Medical Devices

Nonintrusive, Wearable Bioelectrodes for Monitoring the Heart and Brain

September 1, 2007 By: Robert Matthews et al., Quantum Applied Science & Research (QUASAR) Sensors

Improved noncontact electrodes and noninvasive hybrid biosensors enable fast, reliable readings of hearts and brains, without requiring contact with the skin.


The principal technical difficulty in measuring bioelectric signals from the body lies in establishing good, stable electrical contact to the skin. Despite over 40 years of investigation, including the development of several alternative electrode technologies, no reliable method has been developed for making electrical contact to the skin that does not require some modification of its outer layer. This involves the use of conductive pastes or gels, often with abrasive skin preparation of the electrode site, which can be time consuming and unpleasant for the subject.

In response to the Department of Defense's interest in providing fast and reliable readings of soldiers' bioelectric heart and brain signals during training and on the battlefield, Quantum Applied Science and Research (QUASAR) has developed two innovative noninvasive biosensor technologies that require no modification of the skin's outer layer and can be embedded in the DoD's Future Force Warrior warfighters' uniforms and helmets. These technologies form the basis of a biomonitoring sensor suite currently being developed for the U.S. Army's Aberdeen Test Center, in which electroencephalogram (EEG), electroocculogram (EOG), electrocardiogram (ECG), and electromyogram (EMG) signals are measured for monitoring human-computer interactions. The many civilian applications for these technologies include long-term monitoring of subjects who have been diagnosed with cardiac or neurological conditions, immediate readings for those transported in ambulances, and quick evaluations of an athlete's brain after a violent contact incident.

QUASAR's capacitively coupled noncontact electrode (CCNE) technology was described in an earlier article. In this current article, we present advancements upon the CCNE technology, and describe a hybrid biosensor that uses a combination of high-impedance resistive and capacitive contact to the subject.

Capacitive Biosensors
In 2002, QUASAR developed a new class of bioelectric sensors that coupled to the body capacitively. These devices differ from previous capacitive electrodes in that they can tolerate very small capacitances to the source. Thus, the sensors can be operated at a standoff from the skin (up to several millimeters in practice), which makes it possible to measure the ECG through clothing.

Conventional wet electrode technology relies upon a low contact impedance to the body (typically <5 k Ω in clinical settings) to limit common-mode voltages appearing on the body and to reduce interference caused by picking up external signals. The high coupling impedance of the CCNE has necessitated the development of specialized capacitive sensing techniques. In the CCNE sensor (Figure 1), the amplifier electronics are shielded and placed as close as possible to the electrode to limit pickup of external signals.

 

figure
Figure 1. QUASAR's capacitive biosensor compared to dimes

To provide immunity to common-mode (CM) signals appearing on the body, we operate the sensor in conjunction with the company's proprietary common-mode follower (CMF) technology. Due to its ultra-high input impedance (~1012 Ω), the CMF tracks the body-ground potential with a high degree of accuracy. This can then be used to dynamically reduce any CM signals by more than three orders of magnitude.

Since the publication of the previous article, we have made considerable technical improvements to the CCNE technology. A new circuit design has improved sensor performance and also reduced sensor power consumption by the sensors to less than 400 µW. Further improvements include a reduction in the size of the sensor, improved robustness and reliability of the dielectric coating on the electrode, and a simplified manufacturing process to reduce the cost per sensor.

An example of through-clothing ECG data collected using the CCNE sensors is shown in Figure 2. The signal quality is very similar to ECG data simultaneously collected using wet electrodes (included in the figure for comparison). During evaluation, subjects moved through an obstacle course, experiencing moderate to high levels of physical activity. The sensors achieved a 99.8% correct classification of the heart rate (corresponding to 23 points incorrectly classified or missing out of a total of 10,417 points). A Polar S180i heart rate monitoring system installed immediately below the QUASAR sensors on the midline of the chest achieved a 94.5% correct classification of the heart rate.

 

figure
Figure 2. Comparison of ECG measured using QUASAR's CCNE (green) and wet electrodes (blue)

In Figure 3, we compare EEG signal fidelity during the onset of alpha activity as measured by placing a CCNE and a conventional pre-gelled electrode side by side on a subject's forehead. The skin under the wet electrode was prepared with an alcohol wipe (with no abrasion), whereas no skin preparation was used for the CCNE. These results show excellent agreement between the sensors, but the electrode site was specifically chosen to be on the forehead in order to avoid hair. Measuring EEG through hair using capacitive biosensors is problematic because the sensors' high impedance renders them susceptible to triboelectric charge generation due to rubbing between the sensor and hair.

 

figure
Figure 3. Alpha activity observed using QUASAR capacitive sensors (red) and conventional resistive contact electrodes (blue); the two curves are offset to illustrate the temporal correlation between the two sensors

Hybrid Biosensors

figure
Figure 4. QUASAR's hybrid EEG biosensor: original prototype (on left), and new, more compact design (on right); a quarter is shown for scale

Using the experience gained from the CCNE, the company has developed a novel hybrid biosensor technology (Figure 4). These sensors address the issue of triboelectric charging in capacitive sensors by making contact with the skin via a set of 'fingers,' each of which is small enough to reach through hair without trapping hairs beneath the finger, thus ensuring electrical contact to the scalp. The contact impedance between the scalp and each finger can be as high as 107 Ω, which is small enough to reduce triboelectric charge effects to a level too low to affect the measurement. However, this impedance is 103 times greater than that used with conventional resistive contact electrodes, in which case the high impedance techniques used in the CCNE were still found to be necessary.

