Comparison of the effects of and usability of active and active-assistive rehabilitation robots for the upper extremity function among patients with stroke: a single-blinded randomized controlled pilot study

Background Robotic rehabilitation of stroke survivors with upper extremity dysfunction yields different outcomes depending on the robot type. Considering that excessive dependence on assistive force provided by robots may interfere with the patient’s active learning and participation, we hypothesized that the use of an active-assistive robot does not lead to a more meaningful difference with respect to upper extremity rehabilitation than the use of an active robot. Accordingly, we aimed to evaluate the differences in the clinical and kinematic outcomes between active and active-assistive robotic rehabilitation among stroke survivors.

suitable for robotic rehabilitation in stroke survivors who can perform voluntary movement.

Introduction
Approximately 30-66% of stroke survivors suffer from upper extremity dysfunction, which leads to impediment of activities of daily living (ADL) and social participation. [1] Various interventions have been applied for upper extremity rehabilitation, and robotic rehabilitation has been recently popularized. [2][3][4] Robotic rehabilitation has potential advantages regarding high repetition of specific tasks and interactivity, leading to active participation with less burden on therapists. [2,5] Recent systematic reviews have suggested the beneficial effects of robotic rehabilitation on upper extremity dysfunction among patients with stroke. [4,6] Veerbeek et al. described that robotic rehabilitation is more beneficial for the improvement of the motor control and strength of a paretic arm, but not for that of ADL, than is conventional therapy. [6] Furthermore, Mehrholz et al. demonstrated that robotic rehabilitation has more beneficial effects on ADL as well as on arm function and muscle strength than does conventional therapy. [4] However, these conclusions should be considered cautiously because the robots that were included in these reviews are heterogenous, that is, 28 and 24 different rehabilitation robots were included in the systemic reviews by Veerbeek et al. and Mehrholz et al., respectively. We recently showed that the use of end-effector and exoskeleton rehabilitation robots led to significant functional outcome differences stemming from the distinct characteristics of robots; this indicates that the differential effects might result from the type of rehabilitation robot used. [7] Nonetheless, there is a lack of studies that examined the differential effects according to the characteristics of robots. If the discrepant effects during upper extremity rehabilitation are understood according to the characteristics of robots, more suitable robotic rehabilitation may be applied and provided to each patient.
Representatively, robotic devices can be classified as active and active-assistive robotic devices according to the training modality, that is, an active robot does not provide assistive force, while an active-assistive robot supplies assistive force when the participant could not make active movements. [8,9] Robotic active assistance is thought to be beneficial for participants without voluntary movement because they can be trained with ideal path or speed. Nonetheless, active assistance using manipulation for upper limb rehabilitation is too complex to be adopted with ease because the upper extremities are composed of several joints and different muscles, which allow movements with multiple degrees of freedom. Moreover, musculoskeletal problems associated with stroke such as spasticity, contractures, deformity, or hemiplegic shoulder pain make the application of robotic assistance more difficult. Additionally, excessive dependence on assistive force might interfere with active learning and participation for participants who can perform voluntary movement. Therefore, we hypothesized that an active-assistive robot does not make a meaningful difference in terms of upper extremity rehabilitation relative to that made by an active robot. Thus, we aimed to explore whether there is a difference in clinical and kinematic outcomes between active and active-assistive robots during robot-assisted upper extremity rehabilitation of patients with stroke showing a Medical Research Council (MRC) scale score of 3 or 4 for the paretic proximal upper limb. In addition, we assessed usability of robotic assistance. To our knowledge, this is the first clinical trial to directly compare rehabilitative effects between active and active-assistive robots.

Methods
Study design approved by the institutional review boards of a hospital, and all participants provided written informed consent before enrollment. Our study was registered retrospectively with ClinicalTrials.gov (NCT03465267).

Participants
The study enrolled 20 patients with upper extremity dysfunction owing to stroke who were admitted in a rehabilitation hospital between March 2017 and December 2017. The inclusion criteria were: (1) age of > 19 years; (2) presence of hemiplegia owing to ischemic or hemorrhagic stroke; (3)  cybersickness, that is, occurrence of nausea or vomiting while seeing a screen.

