Are you in Canada? Click here to proceed to the HK Canada website.

For all other locations, click here to continue to the HK US website.

Human Kinetics Logo

Purchase Courses or Access Digital Products

If you are looking to purchase online videos, online courses or to access previously purchased digital products please press continue.

Mare Nostrum Logo

Purchase Print Products or eBooks

Human Kinetics print books and eBooks are now distributed by Mare Nostrum, throughout the UK, Europe, Africa and Middle East, delivered to you from their warehouse. Please visit our new UK website to purchase Human Kinetics printed or eBooks.

Feedback Icon Feedback Get $15 Off

Human Kinetics is moving to summer hours. Starting May 31 – August 2, our hours will be Mon – Thurs, 7am – 5pm CDT. Orders placed on Friday with digital products/online courses will be processed immediately. Orders with physical products will be processed on the next business day.

Wearable devices to measure motion

This is an excerpt from Neuromechanics of Human Movement-6th Edition by Roger M. Enoka.

One significant advance in technology in recent years has been the miniaturization of devices that can ­measure motion. The devices have been embedded in accessories that can be attached to the body, such as watches and smartphones. The key components in ­these devices are accelerometers, gyroscopes, and magnetometers. As the name implies, an accelerometer is a device that can ­measure acceleration in one direction. The inclusion of three accelerometers in a single device enables the assessment of acceleration in all three dimensions. Accelerometers comprise a mass attached to a spring and acceleration is sensed by the changes in the length of the spring (sensor) caused by acceleration of the mass. The two most common sensors detect piezoelectric or capacitive signals. Piezoelectric accelerometers use ­either a single crystal or ceramic material to detect a change in electrostatic force that is proportional to the applied acceleration. In contrast, body-­worn devices typically include micro-­electro-­mechanical-­system (MEMS) accelerometers that are constructed out of silicon and use capacitive sensors.

An analogous system exists in our inner ear. Displacements of otoliths relative to hair cells in the inner ear provide information about acceleration. The two otolith organs in each ear detect acceleration in the vertical (utricle) and horizontal (saccule) directions to provide information about the motion and orientation of the head to its surroundings. Accelerometers are commonly used to determine orientation and to detect position, velocity, and vibration.

A gyroscope is a device that is used to ­measure rotation or to maintain orientation. Early gyroscopes comprised a spinning rotor (circular disk) that was suspended inside three rings (gimbals). When the axis about which the rotor is spinning is itself rotated, a force that is proportional to the angular velocity of the rotor ­will be applied to the force-­sensing structures that are supporting the rotor. Gyroscopes integrated in wearable devices are based on MEMS technology and use a vibrating ele­ment instead of a rotating ele­ment. Nonetheless, the physical princi­ples are the same.

A gyroscope ­measures angular velocity about an axis of rotation based on the conservation of angular momentum (amount of rotational motion) princi­ple. The inclusion of three gyroscopes in a single device enables the assessment of angular velocity in three dimensions. In our inner ear, movement of fluid in the three semicircular canals provides information about angular velocity in three dimensions. Gyroscopes are used in many applications, including motion sensing in smartphones, gaming systems, digital cameras, drones, and navigation systems.

A magnetometer is a device that senses the direction of a magnetic field. It is often ­measured with a Hall-­effect sensor, which can detect the orientation of a magnetic field in one direction. Once the downward direction is determined with an accelerometer, the magnetic-­field component in the horizontal plane can be identified from the signals generated by three Hall-­effect sensors (three-­axis magnetometer).

Many smartphones contain miniaturized MEMS magnetometers that comprise a magnetoresistive permalloy sensor, which enables the device to function as a compass. One limitation of the information derived from magnetometers is that it is easily compromised by stray fields from electronic devices, motors, power cables, metal building structures, and so on. Nonetheless, when combined with accelerometers and gyroscopes, the three devices together provide a thorough kinematic assessment of movement.

