Sage Journals: Discover world-class research

Abstract

Shape memory alloys (SMAs) have an innate property that allows them to return to a memorized form when stimulated. They have been recently designed as flexible fabrics for use in wearable assistive devices. This study aimed to develop a rehabilitation assistive glove with good wearability for a hemiplegic patient using a knitted SMA with a high-unidirectional contractile response. We fabricated warp-knitted SMA actuators, which had higher contractile responses than those of weft-knitted SMA. The use of a knitting loom allowed these actuators to fit all the fingers of a patient. We developed a full-fabric glove, placed the SMA wires on the finger regions to form a tendon-driven structure, and attached knitted SMA actuators to the dorsal and palmar sides of the hand to assist flexion and extension. The joint range of motion after glove actuation increased by 11.620% and 141.710% in flexion and extension, respectively. The range of motion improved more with glove rehabilitation than with thermal stimulation or no actuation. The grip force also increased by 52.400% after rehabilitation exercises. The safety of the heat generated during actuation was confirmed, and its applicability as a thermal stimulus was validated (based on increased skin temperature). These outcomes suggest the practical potential of warp-knitted SMAs in fully fabricated wearable assistive devices.

Keywords

Hand assistive device rehabilitation glove shape memory alloy warp knit wearable device

Introduction

Shape memory alloys (SMAs) are intelligent materials with innate properties that allow them to return to memorized forms caused by a phase transformation induced by an external stimulus.¹ SMAs are extensively used in various applications, including actuators, medical devices, and aerospace machinery.² Among them, nitinol, an alloy of nickel and titanium (Ni-Ti), is the most commonly used alloy that exhibits a shape-memory effect.³ SMAs have a significant advantage in that they have the unique characteristics of shape morphing caused by the shape memory effect. However, they also have disadvantages, such as non-linear behavior, complex control, and low-energy efficiency.⁴ Thus, several attempts have been expended to develop various wearable devices using SMAs.

Numerous researchers have designed SMA wires of various shapes to develop wearable devices that can be worn on the human body. A rehabilitation exoskeleton has been designed based on three-dimensional (3D) printing incorporating an SMA wire to facilitate elbow joint motion in stroke patients with weakened muscle strength.⁵ A pronation/supination axis design has been proposed for an arm using an SMA wire.⁶ The researchers developed an exoskeleton rehabilitation device with three degrees of freedom that enabled complex movements along all three axes regardless of the direction of the elbow and shoulder joints. In addition, a hand exoskeleton robot operated based on a tendon using an SMA wire has been proposed for use in rehabilitation exercises and for assisting people with hand disabilities.⁷ A lightweight wearable device that can apply finger pressure with an SMA wire has also been developed.⁸

Recently, some design methods that are flexible but can maintain a driving force have been proposed to apply SMAs to wearables. In particular, some researchers designed SMA wires as fabrics to minimize the burden on the human body. Several techniques that utilize SMAs as textiles, single wires, and springs have been proposed. Incorporating SMAs into textiles significantly expands the achievable strain range, rendering them adaptable for a variety of applications.⁹ An SMA wire is inserted into the fabric, contractile forces are used to supplement the arm strength of the wearer to lift a barbell (with a weight of up to 4 kg) by wearing it as a garment.¹⁰ Smart pants embedded with SMA wires have been proposed to provide ankle support during walking.¹¹ Morphologically adaptable shells comprising glass fibers and SMA fibers exhibited a diverse range of actuation modes, thereby demonstrating their potential utility as dermal actuators within the field of soft robotics.¹² However, numerous SMA wires need to be inserted to assist and support the human body. Therefore, several researchers have proposed SMA wire design methods as knitted fabrics.^13,14 Topographically self-fitting pants have been proposed by designing an SMA wire in a cylindrical garter-knitted structure,¹⁵ and a gripping aid has been proposed using the bending motion of the SMA plain-knitted structure.¹⁶ In another study, the researchers designed a knitted SMA as a double module to provide bending and contraction simultaneously to design an elbow brace for rehabilitation.¹⁷