Intervention
Active-assistive robotic intervention group 7 same virtual reality environment as were those included in the ACAS group.

Outcome Measure
We evaluated FMA to measure impairment, Wolf Motor Function Test (WMFT) to measure activity, Stroke Impact Scale (SIS) to measure participation, according to the International Classification of Functioning, Disability, and Health (ICF) concept. [10] To determine more detailed kinematic outcomes, smoothness and mean speed were measured. Outcome measures were checked at baseline (T0), after 2 (T1) and 4 weeks of the intervention (T2), and 4 weeks after the end of the intervention (T3).

Primary Outcome
The primary outcome measure was WMFT, which quantifies the upper extremity functional activity using 15 functional tasks. [11] WMFT-score is rated on a 6-point scale, with the score ranging from zero to five; thus, the total score ranges from 0-75. WMFT-time is the sum of the time required to perform all 15 tasks. The higher the WMFT-score or the shorter the WMFT-time, the better the motor activity.

Secondary Outcomes
Secondary outcome measures were FMA score, SIS score, and kinematic data. FMA score, which ranges from 0 to 100, is a quantitative indicator of motor impairment following stroke, with a higher scores reflecting a lower impairment. [12] We used FMA-UE (shoulder, elbow, forearm, wrist, and hand; 33 items, 0-66) and FMA-prox (shoulder, elbow, and forearm; 18 items, 0-36). SIS version 3.0, which is a stroke-specific, self-reported questionnaire, has been applied as a health-related quality of life measurement tool to assess participation. [13,14] We measured eight domains of SIS (strength, hand function, ADLs and instrumental ADLs (ADLs/IADLs), mobility, communication, emotion, memory and thinking, and social participation); the score of each domain ranges from 0 to 100; the higher the score the better the health status. In the present study, four domains (strength, physical, ADLs/IADLs, and social participation) that are more relevant to proximal upper extremity function were selected for secondary outcome assessment. We also determined SIS-overall (sum of scores of all eight domains) and SIS-function (sum of scores of ADLs/IADLs and social participation).
With regard to kinematic analysis for detailed information about impairment, we recorded the position of affected upper extremity using the trakSTAR™ system (Ascension Technology Corp, USA), which measures the movement of an electromagnetic sensor tracing 6 degrees of freedom (x, y, and z axes) at 80 Hz of sampling rate during each reaching movement. In the present study, the sensor was attached at the distal phalanx of the middle finger with double-sided tape, and the wire was fixed to the skin with bandage; the reference transmitter was located behind the participant (Fig. 2). Each patient was asked to sit in a chair in front of a table, the height of which was adjusted such that the elbow is flexed at an approximate angle of 90° in the sagittal plane; however, the distance of the table from the participant was maintained such that comfortable reaching is ensured. Participants practiced the reaching task three times to be familiarized with the experimental setup, which is Subsequently, participants were asked to reach from the base button to one of the three different target buttons, subsequently bringing back the upper limb to the base button at their own comfortable speed. Those movements were repeated nine times (three times to reach each target button in a randomized order) with 1 min of rest between each movement. Patients were instructed to limit trunk movements without a trunk restraint.
Subsequently, two kinematic performance indices were computed on the basis of the position data during reaching: spectral arc length (SAL) and mean speed (MSP). SAL is a dimensionless measure reflecting the smoothness, which was calculated using the arc length of the Fourier magnitude spectrum of a movement speed profile. [15] A higher SAL value indicates a smoother and thus a better movement. [16] It is also known as an important marker reflecting motor recovery of patients with stroke. [17] MSP was calculated by dividing the distance of actual trajectory by the time required for reaching from the base button to each target button.

Usability Study
We assessed the usability of the patients with stroke with on the basis of individual interview at the end of the intervention. Usability was also determined on the basis of interviews conducted by the research physical therapists, who were in charge of the robotic intervention, and physiatrists, who observed the robotic rehabilitation at the end of the present study.

