A comprehensive comparison of simple step counting techniques using wrist- and ankle-mounted accelerometer and gyroscope signals

Knee Angle Estimation with Dynamic Calibration Using Inertial Measurement Units for Running

The knee flexion angle is an important measurement for studies of the human gait. Running is a common activity with a high risk of knee injury. Studying the running gait in realistic situations is challenging because accurate joint angle measurements typically come from optical motion-capture systems constrained to laboratory settings. This study considers the use of shank and thigh inertial sensors within three different filtering algorithms to estimate the knee flexion angle for running without requiring sensor-to-segment mounting assumptions, body measurements, specific calibration poses, or magnetometers. The objective of this study is to determine the knee flexion angle within running applications using accelerometer and gyroscope information only. Data were collected for a single test participant (21-year-old female) at four different treadmill speeds and used to validate the estimation results for three filter variations with respect to a Vicon optical motion-capture system. The knee flexion angle filtering algorithms resulted in root-mean-square errors of approximately three degrees. The results of this study indicate estimation results that are within acceptable limits of five degrees for clinical gait analysis. Specifically, a complementary filter approach is effective for knee flexion angle estimation in running applications.

Can Stereotype Threat and Lift Visual Messages Affect Subsequent Physical Activity? Evidence from a Controlled Experiment Using Accelerometers

This study investigated how visual messages conveying stereotype threat or lift influenced physical activity performance. Participants (N = 380) were exposed to a stereotype threat, lift, or control condition image and then engaged in a running task. Accelerometers recorded forward-backward movement, upward-downward movement, and sideways balance. Stereotype threat exposure increased state anxiety relative to the control condition. In addition, forward-backward movement was linked to state anxiety and participants' sex. Moreover, women exposed to stereotype threat who experienced increased state anxiety showed reduced forward-backward movement. Men exposed to stereotype lift displayed higher forward-backward movement. Additionally, stereotype threat visual message exposure increased sideways balance activity for women but not for men. Upward-downward movement was unaffected by stereotype threat or lift. We discuss theoretical and practical implications of how exposure to visual stereotypes can influence physical activity performance.

Continuous Monitoring of Health and Mobility Indicators in Patients with Cardiovascular Disease: A Review of Recent Technologies

Cardiovascular diseases kill 18 million people each year. Currently, a patient's health is assessed only during clinical visits, which are often infrequent and provide little information on the person's health during daily life. Advances in mobile health technologies have allowed for the continuous monitoring of indicators of health and mobility during daily life by wearable and other devices. The ability to obtain such longitudinal, clinically relevant measurements could enhance the prevention, detection and treatment of cardiovascular diseases. This review discusses the advantages and disadvantages of various methods for monitoring patients with cardiovascular disease during daily life using wearable devices. We specifically discuss three distinct monitoring domains: physical activity monitoring, indoor home monitoring and physiological parameter monitoring.

Wrist Accelerometer Estimates of Physical Activity Intensity During Walking in Older Adults and People Living With Complex Health Conditions: Retrospective Observational Data Analysis Study

BACKGROUND

Accurate measurement of daily physical activity (PA) is important as PA is linked to health outcomes in older adults and people living with complex health conditions. Wrist-worn accelerometers are widely used to estimate PA intensity, including walking, which composes much of daily PA. However, there is concern that wrist-derived PA data in these cohorts is unreliable due to slow gait speed, mobility aid use, disease-related symptoms that impact arm movement, and transient activities of daily living. Despite the potential for error in wrist-derived PA intensity estimates, their use has become ubiquitous in research and clinical application.

OBJECTIVE

The goals of this work were to (1) determine the accuracy of wrist-based estimates of PA intensity during known walking periods in older adults and people living with cerebrovascular disease (CVD) or neurodegenerative disease (NDD) and (2) explore factors that influence wrist-derived intensity estimates.

METHODS

A total of 35 older adults (n=23 with CVD or NDD) wore an accelerometer on the dominant wrist and ankle for 7 to 10 days of continuous monitoring. Stepping was detected using the ankle accelerometer. Analyses were restricted to gait bouts ≥60 seconds long with a cadence ≥80 steps per minute (LONG walks) to identify periods of purposeful, continuous walking likely to reflect moderate-intensity activity. Wrist accelerometer data were analyzed within LONG walks using 15-second epochs, and published intensity thresholds were applied to classify epochs as sedentary, light, or moderate-to-vigorous physical activity (MVPA). Participants were stratified into quartiles based on the percent of walking epochs classified as sedentary, and the data were examined for differences in behavioral or demographic traits between the top and bottom quartiles. A case series was performed to illustrate factors and behaviors that can affect wrist-derived intensity estimates during walking.

