The rehabilitation effects of the NMES robotic hand and robotic hand were compared.
Both training systems could significantly improve the motor function of upper limb.
The NMES robot was more effective than the pure robot.
NMES applied on distal muscle could benefit the recovery in the entire upper limb.
Upper limb motor deficits are common after stroke, and observed in over 80% of stroke survivors [1,2]. Various rehabilitation devices have been purposed to assist human physical therapists to provide effective long-term rehabilitation programs [, , ]. Among them, rehabilitation robots and neuromuscular electrical stimulation (NMES) are most widely used in stroke rehabilitation practices. Rehabilitation robots have been recognized as efficient in such cases and could represent a cost-effective addition to conventional rehabilitation services because they provide highly intensive and repetitive training [, , , ]. It has been reported that the integration of voluntary effort (e.g. electromyography, EMG) into robotic design could contribute significantly to motor recovery in stroke patients [6,10]. This is because an EMG-driven strategy can maximize the involvement of voluntary effort in the training, and its effectiveness at improving upper limb voluntary motor functions have been proved by many EMG-driven robot-assisted upper-limb training systems [, , ]. However, rehabilitation robots are unable to directly activate the desired muscle groups, which may only assist, or even dominate limb movement such as continuous passive motions (CPM) . In addition, stroke patients usually cooperate with compensatory motions from other muscular activities to activate the target muscles, which may lead to ‘learned disuse’ . However, NMES can effectively limit compensatory motions by stimulating specific muscles via cyclic electrical currents, which provides repetitive sensorimotor experiences . With the advantage of precisely activating the target muscle, NMES has been reported to be effective in evoking sensory feedback, improving muscle force, and thus promoting motor function in stroke patients [17,18]. Nevertheless, training programs assisted by NMES alone are also suboptimal due to the difficulty of controlling movement trajectories and the early appearance of fatigue [19,20].
Accordingly, various NMES robot-assisted upper-limb training programs which combine these two unique techniques have been proposed to integrate the benefits and minimize the disadvantages [7,12,14,21,22]. The rehabilitation effectiveness of these combined systems has been investigated and reported to be effective in improving motor recovery. Several studies have compared the training outcomes of NMES robot-assisted training and other training programs. For example, Qian et al.  reported that NMES-robot-assisted upper-limb training could achieve better motor outcomes when compared with conventional therapies for subacute stroke patients. Meanwhile, another study which compared the training effects between robot-aided training with NMES and robot-aided training solely using the InMotion ARM™ Robot in the subacute period demonstrated that the active ranges of motion of the NMES robot-training group were significantly higher compared with the robot-training group . Coincidentally, investigations into applications in chronic stroke patients have also been carried out. For instance, Hu et al.  proposed an EMG-driven NMES robot system for wrist training; this combined device improved muscle activation levels related to the wrist and reduced compensatory muscular activities at the elbow, while these training outcomes were absent for the EMG-driven robot-assisted training alone. Indeed, a similar study by another research group also achieved better rehabilitation outcomes on some clinical assessments using the combined system compared to robot-assisted therapy alone .
In the literature, most studies on current rehabilitation devices combining the NMES and robotic systems targeted the elbow and wrist joints [7,, , ], while very few focused on the hand and fingers . In addition, a comparison of the training effects for hand rehabilitation between the NMES robot and other hand rehabilitation devices has not yet been adequately conducted. Indeed, the primary upper-limb disability post-stroke is the loss of hand function, and rehabilitation of the distal joints after stroke is much more difficult than the motor recovery of the proximal joints due to the compensatory motions from the proximal joints . Hence, developing effective rehabilitation devices to minimize compensatory movements for hand motor recovery is especially meaningful for stroke rehabilitation. In our previous work, we developed an EMG-driven NMES robotic hand and suggested it for use in hand rehabilitation after stroke . Our device provides fine control of hand movements and activates the target muscles selectively for finger extension/flexion, and its feasibility and effectiveness have been verified by a single group trial . However, whether the long-term rehabilitation effect of this EMG-driven NMES robotic hand is comparable or even better than other hand rehabilitation devices are still unclear and need to be investigated quantitively. Therefore, the objective of this study is to compare the training effects of hand rehabilitation assisted by an NMES robotic hand and by a pure robotic hand though a randomized controlled trial with a 3-month follow-up (3MFU).
This work was approved by the Human Subjects Ethics Sub-Committee of the Hong Kong Polytechnic University. A total of 53 stroke survivors were screened for the training from local districts. 30 participants with chronic stroke satisfied the following inclusion criteria: (1) The participants were at least 6 months after the onset of a singular and unilateral brain lesion due to stroke, (2) both the metacarpophalangeal (MCP) and proximal interphalangeal (PIP) joints could be extended to 180° passively, (3) muscle spasticity during extension at the finger joints and the wrist joint was below 3 as measured by the Modified Ashworth Scale (MAS) , ranged from 0 (no increase in muscle tone) to 4 (affected part rigid), (4) detectable voluntary EMG signals from the driving muscle on the affected side (three times of the standard deviation (SD) above the EMG baseline), and (5) no visual deficit and able to understand and follow simple instructions as assessed by the Mini-Mental State Examination (MMSE > 21) .
This work involved a randomized controlled trial with a 3-month follow-up (3MFU). The potential participants were first told that the training program they would receive could be either NMES robotic hand training or pure robotic hand training, and all recruited participants submitted their written consent before randomization. Then, the recruited participants were randomly assigned into two groups according to a computer-based random number generator, i.e., the computer program generated either “1” (denoting the NMES robotic hand training group) or “2” (the pure robotic hand group) with an equal probability of 0.5 (Matlab, 2017, Mathworks, Inc.). Fig. 1 shows the Consolidated Standards of Reporting Trials flowchart of the training program.
For both groups, each participant was invited to attend a 20-session robotic hand training with an intensity of 3–5 sessions/week, completed within 7 consecutive weeks. The training setup of both groups is shown in Fig. 2. This robotic hand training system can assist with finger extension and flexion of the paretic limb for patients after stroke. In this work, real-time voluntary EMG detected from the abductor pollicis brevis (APB) and extensor digitorum (ED) muscles were used to control the respective hand closing and opening movements, and the threshold level of each motion phase was set at three times the SD above the EMG baseline at resting state . For example, during the motions of finger flexion, once the EMG activation level of the APB muscle reached a preset threshold, the robotic hand would provide mechanical assistance for hand closing. Similarly, during the motions of finger extension, the robotic hand would assist with hand opening when the EMG activation level of the ED muscle reached a preset threshold. For the NMES robot group, synchronized support from the NMES and the robot were both provided. The NMES electrode pair (30 mm diameter; Axelgaard Corp., Fallbrook, CA, USA) was attached over the ED muscle to provide stimulation during finger extension. The outputs of NMES were square pulses with a constant amplitude of 70 V, a stimulation frequency of 40 Hz, and a manually adjustable pulse width in the range 0–300 μs. Before the training, the pulse width was set at the minimum intensity, which achieved a fully extended position of the fingers in each patient. During the training, NMES would be triggered by the EMG from the ED muscle first and then provided stimulation to the ED muscle to assist hand-opening motions for the entire phase of finger extension, while no assistance from NMES was provided during finger flexion to avoid the possible increase of finger spasticity after stimulation . For the pure robot group, the difference between the two groups was that no NMES was applied in the pure robot group. A detailed account of the working principles of the robotic hand have been described in our previous work [12,30,31].
Continue —-> A comparison of the rehabilitation effectiveness of neuromuscular electrical stimulation robotic hand training and pure robotic hand training after stroke: A randomized controlled trial – ScienceDirect