Kinematic Analysis Using 3D Motion Capture of Drinking Task in People With and Without Upper-extremity Impairments

Summary

This protocol describes an objective method to evaluate the movement performance and sensorimotor function of the upper extremity applied to individuals with stroke and healthy controls. A standardized test procedure, kinematic analysis and outcome variables for three-dimensional motion capture of drinking task are provided.

Abstract

Kinematic analysis is a powerful method for objective assessment of upper extremity movements in a three-dimensional (3D) space. Three-dimensional motion capture with an optoelectronic camera system is considered as golden standard for kinematic movement analysis and is increasingly used as outcome measure to evaluate the movement performance and quality after an injury or disease involving upper extremity movements. This article describes a standardized protocol for kinematic analysis of drinking task applied in individuals with upper extremity impairments after stroke. The drinking task incorporates reaching, grasping and lifting a cup from a table to take a drink, placing the cup back, and moving the hand back to the edge of the table. The sitting position is standardized to the individual's body size and the task is performed in a comfortable self-paced speed and compensatory movements are not constrained. The intention is to keep the task natural and close to a real-life situation to improve the ecological validity of the protocol. A 5-camera motion capture system is used to gather 3D coordinate positions from 9 retroreflective markers positioned on anatomical landmarks of the arm, trunk, and face. A simple single marker placement is used to ensure the feasibility of the protocol in clinical settings. Custom-made Matlab software provides automated and fast analyses of movement data. Temporal kinematics of movement time, velocity, peak velocity, time of peak velocity, and smoothness (number of movement units) along with spatial angular kinematics of shoulder and elbow joint as well as trunk movements are calculated. The drinking task is a valid assessment for individuals with moderate and mild upper extremity impairment. The construct, discriminative and concurrent validity along with responsiveness (sensitivity to change) of the kinematic variables obtained from the drinking task have been established.

Introduction

Kinematic analysis describes the movements of the body through space and time, including linear and angular displacements, velocities, and accelerations. The optoelectronic motion capture systems use multiple high-speed cameras that either send out infra-red light signals to capture the reflections from passive markers placed on the body or transmit the movement data from active markers containing infrared emitting diodes. These systems are considered as 'gold standard' for the acquisition of kinematic data¹. These systems are valued for their high accuracy and flexibility in measurements of diverse tasks. Kinematic measures have shown to be effective in capturing smaller changes in movement performance and quality that may be undetected with traditional clinical scales^2,3. It has been suggested that kinematics should be used for distinction between true recovery (restoration of premorbid movement characteristics) and the use of compensatory (alternative) movement patterns during the accomplishment of a task^4,5.

Upper extremity movements can be quantified using end-point kinematics, generally obtained from a hand marker, and angular kinematics from joints and segments (i.e., trunk). End-point kinematics provide information about trajectories, speed, temporal movement strategies, precision, straightness, and smoothness, while angular kinematics characterize movement patterns in terms of temporal and spatial joint and segment angles, angular velocities, and interjoint coordination. End-point kinematics, such as, movement time, speed, and smoothness are effective to capture the deficits and improvements in movement performance after stroke^6,7,8 and angular kinematics show whether the movements of joints and body segments are optimal for a specific task. Kinematics from people with impairments are often compared with movement performance in individuals without impairments^8,9. End-point and angular kinematics are correlated in a way that a movement performed with effective speed, smoothness, and precision will require good movement control, coordination, and use of effective and optimal movement patterns. For example, a patient with stroke who moves slowly usually also shows decreased smoothness (increased number of movement units), lower maximum velocity, and increased trunk displacement⁸. On the other hand, improvements in endpoint kinematics, such as movement speed and smoothness might occur independently from the changes of compensatory movement strategies of trunk and arm¹⁰. It has been established that kinematic analysis may provide additional and more precise information about how the task is accomplished after an injury or disease, which in turn is essential for individualized effective treatment to reach optimal motor recovery¹¹. Kinematic analysis is increasingly used in clinical studies to describe the movements in people with upper extremity impairments after stroke^8,9, to evaluate motor recovery^7,12,13 or to determine the effectiveness of therapeutic interventions^10,14.

Movement tasks often studied in stroke are pointing and reaching, although the use of functional tasks that incorporate manipulation of real everyday objects is increasing¹. Since kinematics of reaching depend on the experimental constraints such as the selection of objects and the goal of the task¹⁵, it is essential to assess movements during purposeful and functional tasks in which the real difficulties in individual's daily life will be reflected more closely.

Thus, the aim of this paper is to provide a detailed description of a simple standardized protocol used for kinematic analysis of a purposeful and functional task, drinking task, applied to individuals with upper extremity impairments in acute and chronic stages after stroke. Results from the validation of this protocol for individuals with moderate and mild stroke impairment will be summarized.

Protocol

All methods described here have been part of the studies approved by the Regional Ethical Review Board in Gothenburg, Sweden (318-04, 225-08).