Particularly, this knitted structure is well known for its use in producing garments in general and in making fabric by interlooping one or more yarns.¹⁸ Knitting creates continuous stitches by pulling previously formed loops to form new loops.¹⁹ Thus, knitted fabrics have inherently high elasticity, and their properties vary depending on the knitted structure.²⁰ Most knitted SMA-based wearables use a weft-knitted structure.^15–17 SMAs with weft-knitted structures can be transformed into various structures that allow additional compression types beyond uni-directional contraction,¹⁴ but induce low-driving forces. Furthermore, the axial contraction of both the SMA wire and weft rib-knitted SMA, as shown in prior studies, has limitations in wearable devices owing to their low contraction rates. Warp knitting is also extensively used in various fields, such as medicine, engineering, and clothing, and can be designed for the mass production of more complex structures compared with weft knitting.²¹ Thus, in this study, we designed warp-knitted SMA actuators with tendon-driven structures, which have not been attempted before, to provide high shrinkage in the axial direction using relatively small amounts.

The hand comprises many joints, muscles, tendons, and other anatomical structures within a relatively small area; it is agile because of its dynamic anatomical structure, but vulnerable.²² In particular, hemiplegia after a stroke causes loss of upper limb function due to weakness, spasticity and/or contracture, making it difficult for patients to perform activities during their daily lives.²³ Repetitive hand flexion and extension exercises can relieve these symptoms and help restore hand function.²⁴ The use of hand assistive devices for repetitive exercise therapy has a positive effect on the outcome of exercise.²⁵ Alternatively, assistive devices, such as rehabilitation assistive gloves, can improve the function and performance of weakened hands of patients in activities in their daily lives.²⁶ Furthermore, therapy using rehabilitation assistive devices has been demonstrated to be more effective for motor recovery than conventional therapies without assistive gloves.²⁷ Because of the increased interest in treatment using rehabilitation assistive devices, many researchers have recently proposed ways to develop wearable hand assistive devices such as gloves. However, rehabilitation assistive devices with large volumes and rigid materials still limit the continued usage of the product.

Therefore, we aimed to develop a glove with good wearability that can assist the flexion and extension exercise of the hand of a hemiplegic patient by applying the shape-changing technologies of the knitted SMA, leading to high usability. We first proposed a warp-knitted SMA with a high-contraction rate as a tendon-driven structure to show that it can be applied to a rehabilitation glove that can assist hand flexion and extension movements. Thus far, we manufactured the SMA actuators with a high-contraction ratio in the axial direction of the knitted warp-knitted structure using a knitting loom. We then created a glove that can assist the flexion and extension of the hands of hemiplegic patients using a tendon-driven structure. To provide optimal wearability to actual users, we fabricated the wearing part of the glove and the driving part, which is an actuator, with the entire fabric. We hypothesized that the grip strength and range of motion (ROM) of the finger joints in a hemiplegic patient would increase after rehabilitation exercises using our full-fabric rehabilitation glove.

Methods

Materials

We used SMA (DYNALLOY Inc., Irvine, CA, USA), an intelligent material with the ability to return to a memorized form when external stimuli are applied, to develop actuators used in a rehabilitation assistive glove (Table 1). Results from previous studies showed that the knitted structure has shape-changing properties (such as bending and contraction). Accordingly, we attempted to use this SMA wire as an actuator by fabricating knitted structures.^14,17 Specifically, we wrapped threads with 100% polyester (75 deniers, 36 filaments) twice around the SMA wire in the S- and Z-twist directions to prevent SMA wire rotation. This method was used to prevent an electric short circuit and reduce the increase in temperature from Joule heating in the wire during operation.¹⁴ The SMA wire we used was transformed at 70°C and shrank by approximately 4% in the axial direction during operation.¹⁷ The thickness and driving temperature of commercially available SMA wires vary, and SMA wires with lower operating temperatures (e.g., 45°C) are available. However, preliminary test results showed that these wires could not achieve the required driving force, which is an important factor in our rehabilitation assistive gloves. In addition, by wrapping the SMA wire with threads, the temperature transmitted to the outside becomes much lower than the operating temperature. To provide the thermal effect of Joule heating in addition to the driving force of the SMA actuator, a wire with a higher operating temperature range than the appropriate temperature was required; therefore, we used SMA wires with a driving temperature of 70°C.²⁸

Table 1.

