STUK in Leuven, Belgium. Staalhemel is composed of 80 steel segments suspended over the visitors' heads. Visitors wear the wireless EEG headset system, monitoring their brain signals. As they walk through the space, tiny hammers tap rhythmic patterns on the steel plates, activated by their brainwaves. This responsive environment confronts visitors with an acoustic representation of their electrical brain activity (www.staalhemel.com), see fig. 6.
Figure 6: Staalhemel: acoustic representation of brain waves.
Wireless sensors for ambulatory gait analysis
In US and Europe, one in three adults over 65 years old falls each year. And of those, 20% to 30% suffer moderate to severe injuries that make it hard for them to get around or live independently. The European project SMILING (7th Framework Program, grant agreement n. 215493) RP180 targets the improvement of mobility in the elderly by counteracting falls. The main risk factor for falls in the elderly is gait and balance disorders. Current gait analysis systems require invasive optical equipment, and can only be performed in hospital environment. Within the SMILING project, Holst Centre and imec have developed a wireless sensor platform which allows ambulatory gait analysis, a so-called wireless six-dimensional inertial measurement unit platform (6D-IMU).
The platform relies on wireless 6D-IMU modules, see figs. 7 and 8. Each module features a 3D-gyroscope and a 3D-accelerometer, and carries enough processing power to extract relevant information in real-time. The data is wirelessly transmitted to a receiving unit within 10m range, or may be stored in local memory. The 3D-accelerometer consists of a small, thin, low-power, complete 3-axis accelerometer with signal conditioned voltage outputs from analog devices (ADXL330). The sensor measures acceleration with a minimum full-scale range of ± 3g. Rate of turn is measured using a module with three ADXRS610 sensors mounted in the three perpendicular planes. Roll and yaw are set to a sensitivity of 300deg/s and pitch is set to a sensitivity of 800deg/s.
Figure 7: Wireless 6D-IMU sensor for ambulatory gait analysis.
Figure 8: Wireless 6D-IMU sensor for ambulatory gait analysis.
The wireless gait analysis platform has been evaluated in a clinical gait study. The study aimed at evaluating the practical use of the wireless gait analysis platform and its performance in monitoring wirelessly kinematic data from two feet. 20 subjects were involved. In average, the system achieved 1.2% frame losses, which is considered as satisfactory to perform off-line analysis of the data. SMILING project partners are developing algorithms to exploit the 6D-IMU data for portable gait analysis.
Overall, the system was well received by patients and professionals, enhancing comfort and ease of use compared to existing systems. Wireless, miniaturized and wearable, the proposed system opens new perspectives for gait monitoring beyond the lab environment.
Main part of this research is conducted at Holst Centre in Eindhoven (The Netherlands), an open-innovation initiative by imec and TNO.