1. Setting up the Motion Capture System

1. Mount 4 cameras on the wall approximately 1.5 - 3 m away from the measurement area at the height of 1.5 - 2.5 m facing the measurement area. Mount one camera on ceiling just above the measurement area (Figure 1). Start the camera system.
2. Place the L-shape calibration frame on the table with the short axis in line with the edge of the table and the long axis pointing forward.
   NOTE: The coordinate system is defined with X-axes directed forward (anteriorly in the sagittal plane), Y-axis directed laterally (in the frontal plane), and Z-axis directed upward (superiorly, perpendicular to the transversal plane).
3. Open the 3D tracking and data acquisition software (Track Manager), start calibration by selecting Capture | Calibrate, enter the calibration time of 30 s and click OK.
4. Move the wand in all directions throughout the entire measurement area (75 × 75 × 65 cm) above the chair and table to ensure that all 5 cameras capture the wand in as many orientations as possible^16,17. After the calibration, the results show on the screen. Accept calibration residuals below 0.5 mm.
5. Have the subject, wearing a sleeveless top, sit in a in a height adjustable chair with their back against the chair´s back, the upper arm in neutral adducted position, the palm of the hand resting at the table and the wrist aligned to the edge of the table. Check that the knee, hip, and elbow angles are approximately 90°.
6. Place the retroreflective passive markers with double-adhesive tape on the skeletal landmarks¹⁸ on the tested hand (third metacarpophalangeal joint), wrist (styloid process of ulna), elbow (lateral epicondyle), right and left shoulder (middle part of acromion), thorax (upper part of sternum), and forehead (notch between eyebrows).
7. Place two markers on the cup (upper and lower edge).

2. Procedures for Motion Capture of the Drinking Task

1. Place the hard-plastic cup (diameter of 7 cm, height of 9.5 cm) with 100 mL of water 30 cm from the table edge, in the midline of the body. The position of cup on the table is selected deliberately to keep the task performance natural and close to real-life situation.
2. Ask the subject to perform the drinking task in a comfortable self-paced speed by i) reaching and grasping the cup, ii) lifting the cup from the table towards mouth, iii) taking a drink (one sip), iv) placing the cup back on the table behind a marked line (30 cm from the table edge), and v) returning to the initial position with the hand on the edge of the table.
3. Ensure that subject understands the instructions and can reach the cup comfortably with the less-affected arm without leaning forward.
4. Prior each recording, ensure that the start position (initial position) is correct, ask the subject to be ready, start the capture manually and give verbal instruction "you can start now."
5. When the subject finishes the task, stop the recording manually.
6. Record five trials with short pause between each trial (approximately 30 s), starting with the less-affected arm.
7. Check that the data acquisition has been successful (95 - 100% data for each identified marker).
   NOTE: Marker data is automatically transferred in real-time to the data acquisition software (Track Manager). A pre-defined Automatic Marker Identification (AIM) model is used for automatic identification of the markers.
8. When incomplete data are detected, perform extra trials after identifying the problem and adjusting the sitting or marker positions to ensure full visibility of the markers in order to obtain at least 3 successful trials.
   NOTE: Possible problems that may occur are that markers may fall off or they are occluded from the cameras viewing angle, which results in incomplete data. However, the camera and marker set-up, as used in this protocol, produces data loss due to gaps only in very rare occasions. In total, the motion capture session takes approximately 10 - 15 min to complete.

3. Data Analysis

1. Transfer the recorded data from Track Manager directly into Matlab by clicking File | Export | Directly into Matlab.
2. Use the Matlab command at the command prompt: (>> workspace) to see the set of Matlab variables.
   NOTE: Key Matlab variables containing data to be used in the instructions and in creating the analysis are:
   QTMmeasurements.Frames - the number of frames captured
   QTMmeasurements.FrameRate - the number of frames captured per second (240)
   QTMmeasurements.Trajectories.Labeled.Count - number of labels (10)
   QTMmeasurements.Trajectories.Labeled.Labels - labels as defined in the Track Manager
   QTMmeasurements.Trajectories.Labeled.Data - measurement data in a 3D array of 10 x 3 x number of frames, where for each frame and each label the 3 coordinates are recorded
3. In Matlab, filter the x, y, z values using the (butter) and (filtfilt) instructions with a 6Hz second-order Butterworth filter in in both forward and reverse directions, giving a zero-phase distortion and forth-order filtering.
   NOTE: Example
   [b, a] = butter(2, 6/240/2); % Cutoff frequency 6Hz and with respect to ½ sampling frequency
   xfiltered = filtfilt(b, a, QTMmeasurements.Trajectories.Labeled(1,1,:));
4. In Matlab, create a program to use the x, y, z values for each frame sample and each label to calculate the kinematic variables such as tangential velocity of the hand, and joint angles. The kinematic variables are shown in Table 2.
5. In Matlab, create a program to break the sequence of samples into 5 logical phases: reaching, forward transport, drinking, back transport, and returning the hand to the initial position (Figure 2). Definitions for start and end of each phase are showed in detail in Table 1.
6. In Matlab, use the (plot) instruction to create plots of positions, velocities, joints angles, and angle-angle diagrams.

Results

The protocol described in this article has been applied to individuals with stroke and healthy controls^{2,6,8,19,20,21}. In total, kinematic data from 111 individuals with stroke and 55 healthy controls have been analyzed in different studies. The upper extremity impairment after stroke was defi...

Discussion

The protocol can successfully be used to quantify the movement performance and quality in individuals with moderate and mild upper extremity sensorimotor impairments at all stages after stroke. The feasibility of this protocol has been proved in a clinical setting as early as 3 days post stroke, and showed that the system can be used by trained health professional without specific technical qualifications. Technical expertise is, however, needed to create and develop a program for data analysis. From this aspect, the upp...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
5 camera optoelectronic ProReflex Motion capture system (MCU 240 Hz)	Qualisys AB, Gthenburg, Sweden	N/A	Movement analysis system with passive retroreflective markers
Markers	Qualisys AB, Gthenburg, Sweden	N/A	Retroleflective passive circular markers, diameter of 12 mm
Calibration frame and wand	Qualisys AB, Gthenburg, Sweden	N/A	L-shape calibration frame (defines the origin and orientation of the coordinate system); T-shape wand (300 mm)
Qualisys Track Manager	Qualisys AB, Gthenburg, Sweden	N/A	3D Tracking software
Matlab	Mathworks, Inc, Natick, Ca	N/A	Data analysis software