Characteristics of the material used.

Material	Composition	Diameter	Transformation temperature (°C)
SMA	55 wt% Ni and 45 wt% Ti	200 μm	70

Note: SMA = shape memory alloy.

Design and fabrication of rehabilitation assistive glove

Knitted Shape memory alloy

We compared the contraction ratios by designing an SMA wire in knitted structures. A weft knit creates a textile by creating loops horizontally, and a warp knit provides vertical connections as loops are crossed (Figure 1).²⁹ Most previous studies that designed SMA wires in knitted structures used weft knits.^14,16,17 Warp knitting cannot be easily used because the knitting process is difficult to implement. However, there is no significant difficulty in mass production because a weft knit can be knitted through existing commercialized knitting machines.³⁰ Thus, we compared the contraction ratio by designing an SMA wire in weft-knitted and warp-knitted structures. The weft-knitted structure consisted of a plain, rib, garter, and I-cord,^14,31 and an open/closed lap was chosen for the warp knit to provide high-contraction actuation (Figure 2).

Figure 1.

Knitted structures: (a) weft, (b) open-lap, and (c) closed lap warp knitting.

Figure 2.

Experimental knitted modules.

Except for the I-cord, which was made using an I-cord knitter machine (Tulip Company Limited, Hiroshima, Japan), all the weft knits (plain, rib, and garter) were made by hand knitting, and all the warp knits were made with a 3D printed knitting loom. A square knitting room was used with one rod on each short side and ten on each long side. The thickness of each circular rod was 3 mm, and ten rods were arranged at 7 mm intervals on the long side (refer to Figure 6(a)). We turned the SMA wire from right to left around the rod on the front long side, sending the wire toward the rod on the back, and repeated this for five rods. Afterward, we used the needle to cover the next row of wires with the previously finished lower wire. The loops of the first row of warp knit were then created. Next, we completed the warp knit by repeating this on the front and back rods alternately. We attempted to obtain the same results by applying uniform force during the hand-knitting process, and we expect that switching to machine knitting will enable the production of more uniform actuators. The detailed dimensions of each module were measured at their most shrunk states. We then measured the contraction ratios using a drafting ruler. The lengths before and after actuation of the warp-knitted SMA actuators were measured three times. Specifically, we measured the length after stretching to the same length of the knitted SMA actuators and then measured the length after contraction by driving it with heat. All the measurements were taken without an additional load. The difference in the lengths before and after actuation was expressed as a percentage. The structure with a higher contraction ratio was then chosen as the actuator for rehabilitation assistive gloves.

Fabrication of rehabilitation assistive glove

We obtained 3D hand shapes for users by employing a 3D human body scanner to develop glove patterns (Artec Eva, Artec 3D, Luxembourg) (Figure 3(a)). For the glove pattern, we utilized a pattern drafting method,³² which flattens the 3D hand shape into two dimensions. This method is suitable when accurate data cannot be extracted by direct hand measurements, such as in patients with hand stiffness or curl. By modifying the completed basic glove pattern, we designed a more extended version of the glove up to the elbow point to accommodate the SMA’s contraction ratio. Before attaching the operating part, we conducted several rounds of prototyping and modified the glove design to enable the gloves to be donned and doffed easily and independently by the patient (Figure 3(b)). Actuators using the knitted SMA with a large contraction rate were then placed on the back of the hand and palm on the developed glove, and the contraction force of the knitted SMA was used as the driving force in a tendon-driven method (Figure 3(c)).

