10:30
Movement Sensing - Balance - Fall Detection
Chair: Herman van der Kooij
10:30
15 mins
|
3D WIRELESS MEASUREMENT OF SCAPULAR KINEMATICS
Josien van den Noort, Suzanne Wiertsema, Karin Hekman, Casper Schonhuth, Joost Dekker, Jaap Harlaar
Abstract: People with shoulder pathologies often show scapular dyskinesis, i.e. alterations in scapular position and motion [1]. To direct interventions aimed at improving scapular position, motion and muscle forces, an objective and reliable measurement of scapular dyskinesis in clinical setting is important. However, currently available measures are not reliable, not objective, clinically not suitable, invasive, or limited to static positions (e.g. visual based scapular dyskinesis tests, optoelectronic markers, scapula locators or bonepins). An inertial and magnetic measurement system (IMMS: small, wireless sensors containing gyroscopes, accelerometers, and magnetometers) could be a good alternative to measure scapular motion in clinical routine [2]. The aim of this study is to evaluate the intra- and inter-observer reliability (i.e. the precision) of the measurement protocol using IMMS for of scapular motion in 20 healthy subjects.
For intra-observer reliability, scapular motion of 20 healthy subjects without shoulder problems (age 36±11, BMI 22±2) were measured with the IMMS by a physical therapist on two different days (T0 and T1). For inter-observer reliability, a second physical therapist also measured the scapular motion with the IMMS at T1. Four IMMS sensors were placed on the body: scapula (edge of spina), thorax, upper and lower arm. The subjects repeatedly elevated their arm in the sagittal plane (anteflexion) and in the frontal plane (abduction). Kinematics of the scapula with respect to the thorax were measured. Reliability was assessed by calculating the intraclass correlation coefficient (ICC), the standard error of measurement (SEM) and the smallest detectable difference (SDD) of scapular kinematics at 0, 30, 60, 90, 120 and 150 deg of arm elevation [3]. Furthermore, the differences in range of motion (dROM) and the mean differences (to determine the offset) over the entire movement curves were calculated.
The intra- & inter-observer reliabilities of the 3D scapular motion were equal for both the anteflexion and abduction movement (SEM 4±1deg, dROM 4±3deg), and similar movement patterns were observed between the measurements. ICC’s were 0.73±0.12. The highest SDD was found for the inter-observer reliability (17±2deg) in the transversal plane (scapular re/protraction), mainly caused by an offset, i.e. a constant difference between the curves (range 0-22deg, average 8±6deg).
Using the IMMS, measurement could be performed objectively and easily during dynamic tasks as part of clinical routine. Measurement errors are similar to studies using wired IMMS (SEM 1-5deg [2]) or optoelectronic markers (SEM 3-8deg [4]). The use of a scapula locator for optimization of anatomical calibration of the scapular IMMS-sensor might reduce the offset. In the near future, patients measurements will be performed to assess reliability and validity of the IMMS for scapular dyskinesis. This will show whether measurement errors are small enough to enable scapular dyskinesis assessment.
REFERENCES
[1] Warner et al., Clinical Orthopaedics and Related Research 285 (1992) 191-199
[2] Parel et al., Gait & Posture 35 (2012) 636-640
[3] de Vet et al., Journal of Clinical Epidemiology 59 (2006) 1033-1039
[4] Van Andel et al., Gait & Posture 29 (2009) 123–128
|
10:45
15 mins
|
ESTIMATION OF HAND AND FINGER KINEMATICS USING INERTIAL SENSORS
Henk Kortier, Martin Schepers, Victor Sluiter, Peter Veltink
Abstract: It is the objective within the STW PowerSensor project [1] to develop a glove that is able to assess the dynamic interaction of the human hand with its environment in an ambulatory setting. In the first project phase we will concern about measuring hand and finger kinematics. Current instrumented glove systems are often based on resistive or optical sensors that are placed across the various joints of the human hand and therefore often sensitive to misalignments [2]. We present a new kinematic glove that is equipped with inertial sensors placed on top of the phalangeal segments that can be easily customized for subject specific hand dimensions.
The hand, finger and thumb phalanges are equipped with embedded small-scale 3D rate gyroscope and 3D accelerometer chips. In addition, we instrumented the back of the hand and all fingertips with a 3D magnetometer. Using a custom build data acquisition system, data is transferred to a computer via USB and either stored for later analysis or directly processed.
The integrated output of 3D rate gyroscopes provides an estimate of the orientations of respectively hand and phalanges. Subsequently, by knowledge of various phalangeal lengths we can construct the total transformation matrix, which provides us the positions of all joints and fingertips expressed in a common coordinate frame. As the estimation method is prone to orientation drift, we frequently apply measurement updates using knowledge of joint dimensionalities and the direction of the vertical and magnetic north. The kinematic prediction and correction steps are fused into an error Extended Kalman Filter (EKF) algorithm that is able to provide a real time estimate of the total kinematics.
