Over the past few decades, medical electronics has played a key role in supporting personal disease management and simple and advanced diagnostics. Examples range from blood-glucose and blood-pressure monitoring devices to fever management with an electronic thermometer. Several innovations that focus on increasing quality of life for users are being made in this space. The considerable progress made in this area has prompted developers to look beyond personal healthcare for medical electronics applications.
 

A number of applications are emerging that use both conventional and new medical electronics in conjunction with advanced software intelligence known as biofeedback. Devices that incorporate this technology enable users to maintain health or to train for peak performance. Biofeedback already encompasses a diverse range of applications. Simple to complex biofeedback systems and modern semiconductor devices such as ultra-low-power microcontrollers (MCUs), high-end embedded processors, and high-performance analog front ends (AFEs) can contribute to unlimited innovations in the field of biofeedback.
 

Opportunity for Impact
 

In 10 years, there will be approximately 32% more people in the United States who will be older than 65 years than at present, according to statistics from the World Health Report. By 2025, 1.2 billion people in the world will be over 50 years old—twice as many as in 2006. The aging population will drive up healthcare costs. The National Health Expenditure Report 2009 found that current U.S. healthcare spending is more than 17% of gross domestic product (GDP), and Europe is not far behind. Healthcare costs are expected to double in the next 10 years. Similar trends are reflected in the emerging markets. China’s healthcare expenditure increased from 3.7% of GDP in 1995 to 5.6% in 2007. In 2008, India’s government proposed to increase public expenditure on healthcare from 1 to 3% of GDP, as noted by an article in the Economic Times.
 

Figure 1. Looking at 2009 worldwide consumer medical shipments, we can see the percentage revenue split in various consumer medical shipments.

Undoubtedly, the market for medical electronics is growing fast in all major areas, including consumer medical, diagnostics, imaging, surgical, and monitoring equipment. Consumer medical shipments comprised 33% of semiconductor revenue in 2009 (see Figure 1).
 

The section on fitness equipment includes the emerging personal peak-performance gadgets that recently have become popular. Peak performance is defined as augmentation of physical and mental performance of an individual at the optimal level of their physical capabilities. One can reach this state through a regimen of training exercises that enhance qualities such as cognitive skills, emotional control, and concentration and focus.
 

With athletes, this enhanced performance may mean the difference between winning and losing. With CEOs, CTOs, and other executives responsible for making critical decisions, such technology may simply translate to improved stress management.

Essentially, peak-performance fitness equipment are tools that help an individual practice mind-over-matter skills and achieve maximum benefits during routine physical exercise by giving direct, measurable feedback and enabling an optimized daily workout regimen.
 

Personal biofeedback device categories include neurofeedback with electroencephalogram (EEG) and hemoencephalogram (HEG), heart rate variability (HRV), stress and relaxation, electromyogram (EMG) muscle-activity feedback, skin temperature and core temperature measurement, and pulse oximetry. Notice that these are reuse and new-use versions of time-tested diagnostic technologies known to the healthcare industry. An increasing number of emerging fitness products are now geared toward enhancing performance as opposed to general-purpose fitness applications.
 

Neurofeedback System
 

A neurofeedback system is based on a conventional EEG system that measures and logs the brain’s electrical activity. With the help of modern semiconductors instead of an expensive, conventional hospital setup, these systems are designed to be compact, affordable battery-operated systems that can be used at home.
 

Figure 2. Block diagram of a neurofeedback system.

Figure 2 illustrates the block diagram of a neurofeedback system. An AFE is combined with an MSP430G2452 MCU to digitize and manage signal conditioning and flow. This is the primary front-end block of a neurofeedback system. The back end of the neurofeedback system is comprised of elements that provide feedback to users. A simple user feedback system would use audible tones, a color LED, and so forth to communicate with the user. Training sessions would use audio or visual cues to calm or activate the brain. A high-end version of this setup would comprise  a laptop or PC running sophisticated software and some fast Fourier transform (FFT) filters to analyze brain waves and provide user-friendly feedback to help with training.
 

HRV Measurement System
 

An HRV measurement system is similar to a typical fitness equipment heart-rate measurement system. There are two technologies commonly used to measure heart rate: One is based on electrocardiogram (EKG) and the other on optical pulse pick-up, as in a pulse oximeter system.
 

EKG is the most commonly used technology because it can provide the wearer with reliable performance under any circumstance, whether the user is, for instance, jogging or resting. This system needs electrodes connected to the chest or arms. It is easy to develop and works all the time, primarily because EKG signals are typically 1 mV amplitude. With modern low-cost electronics, this process is simple to handle. A good example of an EKG-based heart rate measuring device in use today is the chest strap-based sports watch.
 