Figure 3.

Schematic diagram of the gloves: (a) three-dimensional hand shape, (b) glove design, and (c) actuator structure.

Evaluation

Participants

We recruited one hemiplegic patient from a local welfare center for the disabled. This study was approved by the participating university’s Institutional Review Board Committee (IRB No. 2112/002-019), and written informed consent was obtained from the participant. The inclusion criteria were as follows: a chronic patient who (1) experienced a stroke at least 6 months ago at the time of data collection, (2) had a hemiplegic upper limb with a modified Ashworth scale (MAS) of level 1 for grading spasticity, and (3) did not have cognitive impairment. We selected this patient as a participant (Table 2) because she met the inclusion criteria, did not receive botulinum toxin for 6 months before the study, does not have uncontrolled internal/surgical diseases, and did not participate in other upper extremity functional rehabilitation studies. In addition, this patient can represent many patients because her MAS grade corresponds to 60% of patients who experience hand spasticity after stroke.³³

Table 2.

Demographic characteristics of the participant.

N = 1	Gender	Age	Hemiplegic side	Onset period
Participant	Woman	73 years	Right	9 years

Experimental process

Sensor attachment

Upon the selected participant’s visit to the laboratory, we placed her hemiplegic hand on a desk and measured the joint ROM of the fingers using flex sensors (Resistive Flex Sensors, Spectra Symbol, Mansfield, TX, USA). We conducted a pre-experiment to ensure alignment between the resistance and angle of the sensor. Subsequently, the sensors measured the amount of bending and can be used as goniometers. The sensors were inserted into the proximal interphalangeal (PIP) joint on the palmar side of the glove and the metacarpophalangeal (MCP) joint on the posterior side of the glove (Figure 4(a)). Owing to the short length of the patient’s thumb and little finger, attaching the sensor was infeasible. Thus, the sensor was omitted from the PIP joint of these fingers. In addition, we attached an iButton temperature sensor (DS1923, Dallas Instrument, Lawrenceburg, KY, USA) on the hemiplegic side to measure changes in the skin temperature and the temperature outside the glove during actuation. Four sensors were attached to the dorsal side of the hand and the middle finger inside and outside the glove (Figure 4(b)). These are the regions in which the knitted SMA modules and SMA wire were inserted. The change in the skin temperature inside the glove was measured to evaluate the safety and thermal stimulation of the Joule heating generated during the actuation of the knitted SMA in the developed glove.

Figure 4.

Sensor attachment regions: (a) flex sensor and (b) iButton.

Functional evaluation of joint range of motion

We measured the joint ROM before and after glove operation as an objective functional evaluation test to explore their effectiveness. We instructed the participant to wear the rehabilitation assistive glove and then flex and extend their hand (i.e., flexion exercise vs extension exercise). At this time, we measured the joint ROM at maximum extension and maximum flexion of the participant’s hands. However, we did not explain the details of the glove operation and actuated the glove randomly during the exercise for objective evaluation. The participant slowly flexed for 8 s and released her hand on the hemiplegic side (with the glove on) based on our instructions, repeating the motion three times. A 1-min rest period was allowed between each flexion movement. After the flexion motion had been completed, a 10-min rest period was allowed until the onset of the extension motion. Subsequently, the participant slowly extended and released the same hand, and the motion was repeated three times. A 1-min rest period was also given between each extension movement. The joint ROM was measured using flex sensors inserted into the glove. Two sets of flexion and extension motions were performed to compare the joint ROM before and after glove actuation.