To evaluate the signal fidelity of the hybrid biosensors, we took simultaneous EEG measurements using hybrid sensors side by side with conventional wet EEG electrodes. The hybrid sensors were incorporated into a cap worn by the subject (Figure 5), with wet electrodes positioned 2 cm behind the hybrid sensors. To prepare for the wet electrodes, the scalp was abraded, cleaned with alcohol, and a conductive paste applied. No preparation of the scalp was performed at the hybrid electrode sites.

 

figure
Figure 5. Locations of hybrid biosensors and conventional Ag/AgCl electrodes

The hybrid electrodes were referenced to a CMF placed upon the subject's right earlobe. The ground reference for the wet EEG electrodes was a standard pre-gelled disposable Ag/AgCl electrode placed upon a prepared site on the right ear.

Figure 6 compares EEG signal fidelity between hybrid sensors and Ag/AgCl electrodes positioned at the exact center top of the head (Cz). The 4 second segment in the figure shows alpha activity measured by both electrodes. The plot also includes the Pearson correlation over 0.25 seconds. Very significant correlations, in excess of 90%, are evident throughout the record.

 

figure
Figure 6. Alpha activity observed using QUASAR hybrid sensors (black) and conventional resistive contact electrodes (red); the two curves are offset to illustrate the temporal correlation between the two sensors

Comparisons between a hybrid sensor and an Ag/AgCl electrode positioned at the point midway between the top of the head and the forehead (Fz) for one subject performing a difficult counting task are presented in Figures 7 and 8. The data for a seated subject (Figure 7) were measured 10 minutes prior to the data for a walking subject (Figure 8). Nevertheless, the signals are similar for both electrode technologies, with correlations in excess of 80% observed in both plots. The signal amplitudes are also similar between plots, demonstrating the benefit of the CMF. The CMF reduces the CM signals appearing on the body, effectively removing the larger interfering signals, such as muscle movement, caused by triboelectric charge generation.

 

figure
Figure 7. EEG observed using QUASAR hybrid sensors (black) and conventional resistive contact electrodes (red) while subject was sitting still; the two curves are offset to illustrate the temporal correlation between the two sensors

 

figure
Figure 8. EEG observed using QUASAR hybrid sensors (black) and conventional resistive contact electrodes (red) while subject was walking; the two curves are offset to illustrate the temporal correlation between the two sensors

To obtain an estimate of measurement noise for the hybrid biosensors we placed two hybrid bioelectrodes on unprepared sites on the subjects' scalps. The electrodes were positioned side by side (spacing ~20 mm) to minimize the differential EEG signal detected by each electrode. We simultaneously collected data for wet electrodes on prepared sites.

Click for larger image Figure 9. Measured sensor-sensor noise levels for hybrid biosensors under subject motion (Click image for larger version)

Results for the hybrid biosensors are recorded in Figure 9. The quoted noise figures correspond to 90% of voltage fluctuations from the mean voltage. The results for a still subject and a walking subject are more than an order of magnitude greater than sensor noise in this bandwidth (<1 µVp-p), and therefore represent skin noise on unprepared skin. For comparison, we also present simultaneous measurements from wet electrodes on prepared sites; for this subject the noise of the hybrid electrodes is twice the observed noise between wet electrodes on prepared sites, whereas in some subjects the noise for the hybrid electrodes can be <20% higher than the noise between wet electrodes.

Future Systems
One of the principal benefits of QUASAR's biosensor technology is the ease with which the sensors may be applied. For example, the CCNE sensors can be incorporated into a chair or bed for completely unobtrusive monitoring of ECGs (Figure 10). For EEG, a harness can be used to position the hybrid sensors and hold them comfortably against the scalp. The prototype harness in Figure 11 can be attached by a novice without the aid of an EEG technician by using a simple combing motion to work the sensors through the hair. Ultimately, the sensors in the harness can be embedded in helmets for soldiers and athletes and inserted in a set of headphones or glasses (Figure 12).

 

figure
Figure 10. Real-time measurement of ECG using sensors hidden in a military stretcher with nothing attached to the subject; ECG signal is shown on the computer screen in the foreground

 

figure
Figure 11. Harness for positioning and holding sensors in position on scalp, with sensors shown in the Cz, Fz and F8 positions

 

figure
Figure 12. Two system concepts for unobtrusive EEG monitoring using hybrid sensors

To complete the system, a miniature, ultra-low power data acquisition and wireless unit is currently being developed at QUASAR for use with its cardiac and EEG sensors. This package will power the sensors, digitize the signals, and either store the data to Flash memory or wirelessly transmit the data to a base station. The wireless components operate using exceptionally low power levels, on the order of 50 nJ/bit of acquired data. This system will offer the capability of full deployment of the system on the battlefield and in civilian applications by year's end.


Add Comment




IIoT University


Deep Learning for Vision Using CNNs and Caffe: A Hands-on Tutorial – 9/22/16 – Cambridge, Mass


IDE






Sensors 2017 Call for Speakers


Sensors Midwest


Advertise


Subscribe



Twitter Feed

Find It Fix It Forum

Sensors invites you to join the Findit-Fixit Forum, where you can get answers to your sensing questions—concerning technologies, products, methods, applications, and services--and also offer help to your fellow engineers. The Forum covers all kinds of topics, from the basics to the extraordinary.

Join the discussion!


© Copyright 2016 Questex, LLC. All Rights Reserved. Sensorsmag. Privacy Policy | Terms of Use

If you are having technical difficulties or considerations, please contact the webmaster.