Abstract
The purpose of this study was to
quantitatively compare the difference between unaffected arm and affected arm
of subjects after stroke in voluntary elbow tracking movements. Six subjects were recruited to perform voluntary elbow flexion and extension by following sinusoidal trajectories at different angular velocities (10, 20, 30, 40, 50, and 60 deg/s) on the
horizontal plane. From experiment results, there was an increase of root mean square error (RMSE) in both affected arm and unaffected arm with the
increase of tracking velocity. The RMSEs of unaffected arm were 
deg respectively from 10 deg
/s to 60 deg/s, which is much lower than those of the affected arm (The RMSEs of affected arm were 
deg respectively
from 10 deg/s to 60 deg/s). The
result showed the RMSE can be a useful performance index to evaluate the elbow
control function of subjects after stroke during the tracking test.
1. INTRODUCTION
Clinical scales, such as Ashworth score and Fugl-Meyer scale were often used to evaluate upper limb deficits from different aspects [1, 2]. However, most of these clinical
scales are a semi-quantitative method, which may not be very sensitive to measure and detect small change
during the rehabilitation.
In order to find the better treatment strategies and accurately
evaluate the upper limb improvement of neruomusculoskeletal system during
rehabilitation training, the upper limb function is better to be investigated comprehensively, extensively and quantitatively.
Trajectory-tracking has been a useful tool to evaluate sensorimotor
control function of upper limb which coupling both perception-action and motor
execution [3, 4]. Furthermore, trajectory-tracking
tasks provide a uniform scale for nervous
system to follow and corrective by the feedback across different subjects.
The main objective of this study is to
systematically identify the elbow control ability of subjects after stroke during voluntary elbow tracking tasks, RMSE between elbow angle and target angle is used as a performance index.
2. METHODS
2.1 Experiment procedures
A group of six subjects with stroke (four males and two females) were recruited in this study. The mean age of
subjects was 
years with the
range of 37 – 57 years. Before the test, all subjects were introduced with
the experiment protocol and gave informed consents which followed the procedures established by The Hong Kong
Polytechnic University. In the experiment, subjects were asked to sit beside the
table with the manipulandum (Fig. 1). A strap was used to fix the upper arm to a supporter on the table.
The height of the table was adjusted so that the arm was in horizontal plane with the same height of the shoulder. The
shoulder was in 
abduction and 
flexion. The forearm was
attached to a manipulandum and
the axis of rotation was in line with the elbow joint. There was a computer screen in front of the subjects, which displayed both target and measured elbow joint angle. When the indicator light on the screen turned green, the trial started and the pointer slide moved with
the defined speed in a
sinusoidal trajectory from 
to 
for 36 seconds (elbow angle was defined as 
when the elbow was in full extension). The subjects were instructed to start at
and try their best to follow the target. The measured elbow
angle was displayed in another pointer slide to form the real time feedback.
Fig. 1 Block diagram of experiment setup
Subjects were provided 18 trials structured
in 3 sections, each section consisted of six trials with different angular velocities
(10, 20 30, 40, 50, 60 deg/s) arranged in a random sequence. The subjects had a 30 s rest between each trial and 5 minutes rest between each section. All subjects performed the task with both affected
and unaffected arms.
2.2 Data Analysis
Angular
displacement of the elbow joint was recorded by a flexible electrogoniometer (
RMSE=
(1)
) and the RMSE of affected (*) arm at six velocities (10, 20,
30, 40, 50, 60 deg/s) during elbow tracking movement. Vertical bars indicate standard
deviation. 
deg respectively from 10 deg /s to 60 deg