Rehabilitation exercise assessment

To measure the effects of rehabilitation flexion and extension exercises using the developed glove, we instructed the participant to perform rehabilitation exercises and measured her joint ROM and grip strength. Flexion exercises involve active bending of the finger on the hemiplegic side and the release of the force, whereas extension exercises are used to apply force to spread the finger and then release the force. The time and number of rehabilitation exercises were set according to the condition of the participant with the help of a rehabilitation therapist at the welfare center. Before the rehabilitation exercise, the grip strength and ROM were each measured three times. The grip strength was measured using a hand dynamometer (EH101, Graham, WA, USA) as the highest force value generated by forming the strongest hand grip and maintaining it for 3 s. Joint ROM was measured using flexor sensors before and after flexion and extension, unlike the grip strength. The participant wore rehabilitation gloves and performed flexion and extension rehabilitation exercises without SMA actuation eight times each. After resting for 10 min, flexion and extension rehabilitation exercises were performed eight times during SMA actuation using the same method. We provided heat to the participant’s hands with the developed glove to determine the heat stimulation effects. The same heat was provided when the glove was actuated; however, one side of the SMA actuators was unfixed in the glove to limit the motion assistance of the SMA. These rehabilitation exercises and thermal stimuli were repeated three times. We measured the grip strength and joint ROM three times each. Table 3 summarizes the evaluation process for the rehabilitation exercise.

Table 3.

Summary of rehabilitation exercise.

Step	Before	Rehabilitation exercise (flexion/extension)	After
1	Measurement: grip force and ROM (flexion/extension)	Without SMA actuation	Measurement: grip force and ROM (flexion/extension)
2		Thermal stimuli
3		With SMA actuation

Note: ROM = range of motion, SMA = shape memory alloy.

Data analysis

We analyzed the differences in the joint ROM and skin temperature before and after actuating the developed gloves using the Mann–Whitney U-test to confirm the assistance functionality of the gloves. We then analyzed the differences in the joint ROM and grip force before and after rehabilitation exercise without SMA, rehabilitation exercise with SMA, and using thermal stimulation only based on the Kruskal–Wallis H-test and Mann–Whitney U-test. All analyses were performed using SPSS (version 26.0, IBM, Armonk, NY, USA). The significance level was set at p < .050.

Results

Differences in contraction ratios according to knit structure

We analyzed the differences in contraction ratios according to the SMA knit structures (Figure 5). Table 4 shows that the warp knits, such as open and closed laps, had higher contraction ratios than the weft knits, such as I-cord, rib, and plain knits (18.815 (0.014)% versus 13.804 (0.023)%) (p = .000). Specifically, the contraction ratios of the weft knits exhibited a large trend in the order of the garter, plain, rib, and I-cord knit structures (p = .049). Among the weft knits, the contraction ratio of the I-cord tended to be the lowest (10.859 (0.004)%), and that of the garter knit tended to be the highest (15.698 (0.016)%).

Figure 5.

Results of contraction ratio for SMA knitted structure.

Table 4.

Differences in contraction ratio according to the SMA knitted structure.

Contraction ratio (%, mean (standard deviation (SD))
Weft knit	I-cord	Rib	Plain	Garter	z	p
Weft knit	10.859 (0.004)^b	13.532 (0.014)^{a, b}	15.129 (0.023)^a	15.698 (0.017)^a	7.821	0.049
Warp knit	Open lap		Closed lap		z	p
Warp knit	18.234 (0.004)		19.397 (0.187)		−0.655	0.513

Note: The alphabetical order is the same as the order of mean difference according to the Duncan test; bold item = significant value.

For warp knits, the contraction ratio of the closed loop tended to be higher than that of the open loop (19.397 (0.018)% versus 18.234 (0.004)%); however, the difference was not significant (p = .513). In particular, the warp knits with closed loops have a higher contraction ratio (equal to 3.700%) compared with the contraction ratios for the garter knit used by many researchers.^29,30 In previous studies, weft-knitted actuators were used as assistive devices. However, the weft-knitted structures were knitted perpendicular to the contraction direction. These results indicate that the weft-knitted actuators we developed exhibited strong curling or bending properties, depending on the weft-knitted structures. Conversely, we can infer that the warp-knitted actuators exhibited a higher contraction ratio because they were knitted in the same direction as the shrinkage direction. Thus, we chose a closed-lap structure among the warp knits for our rehabilitation glove’s actuators.