Statistical analysis
We analyzed the participants who completed outcome measurements at T2 at the least. When the results of T3 were not measured, the last observation carried forward method was used; thus, missed outcomes at T3 were filled in with those determined at T2.

Results
A total of 20 patients with stroke participated in the present study from January 1, 2017, to December 31, 2017, and ten participants each were allocated to the ACAS or ACT groups (Fig. 3). One participant of the ACT group dropped out because he was transferred to another hospital without any adverse event; thus, data on 19 participants (10 ACAS group, 9 ACT group) who completed outcome measurements at T2 at the least were analyzed (Table 1).

Primary Outcome
Both the groups showed similar tendencies, that is, WMFT-score improved over the course of 4 weeks of the intervention and declined after its completion, whereas WMFT-time continued to improve over time (Fig. 4). There was a significant effect of time on both WMFT-score (F = 19.754, p < .001) and WMFT-time (F = 7.369, p = 0.002); however, there was no significant effect of group × time interaction on WMFT-score (F = 0.700, p = 0.504) and WMFT-time (F = 0.802, p = 0.457).

Secondary Outcome
There was a significant effect of time on both FMA-UE (F = 6.615, p = 0.004) and FMA-prox (F = 9.746, p < .001) without that of group × time interaction on FMA-UE (F = 0.856, p = 0.434) and FMA-prox (F = 0.388, p = 0.682) (Fig. 4). Furthermore, group × time interaction had a significant effect on SISfunction (F = 4.965, p = 0.013) and SIS-social participation (F = 6.388, p = 0.004), with more improvements in the ACT group than in the ACAS group, but not on SIS scores (Table 2) Table 2 Comparison of the performance between the ACAS and ACT groups at T0, T1 and T2 ACAS group (n = 10) ACT group (n = 9) Time * Group   Variable  T0  T1  T2  T0  T1  T2  F  p-

Usability Study
The usability, in terms of robotic active assistance, mechanical aspects of robot, experience during robotic rehabilitation, and benefits of robotic rehabilitation, was summarized as pros and cons, separately for both the groups in Table 3. Some patients felt that the robotic active assistance was beneficial for their training, as it afforded patient coordination and desirable movement pattern without aggravated compensatory movements of the trunk. However, active assistance was sometimes discordant to the patients' intended movement. The mechanical complexity and high inertia stemming from the manipulator, which provide active-assistive force, make the robotic training more difficult. On the contrary, participants of the ACT group tried to invest more effort to move the limb than did those of the ACAS group, which led to a sense of achievement, fulfillment, and motivation among participants because they could accomplish the tasks without external assistance. The spontaneous and voluntary control of the robot seem be linked to functional improvement in ADL.
The voluntary control of the robot without any ext assistance leads to a feeling of achievement.

Cons
Assistive force sometimes gave the resistance for the intended voluntary movement.
The robotic exoskeleton was too heavy and bulky hampering arm movement.
Assistive force-as-needed function might allow more op movement or the movement that was not possible withou assistance.

Physiatrists and therapists
Pros ACAS robot seems to be better for introducing "ideal smooth and efficient" upper limb movement.
More efforts were required from the participants; thus, motivated voluntary training was fulfilled.

Cons
Assistive force sometimes was not coordinated in terms of timing and context of the virtual environment.
The assistive force caused conflict with the spasticity of participants.
The inertia caused by manipulator was too high for the patients, paradoxically hampering upper limb movement.
Compensatory movements were aggravated, such abnormal posture or overuse of trunk instead of limb because of no assistance from the robots.

