Introduction

For persons with upper limb disabilities due to stroke, musculoskeletal disorders, or other conditions, there exist a number of validated observational outcome measures to assess functional capabilities.  These types of measures are critical to characterizing the efficacy of a specific treatment or rehabilitation regimen.  Only three of these measures have been specifically developed and validated for individuals with upper limb amputations (Lindner, Natterlund, & Hermansson, 2010; Wright, 2009). Others are not yet validated but have begun to be studied in the upper limb amputee population (Jebsen, Taylor, Trieschmann, Trotter, & Howard, 1969; Mathiowetz, Volland, Kashman, & Weber, 1985; Rider & Linden, 1988; Wright, 2009).  One such example is the Jebsen-Taylor Hand Function test (JHFT) (Jebsen et al., 1969; Rider & Linden, 1988).  In this test, subjects are asked to perform seven activities of daily living (ADLs): writing, page turning, picking up small objects, simulated feeding, stacking checkers, moving large light objects, and moving large heavy objects.  Currently, the standard endpoint for each JHFT task is completion time. The quality of movement of the participant is not taken into consideration.  Because the long-term consequences for prosthetic users will depend not only on the acquisition of certain environmental targets, but on the nature of the movements employed to do so, it is important to incorporate standardized, quantitative movement analysis into the assessment of function for the upper limb amputee population.

One way to acquire detailed analysis of movement is through the use of motion capture (Carey, Jason Highsmith, Maitland, & Dubey, 2008; Gates, Walters, Cowley, Wilken, & Resnik, 2016; Hebert & Lewicke, 2012; Kontson, Marcus, Myklebust, & Civillico, 2017; Major, Stine, Heckathorne, Fatone, & Gard, 2014; Metzger, Dromerick, Holley, & Lum, 2012).  Motion capture enables the quantitative analysis of how segments of the body move relative to each other and to a global coordinate system.  While there are several examples of studies that have investigated the use of this technology for the assessment of upper body function, none have collected kinematic data of subjects performing the JHFT.

In this work, we demonstrate the use of motion capture data to capture quantitative information about how subjects complete the JHFT.  To assess the ability of motion capture to capture CMs during administration of the JHFT, subjects performed these tasks with the right hand under normal conditions, and under braced conditions in which the wrist DoFs were reduced.  For the current analysis, only three of the seven JHFT tasks are presented.  In addition to determining any differences in the kinematic trajectories of subjects under normal and braced conditions, this pilot study also serves to provide a preliminary normative data set for the JHFT tasks.    

Methods

Subjects

For the pilot study, 4 subjects (2 males, 2 females, mean age 24.5 years, SD 1.7 years) with no upper limb disability were asked to perform the JHFT.  All subjects were right-hand dominant.  The study was approved by the Institutional Review Board (Research Involving Humans Subjects Committee) at the U.S. Food and Drug Administration (RIHSC #14-086R).  All subjects provided written informed consent prior to participating in the study.   

Tasks

Each subject completed all seven JHFT tasks twice with the dominant hand while in a seated position.  Each of the tasks (except for Task 1 – writing) requires transport or manipulation of multiple objects.  For the current analysis, only data from Tasks 2 (simulated page turning), Task 3 (picking up small objects), and Task 4 (simulated feeding) are presented.  Future work will incorporate all tasks.  A brief description of these tasks is given below. More detailed descriptions of all the tasks can be found in (Jebsen et al., 1969). 

Data Acquisition and Analysis

An eight-camera passive ViconTM motion analysis system was used to acquire and preprocess the motion data (Vicon, Oxford, UK).  The Bonita B10 motion capture cameras collected data at a sampling rate of 150 Hz. Prior to each data collection session, the motion system was calibrated according to the manufacturer’s guidelines.

Twenty-seven reflective markers were placed on the upper body of each subject to create wrist, forearm, upper arm, head, neck, thorax and pelvic segments.  The table on which the tasks were performed was adjusted such that the subject’s arms were positioned at a 90° angle when resting on the table. After calibrating the subject to the upper body model, the subject was given instructions for each task of the JHFT. 