Rehabilitation assistive glove

We knitted ten actuators using a 3D printed knitting loom, five for flexion and five for extension motions. Each actuator was composed of a wire part and a knitted part, and the size of each SMA actuator was determined according to the length of each finger of the user (Figure 6(a)). The base glove to which the actuators were connected was designed to be a long glove up to the lower arm, reflecting the actuator lengths. The palm and dorsal regions of the glove where the knitted SMA actuators were placed were composed of insulating fabric to provide the user with safety due to Joule heating during actuation. In addition, we cut the sideline on the thumb side of the glove and extended the closure using Velcro for ease of donning and doffing by the user who experienced difficulties in moving her hand and arm.

Figure 6.

Rehabilitation assistive glove: (a) SMA actuator and knitting loom, (b) actuators integrated into the glove, and (c) final prototype.

To integrate the actuators into the glove, we first determined the anchoring points at the finger joints and implemented finger extension and flexion based on a tendon-driven mechanism (Figure 6(b)). We fixed the anchoring point with the fabric in the form of a cylinder of the base glove and allowed the wires from the SMA actuators to pass through it. The SMA wire parts were placed on each finger joint, and the knitted parts were placed on the dorsal part of the hand and palm to prevent interference with finger movements. Specifically, the actuators on the dorsal part of the hand and the palm assisted with the extension and flexion of the fingers, respectively, using the contraction characteristics of the SMA actuators during actuation. Table 5 shows the lengths of SMA wire and knit modules inserted into the rehabilitation assistive glove.

Table 5.

Lengths of SMA wires and knit modules inserted into glove.

Length (Wire/Knit, mm)
Region	Thumb	Index	Middle	Ring	Little
Dorsum	9.800/18.000	12.200/19.000	13.000/19.100	11.400/17.150	10.900/17.300
Palm	8.800/18.600	12.900/18.800	11.600/19.900	12.200/18.900	11.600/18.800

Changes in joint ROM after actuation

According to the Mann–Whitney U-test results on the differences in the joint ROMs before and after glove actuation, the joint ROM tended to increase after SMA actuation during flexion and extension (11.620% vs 141.710%), and the increase in joint ROM was significant in the extension (p = .036) (Table 6). Specifically, during flexion, the PIP joint increased by 53.900% after actuation; however, the MCP joint tended to decrease by −34.390%, all of which were not statistically significant (p = .275, .917). During extension, the PIP and MCP joints increased by 223.490% and 82.310%, respectively after actuation compared with their states before actuation; however, these were not statistically significant (p = .050, .251). These results indicate that the increase in finger joint ROMs was more significant when a hemiplegic patient extended her hand with glove actuation than those achieved when she flexed her hand. The decrease in the rate of change in the MCP joint when the patient’s hand flexes can be inferred because the stiff muscles generate resistance when the patient applies force to the hand during flexion motion.

Table 6.

Finger joint ROMs before and after glove operation.

Motion	Finger joint	ROM (°, mean (SD))
Motion	Finger joint	Before actuation	After actuation	Δ (%)	z	p
Flexion	PIP	29.780 (18.500)	45.830 (19.750)	53.900	−1.091	0.275
	MCP	16.400 (20.530)	10.760 (20.140)	−34.390	−0.104	0.917
	Total	21.420 (21.620)	23.910 (20.260)	11.620	−0.420	0.674
Extension	PIP	12.050 (6.470)	38.980 (11.800)	223.490	−1.964	0.050
	MCP	9.950 (6.520)	18.140 (10.790)	82.310	−1.149	0.251
	Total	10.740 (7.300)	25.960 (12.400)	141.710	−2.100	0.036

Note: bold item = significant value.