Activity Trackers

The basic component of most wrist-­ and arm-­worn activity trackers (e.g., Apple Watch, Basis Band, Fitbit, Garmin, Jawbone, Samsung Galaxy Watch, and Xiaomi Mi Band) is a triaxial accelerometer. The acceleration signals are analyzed by an algorithm that can count the number of steps during many, but not all, activities of daily living. When combined with an individual’s age, height, and body mass, the step count data can be used to estimate the distance traveled while wearing the device, average speed, energy expenditure, and the type and duration of exercise. Also, the algorithms can use the acceleration signals obtained while sleeping to estimate sleep quality.

­These devices also can include a number of other sensors, such as an optical sensor (photoplethysmography) to ­measure heart rate, an altimeter to ­measure changes in elevation, an electrodermal sensor to determine the conductivity of the skin (galvanic skin response) as an index of stress, a temperature sensor to ­measure skin temperature, electrical sensors to provide a basic electrocardiogram (ECG) reading, an oxygen saturation sensor (reflectance oximetry) to ­measure blood oxygen levels, and a bioimpedance sensor to determine respiratory rate.

Despite the popularity of ­these devices, ­there is some concern among researchers over the accuracy of the ­measurements. To address ­these concerns, some studies have compared the values obtained with body-­worn activity trackers to ­those ­measured with research-­grade devices. The comparisons have found that some ­measurements made with activity trackers are quite accurate, but some are not. For example, when compared with video-­based visual observations, a waist-­worn activity tracker and a research-­grade (ActiGraph) accelerometer achieved similar accuracy in step count when healthy adults walked on a treadmill at dif­fer­ent speeds (Chow et al., 2017).

However, the accuracy of a step-­count device is influenced by where it is placed on the body, the manufacturer of the device, and the types of physical activities performed by the user (Fuller et al., 2020). Nonetheless, Nelson and Allen (2019) found that two wrist-­worn devices provided acceptable heart-­rate accuracy across a 24 h period of typical activities of daily living when compared with data obtained from an ambulatory ECG.

Like the ­factors that influence step-­count accuracy, estimates of energy expenditure derived from wrist-­ and arm-­worn devices are variable, and the accuracy differs across devices and depends on the type of physical activity performed by the user (Fuller et al., 2020; O’Driscoll et al., 2020). However, the inclusion of heart rate or heat sensors in devices can often improve estimates of energy expenditure relative to ­those based on accelerometry alone. The major challenge in this industry is to develop devices that can recognize changes in the type of physical activity being performed by the user (Dorn et al., 2019).

Wrist-­worn activity trackers that are designed to ­measure sleep quality are reasonably accurate in providing gross estimates of sleep ­parameters and the time spent in sleep stages, but they cannot detect other features of sleep (Haghayegh et al., 2019). In contrast, comprehensive assessments of sleep typically include polysomnography (electroencephalography), eye-­movement signals, cardiac signals (electrocardiography), muscle activity (electromyography), and fin­ger photoplethysmography. In a promising development, a finger-­worn device that incorporates a three-­axis accelerometer, photosensor, and temperature sensors can generate a more complete and accurate description of sleep quality (Altini & Kinnunen, 2021). For example, the device can detect the four stages of sleep (light nonrandom eye movements [NREM], deep NREM, REM, and wake) with 79% accuracy, compared with 57% accuracy for an accelerometer-­based device. ­Future technical developments are expected to improve the accuracy of ­these devices.

Inertial ­Measurement Units

Inertial ­measurement units (IMUs), which typically include accelerometers, gyroscopes, and magnetometers, can be used to ­measure the kinematic characteristics of movement. Most often, IMUs are used during ambulatory activities when multiple devices are attached to study participants. When compared with data obtained using other gait-­analysis approaches, IMU systems achieve good agreement for many gait ­parameters (Alcantara et al., 2021; Rudisch et al., 2021).

In a systematic review of seven studies that compared the two approaches in the assessment of gait, for example, Petraglia and colleagues (2019) found that IMU ­measurements matched the gold-­standard data for speed, step length, step time, stride time, swing time, and cadence, but not stance time. Moreover, an IMU system that comprised six body-­worn sensors (three-­axis accelerometers and gyroscopes) was able to detect balance and gait deficits in ­people with multiple sclerosis who had normal walking speeds and no differences in stopwatch-­timed ­measures on the clinical tests (Spain et al., 2012).

