Protecting Bioelectric Signals from Electromagnetic Interference in a Wireless World

Abstract

The exponential growth of wired and wireless technologies which generate Electromagnetic Interference (EMI) has made obtaining microvolt-level bioelectric signals challenging. While digital filtering algorithms provide a wealth of information and allow Artificial Intelligence (AI) to interpret the data, the process may denigrate the integrity of the original signal. Busy clinicians and researchers have relied on computer-analyzed ECG, losing their ability to discriminate between data of high quality and data contaminated with EMI (noise). Resolving an EMI issue with a microphone is one way to learn the methodology. A step-by-step process of troubleshooting EMI in an audio application provides a framework for understanding the fundamental variables that generate EMI and a better understanding of analog electronics. The troubleshooting methodology applies to resolving EMI issues with all biologic signals including Surface Electromyography (SEMG), EEG, ECG, and Needle EMG. As we enter the age of extended range WIFI and cellular technologies, understanding analog electronics is crucial in ensuring we obtain clean data for more clinically meaningful results.

1. Introduction

The development and widespread use of electrically powered technology has grown exponentially since the 1960s. Modern homes and offices routinely have flat-screen TVs, microwave ovens, WIFI, and air conditioning systems. Technology requires power in either Alternating Current (AC) or Direct Current (DC) form. Cellular, WIFI, satellite, and other Radio Frequency (RF) devices provide communication between technologies. The widespread use of wired and wireless technologies has led to increased Electromagnetic Interference (EMI) due to Electromagnetic Fields (EMF) in its many forms. Electromagnetic Fields (EMF) are generated by both the source of power, and the technology device that is powered. Electromagnetic Interference (EMI), or “electrical noise” is the unwanted disturbance in a circuit caused by EMF [1].

EMI creates problems for the electrophysiologist attempting to extract bioelectric signals from the body, especially when those signals are in the microvolt (μV) range such as Surface Electromyography (SEMG), ECG and EEG [2]. Wireless devices have been quickly replacing wired devices in every area of technology. Smart meters are replacing water, gas and power meters in both residential and commercial applications. WIFI routers and microwave based WIFI blanket many towns and cities. Cellular towers, cellular phones and Bluetooth devices all contribute to EMI. In addition to wireless signals, the impact of the electric vehicle has led to an increase in the transmission of high levels of AC power with powerlines the area we live and work. Wireless devices along with EMF generated by power sources contribute to EMI or EMF, potentially damaging the integrity of bioelectric signals as measured at the human body.

In the past decade, there has been a disturbing trend toward clinicians relying upon computer-analyzed ECG/ECG bioelectric data and making clinical decisions based on erroneous results [3]. Unfortunately, many clinicians no longer understand ECG graphs or understand the subtle meaning of the graphed analog signal. This lack of knowledge means there is a higher probability that they will not be aware of the impact of EMI on electrophysiologic evaluations, and the importance of obtaining a clean signal before being processed by the computer.

There are many methods of filtering out EMI, from hardware-based filtering to software algorithms. The best method of preserving the integrity of a bioelectric signal is to reduce the impact of EMI at the source of the biological signal.

Surface Electromyography (SEMG) is a measurement of muscle activity. A pubmed.gov search of the term yields over 13,500 peer-reviewed studies, including the evaluation for presence or absence of back pain and soft tissue injury [4] and workplace ergonomics evaluations. Since this technology measures muscle activity in the microvolt range, it is extremely susceptible to EMI and is a good tool to use when evaluating a location for EMF/EMI. The device utilizes high gain differential amplifiers, and sensors comprised of a metallic electrode with conductive medium which is placed on the skin above the muscles of greatest interest. When performed with the proper equipment and controlling for EMI, the device can extract a microvolt level signal produced by motor units beneath the skin [5].

EMI can make SEMG extremely difficult to measure, making it a perfect tool for evaluating methods of eliminating EMI. The microphone is another device that allows you to“hear” the presence of EMI (constant “hum” in the speakers), and quickly determine when EMI has been removed. Troubleshooting EMI with a commercially available microphone requires the same troubleshooting steps that apply to eliminating EMI’s impact on bioelectric signals.

2. Problem-solving techniques: methods for removing EMI in practical use-case scenarios

With the significant advancements in sensor technology over the past 10 years, EMI is less of an issue as in the past. That being said, there are additional variables beyond the measurement instrument itself which determines the quality of bioelectric signal obtained. One of these variables is the increasing level of EMI in our environment.

2.1 Obtaining quality signals: Practical examples

Described below are practical examples demonstrating methods of reducing or eliminating EMI in actual practice.

2.1.1 Practical example 1: The CT angiogram and ECG

The author was required to have a CT Angiogram, and arrived for the exam with my own pack of ECG electrodes. At the time, the CT was timed based upon the ECG, and without a clean ECG signal, the test could not be performed.

After looking at the electrodes they were using, I asked the nurse if they had difficulty obtaining a stable ECG signal. She acknowledged that they were having significant challenges with the ECG, and did not know why. They had the ECG serviced, but the difficulties continued. What was wrong?