Preliminary results showed a good estimate of relative orientations during static situations (<1 deg.). The dynamic accuracy is currently evaluated using an optical reference system.
|
11:00
15 mins
|
COGNITIVE PERFORMANCE AND BALANCE CONTROL IN ELDERLY WITH MOBILITY IMPAIRMENTS
Marjon Stijntjes, Jantsje Pasma, Astrid Bijlsma, Marjet van Vuuren, Carolien Jurgens, Marie-Noëlle Witjes-Ané, Gerard Jan Blauw, Carel Meskers, Andrea Maier
Abstract: Background: Impaired balance control is one of the main risk factors for falls in elderly resulting in loss of independence. The association with cognitive performance is of direct clinical interest. We aimed to establish the association of cognitive performance with common measures of balance control in elderly with mobility impairments.
Methods: Cross-sectional study design consisting of 197 patients (mean age (SD) 81.9 (7.1) years) who were referred to the geriatric outpatient clinic for reason of mobility impairments. Actual physical performance was assessed extensively using common clinical tests. Cognitive performance was assessed by Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA) and Visual Association Test (VAT). The group of patients was dichotomized into a group with low cognitive performance (MMSE < 24 points, MoCA < 23 points and VAT < 3 points based on clinically used cut off values) and normal cognitive performance. The ability to maintain ten seconds of quiet stance was tested in side-by-side, semi-tandem and tandem stance, with eyes open and eyes closed. Measurements were performed on a six degrees of freedom force plate to assess Center of Pressure (CoP) movement. Data were analyzed with logistic and linear regression adjusted for age, gender and education.
Results: Short Physical Performance Battery score was (median (interquartile range)) 7 (5-10) out of 12 points and preferred gait speed (mean (SD)) 0.87 (0.29) meters per second. Patients with a low cognitive performance were less able to maintain quiet stance during side-by-side stance with eyes closed compared to the group with normal cognitive performance. This result was consistently found for MMSE (95% confidence interval (CI) 1.17–6.78), MoCA (95% CI 1.17–8.21) and VAT (95% CI 1.02–5.96). Additionally, patients with a low MMSE score were less able to maintain quiet stance during the semi-tandem stance with eyes open (95% CI 1.16–6.86). Cognitive performance was not associated with measures of CoP movement for all test conditions.
Conclusions: In elderly with mobility impairments, low cognitive performance is associated with the ability to maintain quiet stance, especially in conditions without visual control of balance. Confirmation of the clinical importance and absence of the association with measures of CoP movement underlines the urgency for quantitative and causal assessment of the role of cognitive performance in balance control in a clinical setting.
|
11:15
15 mins
|
DAILY PHYSICAL ACTIVITY PATTERNS IN CANCER SURVIVORS: A PILOT STUDY
Josien Timmerman, Marit Dekker, Miriam Vollenbroek-Hutten
Abstract: Background: Low levels of physical activity are associated with low Quality of Life (QoL) and increased chances of morbidity and mortality. In cancer survivors these activity levels have been studied primarily by means of questionnaires, while objective information on actual daily activity levels and their distribution throughout the day is lacking. More insight in actual daily activity is needed to improve post cancer activity behavior advice.
Methods: Daily physical activity was measured with a MTx inertial 3D-motion tracker, for 5 consecutive days from 8 am to 8 pm in both cancer survivors (free from cancer, last treatment ≥ three months ago) and age- and gender-matched healthy controls. Group differences in activity level per day were tested using independent t-tests. In addition, group differences in distribution of activity over the day (morning, afternoon, evening) were analysed using mixed ANOVA. Since chemotherapy is known for affecting physical functioning and activity, associations between having received chemotherapy and objective activity level were explored with point-biserial correlations, and differences in activity patterns between groups (chemo/no chemo/controls) were analyzed separately.
Results: 18 cancer survivors (6 male; mean age 55.7±10.2 yrs) and 18 controls (6 male; mean age 55.2±8.2 yrs) were included. No significant difference was found in daily activity level between cancer survivors (1108±277 counts per minute (cpm)) and healthy controls (1223±304 cpm; p=.129). Also, distribution of physical activity over the day was not significantly different between survivors and controls (p=.191). However, a significant correlation was found between having received chemotherapy and objective daily activity (r=.586; p=.011). Further analysis revealed a trend for decreased activity level in survivors that received chemotherapy (978±195 cpm) compared to controls (1223±304 cpm) and survivors that did not receive chemotherapy (1313±300 cpm; p=.059). Again, no significant difference was found between groups (chemo/no chemo/controls) in distribution of activity over the day (p=.201).
