Difficulty with turning is a major contributor to mobility disability and falls in people with movement disorders, such as Parkinson’s disease (PD). Turning often results in freezing and/or falling in patients with PD. However, asking a patient to execute a turn in the clinic often does not reveal their impairments. Continuous monitoring of turning with wearable sensors during spontaneous daily activities may help clinicians and patients determine who is at risk of falls and could benefit from preventative interventions. In this study, we show that continuous monitoring of natural turning with wearable sensors during daily activities inside and outside the home is feasible for people with PD and elderly people. We developed an algorithm to detect and characterize turns during gait, using wearable inertial sensors. First, we validate the turning algorithm in the laboratory against a Motion Analysis system and against a video analysis of 21 PD patients and 19 control (CT) subjects wearing an inertial sensor on the pelvis. Compared to Motion Analysis and video, the algorithm maintained a sensitivity of 0.90 and 0.76 and a specificity of 0.75 and 0.65, respectively. Second, we apply the turning algorithm to data collected in the home from 12 PD and 18 CT subjects. The algorithm successfully detects turn characteristics, and the results show that, compared to controls, PD subjects tend to take shorter turns with smaller turn angles and more steps. Furthermore, PD subjects show more variability in all turn metrics throughout the day and the week.

Wearable inertial systems have recently been used to track human movement in and outside of the laboratory. Continuous monitoring of human movement can provide valuable information relevant to individuals’ level of physical activity and functional ability. Traditionally, orientation has been calculated by integrating the angular velocity from gyroscopes. However, a small drift in the measured velocity leads to increasing integration error over time. To compensate that drift, complementary data from accelerometers are normally fused into tracking systems using the Kalman or extended Kalman filter. In this study, we combine kinematic models designed for control of robotic arms with state-space methods to continuously estimate the angles of human shoulder and elbow using two wearable inertial measurement units. We use the unscented Kalman filter to implement the nonlinear state-space inertial tracker. Shoulder and elbow joint angles obtained from 8 subjects using our inertial tracker were compared to the angles obtained from an optical-tracking reference system. On average, there was an RMS angle error of less than 8 $^\circ$ for all shoulder and elbow angles. The average correlation coefficient for all movement tasks among all subjects was $r\ge \hbox{0.95}$ . This agreement between our inertial tracker and the optical reference system was obtained for both regular and fast-speed movement of the arm. The same method can be used to track movement of other joints.

BACKGROUND: Difficulty turning during gait is a major contributor to mobility disability, falls and reduced quality of life in patients with Parkinson’s disease (PD). Unfortunately, the assessment of mobility in the clinic may not adequately reflect

Traditionally, human movement has been captured primarily by motion capture systems. These systems are costly, require fixed cameras in a controlled environment, and suffer from occlusion. Recently, the availability of low-cost wearable inertial sensors containing accelerometers, gyroscopes, and magnetometers have provided an alternative means to overcome the limitations of motion capture systems. Wearable inertial sensors can be used anywhere, cannot be occluded, and are low cost. Several groups have described algorithms for tracking human joint angles. We previously described a novel approach based on a kinematic arm model and the Unscented Kalman Filter (UKF). Our proposed method used a minimal sensor configuration with one sensor on each segment. This paper reports significant improvements in both the algorithm and the assessment. The new model incorporates gyroscope and accelerometer random drift models, imposes physical constraints on the range of motion for each joint, and uses zero-velocity updates to mitigate the effect of sensor drift. A highprecision industrial robot arm precisely quantifies the performance of the tracker during slow, normal, and fast movements over continuous 15-min recording durations. The agreement between the estimated angles from our algorithm and the high-precision robot arm reference was excellent. On average, the tracker attained an RMS angle error of about 3° for all six angles. The UKF performed slightly better than the more common Extended Kalman Filter.

Background and purpose: Recent findings suggest that a gait assessment at a discrete moment in a clinic or laboratory setting may not reflect functional, everyday mobility. As a step towards better understanding gait during daily life in neurological populations, we compared gait measures that best discriminated people with multiple sclerosis (MS) and people with Parkinson’s Disease (PD) from their respective, age-matched, healthy control subjects (MS-Ctl, PD-Ctl) in laboratory tests versus a week of daily life monitoring. Methods: We recruited 15 people with MS (age mean ± SD: 49 ± 10 years), 16 MS-Ctl (45 ± 11 years), 16 people with idiopathic PD (71 ± 5 years), and 15 PD-Ctl (69 ± 7 years). Subjects wore 3 inertial sensors (one each foot and lower back) in the laboratory followed by 7 days during daily life. Mann–Whitney U test and area under the curve (AUC) compared differences between PD and PD-Ctl, and between MS and MS-Ctl in the laboratory and in daily life. Results: Participants wore sensors for 60–68 h in daily life. Measures that best discriminated gait characteristics in people with MS and PD from their respective control groups were different between the laboratory gait test and a week of daily life. Specifically, the toe-off angle best discriminated MS versus MS-Ctl in the laboratory (AUC [95% CI] = 0.80 [0.63–0.96]) whereas gait speed in daily life (AUC = 0.84 [0.69–1.00]). In contrast, the lumbar coronal range of motion best discriminated PD versus PD-Ctl in the laboratory (AUC = 0.78 [0.59–0.96]) whereas foot-strike angle in daily life (AUC = 0.84 [0.70–0.98]). AUCs were larger in daily life compared to the laboratory. Conclusions: Larger AUC for daily life gait measures compared to the laboratory gait measures suggest that daily life monitoring may be more sensitive to impairments from neurological disease, but each neurological disease may require different gait outcome measures.

Although the use of wearable technology to characterize gait disorders in daily life is increasing, there is no consensus on which specific gait bout length should be used to characterize gait. Clinical trialists using daily life gait quality as study outcomes need to understand how gait bout length affects the sensitivity and specificity of measures to discriminate pathological gait as well as the reliability of gait measures across gait bout lengths. We investigated whether Parkinson’s disease (PD) affects how gait characteristics change as bout length changes, and how gait bout length affects the reliability and discriminative ability of gait measures to identify gait impairments in people with PD compared to neurotypical Old Adults (OA). We recruited 29 people with PD and 20 neurotypical OA of similar age for this study. Subjects wore 3 inertial sensors, one on each foot and one over the lumbar spine all day, for 7 days. To investigate which gait bout lengths should be included to extract gait measures, we determined the range of gait bout lengths available across all subjects. To investigate if the effect of bout length on each gait measure is similar or not between subjects with PD and OA, we used a growth curve analysis. For reliability and discriminative ability of each gait measure as a function of gait bout length, we used the intraclass correlation coefficient (ICC) and area under the curve (AUC), respectively. Ninety percent of subjects walked with a bout length of less than 53 strides during the week, and the majority (>50%) of gait bouts consisted of less than 12 strides. Although bout length affected all gait measures, the effects depended on the specific measure and sometimes differed for PD versus OA. Specifically, people with PD did not increase/decrease cadence and swing duration with bout length in the same way as OA. ICC and AUC characteristics tended to be larger for shorter than longer gait bouts. Our findings suggest that PD interferes with the scaling of cadence and swing duration with gait bout length. Whereas control subjects gradually increased cadence and decreased swing duration as bout length increased, participants with PD started with higher than normal cadence and shorter than normal stride duration for the smallest bouts, and cadence and stride duration changed little as bout length increased, so differences between PD and OA disappeared for the longer bout lengths. Gait measures extracted from shorter bouts are more common, more reliable, and more discriminative, suggesting that shorter gait bouts should be used to extract potential digital biomarkers for people with PD.