Discussion

The movements of short-term learning depended on the dynamics of both arm and hand-held objects in our study. This study investigated learning effect in a tracking trajectory task, continuous force feedback assistance that afforded continuous monitoring and adjustment of motor output. We focused on two indices to discuss the tracking performance of force feedback assistance, namely, tracking error and variation in the short-term learning. The meaning of the tracking error and variation could be deemed tracking accuracy and tracking stability. First, the subjects of two groups didn’t significantly differ in the earlier evaluation, thus they had the same basic function of right-hand. Furthermore, the subjects of assist group answered that they could correct the deviation easily by the real-time force feedback after experiment.

In the learning stage, because there were different effects of force feedback assistance between four areas, we discussed tracking performance in local aspect.

From the analysis of tracking error, assist group had better tracking accuracy for flexion and extension in comparison with non-assist group. Thus the non-assist group might consume most of their time in adjusting the accurate position for flexion and extension because subjects didn’t have the force feedback assistance. Thus incidence of under-correction or over-correction in the non-assist group was also greater than the assist group. The same results could be found from the analysis of curve fitting, initial tracking error indices were differ for flexion and extension. Subjects couldn’t track a trajectory so straight for flexion and extension [42]. In the phase condition, both groups improved the tracking accuracy significantly for abduction and extension with the increase in motor learning. There were similar learning effect at four movements whether or not force feedback assistance. This could be inferred as below.

37

For abduction, both groups improved the tracking accuracy with learning , since the operation angle at the elbow was approximately from 100∘to 110∘, the better mechanical advantages for the arm and biceps occurred between those angle at the elbow [43] [44]. For extension, although non-assist group improved tracking accuracy, it was still not better than assist group. The reason may be the movements for extension needed greater inertia to move the arm, and the greater inertia the harder it is to tracking trajectory. [45] [46]. For flexion, the effect of inertia influenced both groups more serious. Non-assist group could not learn how to adjust to promote tracking accuracy. It could be emphasized that force feedback assistance could be used, especially for improving the tracking error caused by greater inertia to move arm.

From the analysis of variation, assist group had better tracking stability for adduction, flexion and extension. Non-assist group had worse tracking stability at beginning of tracking task, but after the first block, they gained similar tracking stability as assist group with learning. The results of curve fitting analysis could further confirm that force feedback assistance could help effective enhance tracking stability. Especially for abduction, it had better tracking stability than the other movements and also could improved tracking stability with learning. Consequently, it could be responded to the results of tracking error that operation angle at the elbow was the better mechanical advantages for the arm to do the tracking task. However, for adduction, flexion and extension, there was no improvement for tracking stability.

It may be the intrinsic instability of the hand still not improved and intrinsic instability never disappearance. [47] [42]. On the above discussion, we can know that force feedback assistance theory may be helpful to validate motor tracking task.

Yamamoto and Kitazawa [48] indicate that tactile signals were referred correctly to vertically aligned locations, which are reported to lead to normal judgments. In

38

object manipulation, motor actions of the arm and hand, the motor system receives important sensory information from tactile receptors in the glabrous skin of the volar aspect of the hand [49] [50] [51]. In addition, muscles stabilizing the wrist must also be coordinated with muscles driving the arm when moving handheld loads. Moreover, recent research demonstrated that external focus is more effective than internal focus and examined importance of force feedback from the tool in the referral by manipulating the direction of force feedback in a virtual reality [52] [53] [54] . Computer controlled virtual environments, usually including a robotic manipulandum for force feedback, are often employed in this research as they allow experimenters to precisely control the parameters of their subject’s mechanical and visual environment [55] [56].

This study also discussed whether the learning effect retains after removing the warning of force feedback by comparing the later evaluation with the earlier evaluation in the assist and non-assist groups. First of all, consider the same condition as the learning stage, that is counter-clockwise tracking task without disturbance force, the assist group improved the tracking error and variation after removing the force feedback in the later evaluation. The non-assist group was learning the motor skill by themselves, the tracking performances were also increasing at the later evaluation.

Both groups improved the tracking performance similarly. There is no significant improvement for both groups doing the task with different conditions as the learning stage, such as tracking task at clockwise direction or with disturbance force. After the learning process only the same movement condition can be improved.

Compared the tracking performance at the end of learning stage and the later evaluation in the same condition of task, the tracking error and variation were worse in the later evaluation. That is after the removal of force feedback assistance, motor performance may not maintain as force feedback assistance. We assume that the

39

subject doesn’t control the joystick with force feedback stably in short-term learning.

The subjects needed the force feedback to warn them of the deviation during the last block , even though the tracking error of assist group decreased in the short-term learning. Thus the tracking performance of assist group decreased after removing force feedback.

In the behavior experiment, the external force feedback assistance can reduce the tracking error and variation for motor skill learning especially for flexion and extension. And then we investigated the physiological signals changes under force feedback assistance. However, there was much difference of physiological activity between two groups to make objective comparison the force feedback effect.

Therefore, we designed experiment 2 to compare the non-assist and assist condition in each subject to reduce the difference from subjects’ effect. Since there were both difference for tracking error and variation for flexion and extension, we compared the physiological activity mainly at these sides. The physiological activity, such as oscillatory brain activity and muscle activity during the preparation, execution, and post-movement stage of a tracking task, were investigated.

40

3. Physiological Signals Activities with