Conclusions: The findings of this study suggest that especially cancer survivors who received chemotherapy as part of their treatment, have low physical activity levels. However, this study did not found support for a deviating daily activity pattern in cancer survivors compared to healthy controls. Since information on daily activity patterns is important for the development of new interventions to improve post cancer activity behavior, additional research is needed to determine if a larger study sample supports the findings of this pilot study.
|
11:30
15 mins
|
CAMERA-BASED FALL DETECTION SYSTEM: PRACTICAL DESIGN ISSUES
Glen Debard, Peter Karsmakers, Mieke Deschodt, Ellen Vlaeyen, Eddy Dejaeger, Koen Milisen, Toon Goedemé, Tinne Tuytelaars, Bart Vanrumste
Abstract: More than thirty percent of persons over 65 years fall at least once a year and are often not able to get up again unaided [1][2]. The lack of timely aid can lead to severe complications such as dehydration, pressure ulcers and death. A camera-based fall detection system can provide a solution. For this, several new algorithms have been proposed in the literature recently [3-6]. However, these algorithms are evaluated almost exclusively on data captured in controlled environments, under optimal conditions (simple scenes, perfect illumination and setup of cameras), and with falls simulated by actors.
In contrast, we collected a dataset based on real life data, recorded at the place of residence of three older persons over several months. We showed that this poses a significantly harder challenge than the “artificial” datasets used in previous studies. The image quality is typically low. Falls are rare and vary a lot both in speed and nature. We investigated the variation in environment parameters and context during the fall incidents. We found that various complicating factors, such as moving furniture or the use of walking aids, are very common yet almost unaddressed in the literature. Under such circumstances and given the large variability of the data in combination with the limited number of examples available to train the system, we posit that simple yet robust methods incorporating, where available, domain knowledge (e.g. the fact that the background is static or that a fall usually involves a downward motion) seem to be most promising. Based on these observations, we propose a new fall detection system. It is based on background subtraction and simple measures extracted from the dominant foreground object such as aspect ratio, fall angle and head speed. All tests are executed using real life data, which has been recorded at the home of 3 elderly, containing 24 falls. Experiments indicate that a fall detector based on a combination of aspect ratio, head speed and fall angle is preferred. The preliminary detector still has a substantial false alarm rate with a precision of 0.26(±0.07) and a promising recall of 0.9(±0.2). We discuss these results, with special emphasis on particular difficulties encountered under real world circumstances. More details can be found in [7].
|
11:45
15 mins
|
ESTIMATING FALL RISK AND STABILITY FROM TRUNK ACCELERATIONS OF DAILY LIFE ACTIVITIES
Sietse Rispens, Kimberley van Schooten, Mirjam Pijnappels, Andreas Daffertshofer, Peter Beek, Jaap van Dieën
Abstract: Algorithms, based on movement patterns obtained from inertial sensors, can be used to quantify dynamic instability [1] and increased fall risk [2]. As it stands, fall risk can only be estimated under controlled conditions and on a substantial set of steady-state data. However, features extracted from trunk accelerations during daily life might improve risk estimations, since these features are based on situations in subjects’ real life and they can be complementary to questionnaire or laboratory based fall risk predictions. This study therefore aims to discover what measures, based on daily life trunk accelerations, are reliable and associated to self-reported fall history.
One week of trunk acceleration data was obtained with the MoveMonitor (McRoberts) in 29 older adults. From these data, we extracted the locomotion episodes, from which a number of measures, including maximum Lyapunov exponents, inter-stride variabilities and spectral features were calculated. Associations between these measures and self-reported number of falls in the preceding year were calculated by Pearson’s correlation. Reliability of the measures obtained from walking episodes during weekly activities was assessed by intra class correlations (ICC), comparing first and second measurement weeks of 56 subjects.
The best association with fall risk was shown by the relative amount of spectral power below 0.5 Hz in the medio-lateral accelerations, having a Pearson’s correlation of 0.62 (p=0.0004) with the self-reported number of falls in the preceding year and an ICC of 0.80. Other features showing a substantial correlation with fall history were the median stride time (correlation 0.38 (p=0.047), ICC 0.92) and the median duration of the locomotion episodes (correlation 0.41 (p=0.030), ICC 0.52).
From the results of our study we conclude that daily life trunk accelerations contain reliable information, which, if validated by future research, may contribute to fall risk assessment.
[1] J. B. Dingwell, et al., "Local dynamic stability versus kinematic variability of continuous overground and treadmill walking," Journal of Biomechanical Engineering-Transactions of the Asme, vol. 123, pp. 27-32, Feb 2001.
[2] M. J. P. Toebes, et al., "Local dynamic stability and variability of gait are associated with fall history in elderly subjects," Gait & Posture, vol. 36, pp. 527-531, 2012.
|
|