Heart rate variability is simply a heart rate logged and a trend computed over time. In an individual’s perfectly relaxed state, there is minimal-to-no HRV. A low heart rate combined with no HRV is the feedback for best relaxation, experts say.
 

Figure 3. Block diagram of a heart-rate variability measurement system.

Figure 3 illustrates the block diagram of an HRV system. In this setup, an AFE amplifies EKG signals by 1000. The MCU is digitized and further processed. HRV is calculated first by computing heart rate and trending it over time with the logged heart rate information in the MCU’s memory. The HRV is displayed in a LED bar graph in its simplest self-contained implementation. For a more sophisticated setup, the HRV is sent to a PC or smartphone device via USB or wireless connectivity.
 

An example of an EKG-based heart-rate measurement system reference design using an MSP430G2452 is available here.
 

The circuit can easily be extended to perform HRV measurement. Another method to calculate HRV is to measure heart rate using technology commonly used in conjunction with a pulse oximeter. Figure 4 illustrates an optical pulse pick-up system based on pulse oximeter technology. A complete application note with schematics and source code featuring this design, titled “A Single-Chip Pulsoximeter Design Using the MSP430,” is available for download here.
 

Figure 4. Block diagram of a single-chip pulse oximeter design.

Respiratory Waveform Measurement
 

Several methods can be adapted for respiratory waveform measurement. A basic version is to use a wearable chest strap with a sensor, such as a piezo element. As the chest expands and contracts during an inhalation and exhalation cycle, the piezo element generates significant signal voltage that easily can be processed by a simple MCU to compute the respiratory rate.
 

A second method is to measure the exhalation air temperature using a small thermistor close to the nasal airway. But this involves some inconvenience to the individual and isn’t a popular method. Nevertheless, it’s an approach that is easy to develop and a system that can help algorithms compute calories burnt by the individual.
 

A third method uses the principle of impedance pneumography. Such a setup could be realized by an MCU. The low current, which is on the order of a few microamps, of a 50-kHz ac excitation is passed via two chest electrodes placed 6–8 in. apart. This method can be generated using one of the pulse width modulation (PWM) channels on the MCU. The chest forms part of the impedance path for this flow of current. As you might guess, the chest’s impedance varies depending on the amount of air in the lungs. This is simply the variation created by inhalation and exhalation, so the respiratory waveform and respiratory rate can be easily computed.
 

Figure 5. Derivation of a respiratory waveform from a chest EKG waveform.

Another approach is the derivation of respiratory waveform from the chest EKG, which doesn’t require any added electrodes in addition to a conventional EKG electrode setup to measure HRV. If there’s an HRV system in place, this method can be realized almost for free. As shown in Figure 5, the upper trace is the amplified and low-pass filtered EKG signal. You’ll notice the change in the amplitude in the QRS peaks. This variation is actually caused by the thorax region conductivity change because of inhalation and exhalation. When the individual inhales, the lungs are filled with air. This results in an increased impedance of the thorax region, resulting in lower amplitude EKG waveform and vice-versa. A simple amplitude modulation (AM) demodulation scheme can reveal the respiratory waveform as shown in the lower trace of Figure 5.
 

MCU and MEMS Sensors
 

There are peak performance applications for practically every fitness regimen, including activities such as weight lifting, push-ups, and gym workouts (see Figure 6). Several MEMS accelerometers can be placed on strategic locations of the body, such as biceps, triceps, abs, hamstrings, and so forth, to get a measurement and feedback on muscle contraction and expansion caused by the workout. In turn, the feedback helps the individual to adjust the workout strategy to maximize the muscle differential movements. One may even realize that a workout routine used for years was not optimal after all. MEMS sensors are a good alternate for EMG systems.
 

Figure 6. Example setup of a MEMS body area network.

EMG systems provide the same feedback, albeit with a more expensive and uncomfortable setup. Additionally, several MEMS sensors can be designed as a wireless patch and can be networked with schemes such as a body area network (BAN).
 

Today’s MEMS sensors and RF wireless technologies are opening up new possibilities in the personal healthcare and peak performance applications. Inertial sensors, such as accelerometers and gyros, are currently designed into applications such as stability monitoring and exercise feedback. Orientation sensors, such as e-compass and global positioning systems (GPS), are used for in-range monitoring, alarm, and guidance to monitor and track Alzheimer patients, for example.
 

Traditional medical diagnostic technologies have found new use in the field of biofeedback. Modern biofeedback equipment is portable and affordable. In addition, MEMS inertial sensors, such as accelerometers and gyros, have enabled sophisticated biofeedback technologies. Wireless technologies, especially in body area network applications, have also spurred several new products in this fast-emerging biofeedback industry.

Murugavel Raju is the worldwide manager of the systems solutions development for the MSP430 MCU product group at Texas Instruments in Dallas, Texas.