Discussion
We demonstrated that the ACAS and ACT rehabilitation robots had distinct effects on different domains among patients with chronic stroke showing an MRC scale score of 3 or 4 for the affected proximal upper extremity muscle strength. In terms of the impairment and activity domains, there were no differences between the two groups. On the contrary, for the participation domain, the ACT rehabilitation robot showed more beneficial effects than did the ACAS rehabilitation robot on SISfunction and SIS-social participation. Kinematic analysis demonstrated that the ACT group showed better lasting effects on smoothness, while the ACAS group showed immediate effects on speed.
On the basis of our finding, we could say that our results represent the effects of active-assistance during robotic rehabilitation, because other factors of each group, such as the dose (time), task (three-dimensional task), software platform (game-based virtual reality environment), and mechanical structure (exoskeleton type), were comparable. There have been few studies focusing on robotic rehabilitation using assistive force. A previous study compared active-assistive robotic reaching training and non-robotic free-reaching training. [18] In terms of clinical outcomes, no between-group differences were found. On the contrary, kinematic analysis demonstrated that active-assistive robot training improves the smoothness but not the range of motion and straightness, indicating the subtle effects of active-assistive robot. A recent study compared the effects of robotic path assistance and/or weight support on upper extremity kinematics among patients with stroke. [19] They showed that path assistance led to a faster movement in the high functioning group and that a combination of path assistance and weight support led to a smaller error in the low functioning group. However, path assistance was not superior to weight support alone with regard to upper extremity kinematics of especially the lower functioning group, when considering a trade-off between speed and error.
Collectively, results of the previous studies and the current study indicate that ACAS and ACT rehabilitation robots showed no difference in their effects on clinical measures of parameters including impairment and activity, but they have distinct kinematic effects. There might be several explanations for these findings. First, our study population had an ability to move their affected arm without necessarily requiring external assistance although some patients of the ACT group said that robotic active assistance might be more helpful for their training intensity and quality. Second, the active-assistive function was not excellent enough to alleviate the fundamental issues of the upper limb function; therefore, other impairment or activity cannot be attributed to its effects. [20] Robotic assistance was not well coordinated with the motion of participants, thus impeding the intended voluntary movement in some patients, especially among those with spasticity. In a similar vein, the therapists who participated in this study emphasized that the alignment of axis is important to minimize those conflicts. Third, a higher inertia owing to the weight of the manipulator that supplied assistance, hampered the patient's movement, thereby offsetting the effect of the assistance. Fourth, the kinematic analysis had detected distinct feature that were not explained by clinical scale scores. [21] Notably, the ACT rehabilitation robot showed more beneficial effects with respect to SIS-function and SIS-social participation than did the ACAS rehabilitation robot. Active assistance could induce "motor slacking" of participants, which is a tendency to minimize metabolic and movement-related costs, thereby preventing active participation and simultaneously developing dependence on the robot. [22] Motor slacking also possibly affects motivation, attention, effort, and active engagement, which are related to motor cortex excitability and motor plasticity. [20,23] Robotic assistance of the ACAS group decrease the loads on the participants' motor systems, which impedes the learning of the fundamentals essential for performing the task. [20] On the contrary, the ACT group might experience more achievement, resulting in the improvement of participation, as reflected by SIS, but not that of impairment and activity, as reflected by FMA and WMFT. Similar results were found by a previous study that used a self-powered robot, which manipulated the participants' affected arm using their unaffected arm and induced a higher degree of muscle activation in the affected arm than did externally powered robots, indicating the role of active participation. [22] In addition, those active engagements might induce learning and lasting effects, as shown by the lasting effects of smoothness after intervention in the ACT group.
There are four limitations to this study. First, for the ACT group, we used a robot that supports the limb with gravity compensation. Nonetheless, most rehabilitation robots provide weight support of the

Conclusion
Our findings suggested that the active-assistive robot did not provide a significantly higher advantage than did the active robot with regard to improvement of impairment and activity. It had a rather lower effects on participation, although there were differences with regard to kinematic results. Moreover, considering the complex nature and high price of active-assistive robots, active robots could provide

Consent for publication
Consent for publication is acquired by all the authors.

Availability of data and materials
The dataset used in the present study is available from the corresponding author on reasonable request. study design, prepared the research protocol, interpreted the data, and wrote and edited the manuscript. All authors have read and approved the final manuscript.    Supplementarytable.docx