Fig 1.
Set-up of the validation experiment: 23 opto-electronic markers (Optotrak).
(A) and 11 inertial sensor units of the PowerGove [12] (B) were placed on the hand and finger segments of the thumb, index and middle finger. Sensor units on the ring and small fingers were not applied in this study. The arm was placed on a custom-made arm-rest with a palmar position of the hand of 45deg. The directions of the segment axes are described in (C).
Table 1.
Finger movement tasks.
Fig 2.
Examples of finger kinematics and positions during various finger tasks, measured simultaneously with the PowerGlove (solid lines) and opto-electronic markers (Optotrak system, dashed lines).
(A) index PIP/DIP flexion task with the largest joint flexion in the PIP joint; (B) fast task: index MCP/PIP/DIP flexion task with the largest joint flexion in the PIP joint; (C) thumb/index tapping task with wrist flexion, showing the index finger joint angles with the largest joint flexion in the MCP joint; (D) thumb/index tapping task without wrist flexion, showing the amplitude between the tip of the thumb and index finger; (E) index mcp ab/adduction task, showing the MCP joint ab/adduction angle; and (F) index circular pointing task showing both MCP joint flexion and ab/adduction angles.
Table 2.
Outcomes (median and standard deviation) of the comparison of the PowerGlove versus the opto-electronic marker system for the index finger during various finger kinematic tasks and for the thumb tip position with respect to the index fingertip position.
Fig 3.
Boxplots of root mean square differences (dRMS) for index finger kinematics (mcp/pip/dip joint angles) between the PowerGlove and the opto-electronic marker system of all 3 included subjects separately per task (1,2,3).
Flexion task boxplot includes 15–23 trials per subjects (A), fast task boxplot includes 5–9 trials per subject (B), ab/adduction task boxplot includes 1–2 trials per subject (C) and circular pointing task boxplot includes 2–3 trials per subject (D).
Fig 4.
Boxplots of relative marker movement of the 3 markers on a finger segment (averaged over proximal, medial and distal phalanges of the index).
(A); and the mean effect on the segment orientation measured with the opto-electronic marker system (averaged over the 3 phalanges) (B) (flex = index finger flexion tasks; fast = fast flexion tasks; abad = ab/adduction task; point = circular pointing task).
Fig 5.
Root mean square difference (dRMS).
(A) of the norm of angular velocity as measured directly with the gyroscopes of the PowerGlove versus angular velocity determined from the segment orientation derived from the PowerGlove algorithm after processing of the data (expressed as percentage of maximal angular velocity) for the different movement tasks (flex = flexion tasks; fast = fast flexion tasks; abad = ab/adduction task; point = circular pointing task). (B): example of the different signals during a fast flexion task showing good correspondence.