Changes in joint ROM after rehabilitation exercise

The result of rehabilitation exercises for flexion and extension of the hand of a hemiplegic patient indicates that when there was no actuation of SMA, the finger joint ROMs decreased after flexion and extension exercises (−3.340 (8.490)° versus −2.080 (3.360)°) (Table 7). In contrast, after thermal stimulation and SMA actuation, the finger joint ROMs increased after flexion and extension exercises. According to the results of the Kruskal–Wallis H-test on the differences in the joint ROMs before and after rehabilitation exercise, the ROM tended to increase the most after actuating the SMA for both the flexion and extension exercises (8.400 (19.200)° versus 7.670 (15.940)°, respectively), followed by thermal stimulation only. Specifically, after flexion exercise of the MCP joint, the change in the joint ROM during SMA actuation significantly exceeded that during exercise without SMA (p = .046). However, in the PIP joint, the ROM after thermal stimulation tended to be larger than that after SMA actuation (16.030 (5.010)° versus 14.880 (29.070)°); however, the difference was not significant (p = .561). After the extension exercise, the joint ROM yielded the most significant increase after SMA actuation in both the PIP and MCP joints (16.610 (22.620)° versus 2.300 (4.770)°). These results show that rehabilitation exercises using thermal stimulation and SMA actuation increased the joint ROM more than exercises without SMA actuation. We can infer that a hemiplegic patient’s finger joint ROM decreased owing to muscle stiffness during flexion and extension exercises without the actuation of a rehabilitation glove. Additionally, these results suggest that the combined effects of thermal stimulation and SMA contraction during actuation enhance the finger joint ROM more than just heat stimulation alone, particularly in extension motion.

Table 7.

Differences in the finger joint ROM before and after rehabilitation exercise.

Motion	Finger joints	ROM (°, mean (SD))
Motion	Finger joints	Without SMA	Thermal stimuli	With SMA	Kruskal–Wallis H	p
Flexion	PIP	−2.080 (13.320)	16.030 (5.010)	14.880 (29.070)	1.156	0.561
	MCP	−4.100 (2.680)^b	2.050 (3.090)^{a, b}	4.510 (6.490)^a	6.180	0.046
	Total	−3.340 (8.490)	7.290 (7.820)	8.400 (19.200)	5.735	0.057
Extension	PIP	−3.660 (2.680)	5.880 (8.700)	16.610 (22.620)	3.467	0.177
	MCP	−1.130 (3.380)	1.640 (0.930)	2.300 (4.770)	2.060	0.357
	Total	−2.080 (3.360)	3.230 (5.750)	7.670 (15.940)	5.715	0.057

Note: bold items = significant value; the alphabetical order is the same as the order of the mean difference based on the Duncan test.

Changes in grip force after rehabilitation exercise

We conducted the Mann–Whitney U-test to analyze the differences in the hand grip forces generated before and after rehabilitation exercises with and without glove actuation (Table 8). We found that the difference in the grip strength according to the actuation of the developed glove and thermal stimuli before the rehabilitation exercise was not significant (p = .268). The grip force increased after the rehabilitation exercise and thermal stimuli. Specifically, the grip strength after the flexion and extension rehabilitation exercises increased without SMA actuation, after thermal stimuli, and with SMA actuation by 21.900%, 45.900%, and 52.400%, respectively (p = .046, .040, 0.036). These results demonstrate that the grip force after the rehabilitation exercise exhibited a tendency to increase during the actuation of the developed glove. Using the rehabilitation glove developed in this study had positive effects on the grip forces during the flexion and extension rehabilitation exercises of the hand of a patient with hemiplegia. In addition, although the joint ROM tended to decrease after rehabilitation exercises in cases in which SMA actuation was not used, the grip strength increased after rehabilitation exercises (no SMA actuation). We can infer that this result is attributed to the fact that the grip strength is a measure of the strength of the entire finger.

Table 8.

Differences in the grip forces before and after rehabilitation exercises.