RESULTS

Participants averaged 107.7 (SD 55.8) LONG walks with a median cadence of 107.3 (SD 10.8) steps per minute. Across participants, wrist-derived intensity classification was 22.9% (SD 15.8) sedentary, 27.7% (SD 14.6) light, and 49.3% (SD 25.5) MVPA during LONG walks. All participants measured a statistically lower proportion of wrist-derived activity during LONG walks than expected (all P<.001), and 80% (n=28) of participants had at least 20 minutes of LONG walking time misclassified as sedentary based on wrist-derived intensity estimates. Participants in the highest quartile of wrist-derived sedentary classification during LONG walks were significantly older (t₁₆=4.24, P<.001) and had more variable wrist movement (t₁₆=2.13, P=.049) compared to those in the lowest quartile.

CONCLUSIONS

The current best practice wrist accelerometer method is prone to misclassifying activity intensity during walking in older adults and people living with complex health conditions. A multidevice approach may be warranted to advance methods for accurately assessing PA in these groups.

Cross-Leg Prediction of Running Kinematics across Various Running Conditions and Drawing from a Minimal Data Set Using a Single Wearable Sensor

The feasibility of prediction of same-limb kinematics using a single inertial measurement unit attached to the same limb has been demonstrated using machine learning. This study was performed to see if a single inertial measurement unit attached to the tibia can predict the opposite leg’s kinematics (cross-leg prediction). It also investigated if there is a minimal or smaller data set in a convolutional neural network model to predict lower extremity running kinematics under other running conditions and with what accuracy for the intra- and inter-participant situations. Ten recreational runners completed running exercises under five conditions, including treadmill running at speeds of 2, 2.5, 3, and 3.5 m/s and level-ground running at their preferred speed. A one-predict-all scheme was adopted to determine which running condition could be used to best predict a participant’s overall running kinematics. Running kinematic predictions were performed for intra- and inter-participant scenarios. Among the tested running conditions, treadmill running at 3 m/s was found to be the optimal condition for accurately predicting running kinematics under other conditions, with R2 values ranging from 0.880 to 0.958 and 0.784 to 0.936 for intra- and inter-participant scenarios, respectively. The feasibility of cross-leg prediction was demonstrated but with significantly lower accuracy than the same leg. The treadmill running condition at 3 m/s showed the highest intra-participant cross-leg prediction accuracy. This study proposes a novel, deep-learning method for predicting running kinematics under different conditions on a small training data set.

Validation of open-source step-counting algorithms for wrist-worn tri-axial accelerometers in cardiovascular patients

BACKGROUND

Accurate quantification of daily steps in a cardiovascular patient population is of high importance for primary and secondary prevention. While sensor derived step counts have been sufficiently validated for hip-worn devices and commercial wrist-worn devices, there is a lack of knowledge on validity of freely available step counting algorithms for raw acceleration data collected at the wrist.

RESEARCH QUESTION

How accurate are step-counting algorithms for wrist worn tri-axial accelerometers in a cardiac rehabilitation training setting?

METHODS

Two step counting algorithms (Windowed Peak Detection, Autocorrelation) for tri-axial accelerometers (Axivity AX-3), were tested. Steps were recorded by chest-mounted GoPro video cameras as gold standard. Cardiovascular patients without neurological impairments enrolled in an ambulatory rehabilitation program were recruited. Recordings were performed during one 45-90 min outdoor physical therapy session of which 5-min segments of six movement categories, namely Walking, Running, Nordic, Stairs, Arm Movement [AM] With [+] and Without [-] Walking [W] were identified and analyzed. Mean absolute difference and mean absolute percentage error [MAPE] with regard to true steps measured from video are reported to report accuracy.

RESULTS

Training sessions of 22 patients were recorded and analyzed. Steps were overestimated during AM-W and underestimated during Walking, Running and Stairs. Windowed Peak Detection algorithm was more accurate during AM+W and AM-W and Autocorrelation performed better during Nordic. A MAPE of close or below 10% was achieved by both algorithms for the categories: Walking, Running, Stairs and Nordic.