The hospital-supplied electrodes had a very high viscosity “gummy bear” type conductive medium. The author-supplied electrodes had a low viscosity gel form of conductive medium. Having the most high-tech ECG would not matter, as the weakness in this system was in the $5.00 pack of electrodes.

To obtain a clean signal, the skin-electrode interface has to have a low enough impedance that the ECG amplifiers would be capable of acquiring the ECG signal without being “saturated” with EMI from all the other electronics in the room. I explained that the hospital supplied electrodes may work if given enough time for the conductive medium to heat up enough to permeate the skin. The gel-type medium on the author supplied electrodes instantly permeated the skin due to its low viscosity.

2.1.2 Tools for evaluating and reducing EMI

These first two devices are crucial in detecting EMI/EMF issues at the physical location of data collection. The inverse square law applies to EMI. In other words, by moving a few feet away from the source of the EMI may resolve the problem. Therefore, you must test for EMI at the location where you will be performing the data collection. You will see the manufacturer and model number of the devices used in troubleshooting.

1. A simple electrical outlet ground tester confirms the AC outlets are properly grounded.

2. An EMF meter measures RF, EMI, Powerline EMF, and cellular EMI.

3. A Faraday Cage or EMI shielding. There are commercially available shielding paints and cloth.

4. Smart-meter shield, if a smart meter is close to your data collection site.

2.1.3 Resolving EMI issues with a commercially available USB-powered microphone

A demonstrative method of eliminating EMI was in resolving an EMI issue with one of the most popular, commercially available microphones (Blue Microphones model: Yeti). The popularity of this microphone is based upon its simplicity and excellent sound quality (once EMI is removed). It is USB-powered and does not rely upon a separate amplification system.

The approach to resolving EMI issues with a microphone is identical to resolving EMI issues with SEMG or any low-level bioelectric signal. The microphone is a perfect example, as the process for eliminating EMI is the same for the microphone for bioelectric signals. There is no need to learn to operate an oscilloscope as the human ear can hear the change in sound quality when the EMI issue is resolved.

Upon plugging the device into my computer, and attempting to record, I immediately noticed a problem with a low-frequency” hum”. In an online search, I found every expert on the microphone recommending the use of audio editing software with a noise reduction algorithm.

The resultant processed audio demonstrates the exact problem with digital signal filtering: The digital noise reduction definitively removed the hum, but concomitant with this approach was the removal of the subtle, rich, and warm qualities of the human voice heard in the original recording. The use of sound as an analogy demonstrates the impact digital signal processing may have on biologic signals, as sound has much of the same spectral and amplitude characteristics.

In a google search, the experts on this microphone almost unanimously recommended using audio editing software filters and equalizers to recover the unique vocal attributes removed during the noise reduction processing. In following this approach, the resultant output although improved lacked all the qualities of the original recording, degrading the sound quality along with removing the hum. Is there a better way than to use software to filter out the EMI?

Ideally, if we could remove the source of EMI, there would be no need for post-recording filtering. The problem-solving process presented below applies to any bioelectric signal contaminated by EMI (e.g. SEMG, EMG, ECG etc.) As with all problem-solving, it is essential to change one variable at a time (Figures 1–5).

Figure 1.

Grounded 3 prong AC adapter.

Figure 2.

Ungrounded AC adapter.

Figure 3.

The Inverse Square law.

Figure 4.

Faraday cage concept.

Figure 5.

Ferrite Core mini USB cable.

These are the steps I followed in resolving the EMI issue:

1. Use an electrical ground tester (Klein Tools Model RT 310) to confirm that all electrical outlets are tied to earth ground. Although this issue is more likely to occur in an older residential home, it is still possible the location was not properly wired. An ungrounded AC circuit is a source of EMI and needs to be resolved immediately. The ground appeared good in all outlets throughout my office setting.

2. The laptop had an ungrounded AC Adapter (a two-prong plug). An ungrounded AC adapter can be a significant source of EMI, as there is no path to ground. As many computers are utilizing the USB-C connector for power, finding a replacement grounded AC adapter was easy. Another approach while waiting for the AC adapter to arrive was to run the computer on battery power. It is crucial to have t he AC adapter unplugged from the AC outlet and the computer. The only problem is when operating on battery power, the internal power supply may generate enough EMI to cause interference. It is always best to have the system grounded. Unfortunately, upgrading to a grounded AC adapter did not resolve the problem. NOTE: Many believe that there is an adapter that allows conversion from ungrounded AC plug to a grounded AC plug. This is false. There is an adapter which allows conversion from grounded to ungrounded (3 prong to 2 prong plug) but it is impossible to add a ground connection.

3. The Faraday cage is essentially a metal box surrounding the subject with the metal connected to the earth ground connection on the wall outlet. The Faraday cage essentially shields the subject from EMI, drawing it to earth ground. In a location with severe EMI issues, the lab room was painted with a commercially available shielding paint designed with a grounding plate (e.g. Gigahertz Solutions Manufacturer number 863–091 with grounding plate 863–138). It is recommended that a professional perform the installation of such paint. Also, there are manufacturers who provide cloth that can be used to build a Faraday Cage (JJ Care part FF44x20), but the process of building a proper Faraday cage is more complex than it appears [6], and without proper understanding of the principles it is the position of the author that these should be limited to use in clinic or lab settings. The Faraday cage I built for the microphone was unsuccessful.