A major advantage of IMU systems over laboratory-­based approaches is that they enable the ­measurement of movement kinematics in a wide range of conditions (Renggli et al., 2020). Also, IMU systems can ­either be customized by the user or be standardized by a manufacturer for a specific application. Customization of an IMU system is useful when studying unique protocols or when assessing the merits of a new commercial product. In contrast, the advantage of standardized systems is that they minimize the burden on the investigators when evaluating ­performance on well-­defined tasks, such as clinical tests of motor function (Horak et al., 2015). Such systems include specific instructions on the placement of multiple IMU devices, hardware that can capture and upload the wireless signals, and software that analyzes the data to produce a prescribed range of outcome variables.

One example of a standardized IMU system is a product known as Mobility Lab (Washabaugh et al., 2017). It comprises six IMU sensors that contain accelerometers, gyroscopes, and magnetometers; a transmitter; and a laptop to rec­ord the signals. The system can ­measure limb kinematics during tests of walking ­performance, postural sway, balance while turning, dynamic balance during standing and sitting tasks, and trunk and lumbar range of motion. Each of ­these assessments includes from 2 to 14 outcome variables, which can be combined into defined sets to provide a single value that can be compared with a normative database. For example, the Mini-­Balance Evaluation Systems Test (Mini-­BESTest) combines 14 items to provide an assessment of dynamic balance (Franchignoni et al., 2010). The Mini-­BESTest is a useful clinical tool to ­measure the balance capabilities of vari­ous groups of individuals, including persons with multiple sclerosis (Potter et al., 2019) and ­those with Parkinson’s disease (Lopes et al., 2020).

One limitation of IMU devices is that most are not able to ­measure limb rotations, such as during reaching actions, or quantify fine motor activities, such as wrist and fin­ger movements. In contrast to most motion-­analysis systems that ­measure body position directly, IMUs use the recorded signals to estimate position. ­Because the ­measurements are made relative to the device itself instead of an absolute reference, small errors can accumulate over time and compromise the accuracy of the output signals. However, careful calibration of each IMU at the beginning of a test can minimize this error. Also, emerging algorithms have the potential to reduce some of the limitations of IMUs and expand their utility for a broader range of applications (Schwerz de Lucena et al., 2021; Slade et al., 2022).

Motion Analy­sis With Smartphones

­Because many smartphones include accelerometers, gyroscopes, and magnetometers, they can be used to ­measure the kinematic characteristics of some movements. The advantages of smartphone technology is the widespread availability of the devices, their low cost compared with other motion-­analysis systems, and the ease with which they can be used outside clinical settings.

­There are several challenges, however, including data acquisition and reliance on a single device. The first issue has been resolved by the development of applications that can be installed on a phone and used to acquire data that can be transmitted to a computer. The second issue is being addressed by the development of algorithms that can use the acquired signals to evaluate a user’s ­performance on a prescribed task and generate an outcome ­measure.

This technology is continuing to evolve. It appears to be reliable and to provide valid ­measures for some, but not all, movements. For example, Mansson and colleagues (2021) found that ­measurements of leg strength performed by older adults with a smartphone ­were in good agreement with the results obtained in a clinical assessment, but the results for two balance tests during upright standing produced mixed correlations with the clinical tests. Similarly, Hseih and Sosnoff (2021) found strong correlations between smartphone and accelerometer (placed on the smartphone) recordings when ­people with multiple sclerosis stood upright but found weak-­to-­moderate correlations between the smartphone and force plate recordings. ­These results underscore the potential for smartphones in ­these settings but suggest that additional technical improvements are needed.

Despite the limitations of smartphones as research tools, they can be used to provide motivational support for participants in physical activity programs, and they offer an accessible solution for analyzing movement in an educational setting. For example, Phyphox (Physical Phone Experiments) is a freely available application that was developed by physicists at the RWTH Aachen University in Germany. The application is available for ­either Android or Apple smartphones. Phyphox has extensive web-­based support for 35 dif­fer­ent experiments that can be performed using smartphone sensors (accelerometer, gyroscope, magnetometer, light sensor, pressure sensor, video, and microphone). The application and its associated website are an excellent resource for student proj­ects.

More Excerpts From Neuromechanics of Human Movement 6th Edition