SIGNIFICANCE

Both algorithms provided accurate results for estimation of step counts in a controlled setting of a cardiovascular patient population. The quantification of daily number of steps recorded by wrist-worn accelerometers delivering raw data analyzed by freely available algorithms is a cost-effective option for research studies.

Comparison between Accelerometer and Gyroscope in Predicting Level-Ground Running Kinematics by Treadmill Running Kinematics Using a Single Wearable Sensor

Wearable sensors facilitate running kinematics analysis of joint kinematics in real running environments. The use of a few sensors or, ideally, a single inertial measurement unit (IMU) is preferable for accurate gait analysis. This study aimed to use a convolutional neural network (CNN) to predict level-ground running kinematics (measured by four IMUs on the lower extremities) by using treadmill running kinematics training data measured using a single IMU on the anteromedial side of the right tibia and to compare the performance of level-ground running kinematics predictions between raw accelerometer and gyroscope data. The CNN model performed regression for intraparticipant and interparticipant scenarios and predicted running kinematics. Ten recreational runners were recruited. Accelerometer and gyroscope data were collected. Intraparticipant and interparticipant R² values of actual and predicted running kinematics ranged from 0.85 to 0.96 and from 0.7 to 0.92, respectively. Normalized root mean squared error values of actual and predicted running kinematics ranged from 3.6% to 10.8% and from 7.4% to 10.8% in intraparticipant and interparticipant tests, respectively. Kinematics predictions in the sagittal plane were found to be better for the knee joint than for the hip joint, and predictions using the gyroscope as the regressor were demonstrated to be significantly better than those using the accelerometer as the regressor.

Generation and validation of ActiGraph GT3X+ accelerometer cut-points for assessing physical activity intensity in older adults. The OUTDOOR ACTIVE validation study

The study of physical activity in older adults is becoming more and more relevant. For evaluation of physical activity recommendations, intensity-specific accelerometer cut-points are utilized. However, research on accelerometer cut-points for older adults is still scarce. The aim of the study was to generate placement-specific cut-points of ActiGraph GT3X+ activity counts and raw measures of acceleration to determine physical activity intensity in older adults. A further aim was to compare the validity of the generated cut-points for a range of different physical activities. The study was a single experimental trial using a convenience sample. Study participants were 20 adults aged 59 to 73 years. Accelerometers were worn at six different placements (one on each wrist, one on each ankle, and two at the hip) and breath-by-breath indirect calorimetry was used as the reference for energy. The experiment comprised of two parts; a) The first required participants to walk on a treadmill at incremental speeds (3.0-5.0 km·h-1), and b) Five different everyday activities (reading, cleaning, shopping, cycling, aerobics) were staged in the laboratory setting. Accelerometer cut-points (activity counts, raw data) were derived for each of the investigated placements by linear regression using the treadmill part. Performance of the cut-points was assessed by applying the cut-points to the everyday activities. We provide cut-points for six placements and two accelerometer metrics in the specific age group. However, the derived cut-points did not outperform published ones. More research and innovative approaches are needed for improving internal and external validity of research results across populations and age groups.

The Effect of Sensor Placement and Number on Physical Activity Recognition and Energy Expenditure Estimation in Older Adults: Validation Study

BACKGROUND

Research has shown the feasibility of human activity recognition using wearable accelerometer devices. Different studies have used varying numbers and placements for data collection using sensors.

OBJECTIVE

This study aims to compare accuracy performance between multiple and variable placements of accelerometer devices in categorizing the type of physical activity and corresponding energy expenditure in older adults.

METHODS

In total, 93 participants (mean age 72.2 years, SD 7.1) completed a total of 32 activities of daily life in a laboratory setting. Activities were classified as sedentary versus nonsedentary, locomotion versus nonlocomotion, and lifestyle versus nonlifestyle activities (eg, leisure walk vs computer work). A portable metabolic unit was worn during each activity to measure metabolic equivalents (METs). Accelerometers were placed on 5 different body positions: wrist, hip, ankle, upper arm, and thigh. Accelerometer data from each body position and combinations of positions were used to develop random forest models to assess activity category recognition accuracy and MET estimation.