4. Testing the microphone/computer system at a different location was done to eliminate any source of EMI not obvious. Changing locations has exposed issues for other offices, but did not eliminate the EMI issue with the microphone. In one scenario the dry cleaner next door was using industrial size washers and dryers generating so much EMI that the only solution was to move the entire SEMG system to the opposite end of the office. The good news is that the inverse square law applies to EMI (the impact of EMI drops off rapidly the further you move away from the source of interference).

5. Lights may be a significant source of EMF. Fluorescent lights have always been a source of EMI. Newer LED lighting systems generate significant EMI as well. It is most likely the power supplies which generate the EMI but turning off LEDs can reduce EMI significantly. In my testing with the hum, I also connected a SEMG system to evaluate the EMI of LED lights [7]. Turning off the lights had no impact on the low frequency hum, but did cause significant EMI in the SEMG signal.

6. The next step involved unplugging all electronics in the office, whether near the microphone or not. The most likely culprit? The flat screen TV: If a Plasma TV, the power supplies generate quite a bit of EMI. You will note that plasma TV manufacturers typically equip all important cables including both power cords and those for connection to the source of entertainment (e.g. HDMI cables for DVD players) with ferrite wrapped cores to reduce the impact of EMI [8].

7. Many TV manufacturers have been shipping large screen TVs with ungrounded 2 prong AC power plugs. This makes them the most likely source of EMI when performing any low-level signal measurement. Since monitors and TVs have always been a source of EMI, but I simply made sure to move the microphone at least 8 feet from the large screen TV I was using as a monitor. Moving a few feet away from the screen drops the EMI dramatically as the inverse square law applies.

8. Cellular phones generate significant EMI. It is important when gathering any low-level biologic signal to keep the phone at least 6–8 feet away from the sensors. Many believe if the ringer is off, that the phone will not generate EMI. When a call comes in, whether the ringer is on or off, the wattage increases significantly and can be seen directly on the graph of an SEMG measurement. I also moved my cellular phone at least 6 feet from the microphone but this had no affect.

9. Any device with a motor may be a source of EMI whether it is turned on or not. Massage chairs or motorized tables as found in a chiropractic or physical therapy office need to be unplugged as part of the process of eliminating them as a source of EMI. With the lights off, the laptop powered by a grounded AC adapter and after unplugging all electronics including the massage chair and TV, the microphone issue was not resolved.

10. The next variable involved the Smart Meter proliferating throughout the US. The device sends data to the utility company but appears to generate significant EMI. I measured the level of EMI using an EMF meter (GQ Electronics, Inc. Model EMF-390), and found the smart meter was generating such significant EMI that I felt the need to apply a shield. Initially I was skeptical of these forms of Faraday cages, as the claims being made by some seemed ludicrous. I was wrong. The measured level of EMI dropped by 14 times with the shield installed and grounded (Smart Meter Guard, Model SMG1). The smart meter still functioned properly, but the microphone EMI issue was not resolved.

11. I decided to take an inventory of the system disconnected to evaluate possible sources of EMI that were inherent in the design. What I immediately noticed was that the USB A cable to Mini USB cable (which powered and transmitted data from the microphone to the PC) was a standard cable without a ferrite core. I purchased a male USB-2.0 A to male Mini USB cable (Monoprice model 105,447), manufactured with a ferrite core wrapped around the Mini-USB end of the cable. The problem was resolved, and this wonderful microphone was now usable without any digital filtering. Attenuating EMI at the source will always lead to a cleaner, more clinically valuable signal and reduce the need for as much signal processing. The process above applies to any EMI issue with any device.

2.1.4 Miscellaneous means of reducing EMI

Improving the interface between the human and measurement sensor is always critical. Without proper conductivity, bioelectric signals are difficult to measure.

3. Conclusions

The current trend in biomedical engineering is to focus on development of algorithms that provide the healthcare provider, researcher and layperson with valuable biological data to aid in maintaining better health and automating the evaluation of bioelectric signals. The focus of this chapter provides a reminder to be aware of the environmental effects on any biologic signal and not to completely rely upon computers for automated analysis. Although the computer will definitively do a much better job in the future, there may be a variable not taken into consideration.

Additionally, it is important to always attempt to properly attach sensors to the body. By doing so, we increase the probability of obtaining a clean signal at the sensor/human interface such that the original signal is faithfully reproduced in digital format. With the proliferation of wireless technologies throughout the globe, the importance of preventing Electromagnetic Interference (EMI) from contaminating the biologic signal is of utmost importance.

The process of eliminating EMI is presented in the simple case study of a microphone, but the algorithm applied to resolving EMI issues with a microphone apply to all human bioelectric signals. It is the hope that this chapter provide some a different perspective that may aid the researcher and clinician in developing a methodology which leads to the cleanest, noise-free signal possible when gathering bioelectric signals.