Three angles (right elbow flexion, right shoulder abduction, torso lateral flexion) were calculated from the Vicon upper body model using YXZ Euler angles derived from the comparison of relative orientations of two segments (Vicon Plug-In-Gait, Oxford, UK).  These data were filtered using a 4th order, zero lag, lowpass Butterworth filter at 6 Hz.  The motion capture data for each subject were segmented into trials: trial start was defined as the initiation of the approach to pick up an object and trial end was defined as the release of the object.  For the three tasks presented in the current work, the last trial from each subject was used to determine the range of motion (RoM) and maximum angle value.  Wilcoxon’s rank sum test was applied to test for statistical differences in completion time, RoM, and maximum angle between normal and braced conditions.

Results

The standard endpoint for the JHFT tasks is completion time.  Figure 1 shows the completion times for subjects performing Tasks 2, 3, and 4 of the JHFT under normal and braced conditions.  Completion time decreased for each task under the braced conditions, with statistically significant differences in completion time between conditions in Task 2 and 3.

Kinematic trajectories for right elbow flexion, right shoulder abduction, and torso left lateral flexion were calculated for the last trial of each task.  RoM and maximum angle from each subject’s trajectory for each task were computed for normal and braced conditions and are shown in Table 1. For Task 2, significant increases in the maximum torso lateral flexion and elbow flexion angles were seen when subjects performed the task under braced conditions. In Task 3, significant increases in both RoM and maximum angle value were observed in all calculated angles, except maximum angle of torso left lateral flexion.  This can be seen in Figure 2 showing the average trajectories for Task 3.  For example, Figure 2A shows elbow flexion remains fairly constant prior to subjects dropping the object in the can under normal conditions.  Under braced conditions, subjects significantly increased elbow flexion at the beginning of the trial in order to pick up the object before transporting and releasing it in the can.  There were no significant differences between normal and braced conditions for Task 4. 

Table 1: RoM and maximum angle for subjects performing Task 2, 3, and 4 of JHFT under normal and braced conditions.  Significant differences between conditions are highlighted in grey.

		Range of Motion		Max Angle
		Normal	Braced	Normal	Braced
R Sho Abduction	Task 2	9.5 ± 6.7	14.6 ± 8.1	46.2 ± 18.0	66.2 ± 21.1
	Task 3	5.8 ± 1.4	11.2 ± 3.1 *	36.8 ± 9.2	54.8 ± 5.3 *
	Task 4	11.6 ± 8.8	21.0 ± 11.5	51.4 ± 21.1	56.4 22.8
R Elb Flexion	Task 2	18.1 ± 6.1	32.3 ± 16.6	93.0 ± 7.8	118.2 ± 7.1 *
	Task 3	17.9 ± 2.3	26.7 ± 3.1 *	100.8 ± 7.7	116.3 ± 5.5 *
	Task 4	25.4 ± 11.3	28.9 ± 16.1	108.0 ± 13.4	110.7 ± 11.8
Torso Left Lateral Flexion	Task 2	1.4 ± 0.8	4.8 ± 3.3	1.1 ± 1.5	3.6 ± 0.9 *
	Task 3	0.9 ± 0.5	3.7 ± 1.3 *	4.3 ± 3.0	6.1 ± 1.9
	Task 4	2.2 ± 2.8	3.0 ± 1.6	1.8 ± 1.8	7.0 ± 4.3

Discussion

Preliminary results for this study indicate that motion capture is able to detect compensatory movement during task performance under braced conditions compared to normal conditions. Significant differences between the conditions were observed in right elbow flexion, right shoulder abduction, and torso left lateral flexion. Although there were no significant differences observed between normal and braced condition for the last trial of Task 4, Table 1 shows that subjects required more RoM at larger angles to complete this task under braced conditions.  With a larger sample size, significant differences may be observed.  Future work will focus on the analysis of other relevant kinematic parameters (e.g. shoulder flexion, shoulder rotation, head flexion, etc.) for all JHFT tasks.  Inter-subject and intra-subject kinematic variability will also be evaluated.