RESULTS

Model performance for both MET estimation and activity category recognition were strengthened with the use of additional accelerometer devices. However, a single accelerometer on the ankle, upper arm, hip, thigh, or wrist had only a 0.03-0.09 MET increase in prediction error compared with wearing all 5 devices. Balanced accuracy showed similar trends with slight decreases in balanced accuracy for the detection of locomotion (balanced accuracy decrease range 0-0.01), sedentary (balanced accuracy decrease range 0.05-0.13), and lifestyle activities (balanced accuracy decrease range 0.04-0.08) compared with all 5 placements. The accuracy of recognizing activity categories increased with additional placements (accuracy decrease range 0.15-0.29). Notably, the hip was the best single body position for MET estimation and activity category recognition.

CONCLUSIONS

Additional accelerometer devices slightly enhance activity recognition accuracy and MET estimation in older adults. However, given the extra burden of wearing additional devices, single accelerometers with appropriate placement appear to be sufficient for estimating energy expenditure and activity category recognition in older adults.

Cut points of the Actigraph GT9X for moderate and vigorous intensity physical activity at four different wear locations

Accelerometer cut points are an important consideration for distinguishing the intensity of activity into categories such as moderate and vigorous. It is well-established in the literature that these cut points depend on a variety of factors, including age group, device, and wear location. The Actigraph GT9X is a newer model accelerometer that is used for physical activity research, but existing cut points for this device are limited since it is a newer device. Furthermore, there is not existing data on cut points for the GT9X at the ankle or foot locations, which offers some potential benefit for activities that do not involve arm and/or core motion. A total of N = 44 adults completed a four-stage treadmill protocol while wearing Actigraph GT9X sensors at four different locations: foot, ankle, wrist, and hip. Metabolic Equivalent of Task (MET) levels assessed by indirect calorimetry along with Receiver Operating Characteristic (ROC) curves were used to establish cut points for moderate and vigorous intensity for each wear location of the GT9X. Area under the ROC curves indicated high discrimination accuracy for each case.

Effects of using activity diary for goal setting in occupational therapy on reducing pain and improving psychological and physical performance in patients after total knee arthroplasty: A non-randomised controlled study

Background

Psychological factors have been reported to affect chronic pain and may lead to inactivity after total knee arthroplasty. This study aimed to determine whether the use of an activity diary for goal setting during occupational therapy would reduce pain, and improve psychological and physical performance in patients after total knee arthroplasty.

Methods

A total of 41 total knee arthroplasty participants from two cohorts were recruited in the study and allocated by convenience to either the experimental group using an activity diary (n = 20) or the control group (n = 21). Occupational therapy intervention (1–2 weeks postoperatively) to promote goal achievement was performed in both groups, and self-monitoring was performed in the diary group by using the activity diary. The outcome indices were Canadian Occupational Performance Measure, pain (resting pain, walking pain), pain catastrophizing (rumination, helplessness, and magnification), anxiety, depression, pain self-efficacy, and physical activity level. Data were evaluated by using analysis of variance analyses with post hoc tests.

Results

A time-by-group interaction emerged for Canadian Occupational Performance Measure, walking pain, pain catastrophizing, anxiety, depression, and physical activity level (p < 0.05), both favouring the diary group. The diary group also showed greater improvement in Canadian Occupational Performance Measure, walking pain, anxiety, and physical activity levels at four weeks postoperatively, compared to the control group (p < 0.05).

Conclusion

The use of the activity diary in this study increased occupational therapy effectiveness, reduced patients’ pain, and prevented a decline in physical performance. We believe that the use of an activity diary is an effective and feasible addition for total knee arthroplasty patients.

Methodology and validation for identifying gait type using machine learning on IMU data

With the rising popularity of activity tracking, there is a desire to not only count the number of steps a person takes, but also identify the type of step (e.g., walking or running) they are taking. For rehabilitation and athletic training, this difference is important to the prescribed regiment. Fourteen healthy adults walked, jogged and ran on a treadmill at three different constant speeds (1.21, 2.01, 2.68 m/s) for 90 s. An inertial measurement unit (IMU) with accelerometer and gyroscope was affixed to their left ankle. Collected acceleration and angular velocity data were partitioned into individual time-normalised strides. These data were used as features in the artificial neural network (ANN) that classified the type of stride. Several ANN models were tested: using only acceleration, only angular velocity and both. Using primarily acceleration data in the trained ANN yielded the best results (>94% correct stride-type identification) after cross-validation. The ANN models were able to accurately classify the gait type of each stride using a single wearable IMU. The accuracy of the method should improve further as more data is added to the ANN training.