Grip force (N, mean (SD))
Without SMA	Before	After	Δ (%)	t	p
Without SMA	3.470 (0.410)	4.230 (0.090)	21.900	−1.993	0.046
Thermal stimuli	Before	After	Δ (%)	t	p
Thermal stimuli	3.050 (0.250)	4.450 (0.150)	45.900	−1.901	0.040
With SMA	Before	After	Δ (%)	t	p
With SMA	3.130 (0.240)	4.770 (0.400)	52.400	−1.882	0.036

Note: bold item = significant value.

Changes in skin temperature after glove operation

According to the Mann–Whitney U-test conducted to analyze the differences in the skin temperature and the temperature outside the glove before and after glove actuation (Table 9), the temperatures tended to increase after actuation. The skin temperature inside the glove increased by 4.667°C (15.490%) and 4.656°C (15.730%) on the dorsal hand side and the fingers, respectively, and the difference was significant (p = .005, .001). After actuating the glove, the temperature outside the glove increased by 13.100°C (66.970%) on the dorsal hand side (p = .002). The temperature outside the glove at the finger region increased by 3.768°C (14.810%); however, the difference was not significant (p = .115). Among the regions with knitted SMA, the skin temperature inside the glove increased to 34.800°C, and the temperature outside the glove increased to 42.300°C. These results indicate that the outside temperature of the glove (with the inserted SMA) continued to increase owing to the Joule heating generated during SMA operation; however, the temperature inside the glove (i.e., skin temperature) was maintained below ∼35°C owing to the insulation fabric. Therefore, we can infer that temperature increases from Joule heating resulting from SMA actuation are within a safe range when the glove developed in this study is used.

Table 9.

Differences in skin temperature before and after SMA actuation.

Temperature (°C, mean (SD))
Measurement region		Before	After	Δ (%)	z	p
Skin	Dorsum	30.133 (0.411)	34.800 (0.283)	15.490	−14.000	0.005
Skin	Finger	29.677 (0.450)	34.333 (0.660)	15.730	−28.000	0.001
Outside the glove	Dorsum	29.201 (1.042)	42.300 (1.219)	66.970	−20.263	0.002
Outside the glove	Finger	25.433 (1.247)	29.201 (1.042)	14.810	−2.694	0.115

Note: bold item = significant value.

Discussion

This study demonstrated the hand rehabilitation effects of a developed rehabilitation glove with integrated knitted SMA actuators for a patient with hemiplegia. We designed the SMA wire with a warp-knitted pattern and observed a higher contraction ratio compared with the weft-knitted design. We designed a full-fabric rehabilitation glove for wearability and ease of donning/doffing and inserted a warp-knitted SMA as a tendon-driven actuator. We demonstrated that the flexion and extension of the fingers of a hemiplegic patient can be assisted when performing ROM exercises using the developed rehabilitation glove. We proposed that the warp-knitted SMA can be used as a flexible and lightweight actuator to aid flexion and extension rehabilitation exercises.

We designed the SMA as a knitted structure and demonstrated that the warp-knitted structures yielded larger contraction ratios than the weft-knitted structures. In previous studies, wearable products were mainly developed using garters,^13,15 plains,^14,16,17 and ribs^14,16 in weft knits. However, we found that this warp-knitted structure yielded a higher contraction ratio than the weft-knitted structure. Based on the results of this study, we recommend the use of warp knits for wearable products that require a high shrinkage ratio. Warp-knit technology is generally implemented with a warp-knitting machine and generates structures that are extensively used in the production of medical and industrial textiles because (a) they can be transformed into complex forms, such as multi-axial warp knitting, and (b) the technology has high-productivity advantages.^21,34 In particular, we can infer that among the warp-knitted structures, the higher contraction ratio in the closed-lap structure is because the loop legs cross each other in this structure, unlike the open-lap structure.³⁵ Furthermore, the Raschel and Tricot warp-knitting machines can knit up to 1500 denier and 240 denier yarns, respectively.²⁹ Thus, the developed warp knitting SMA module has the potential to be applied as a mass production technology using a warp knitting machine with a larger gauge than the gauge commonly used in fabric production. Additionally, during the knitting process, machine knitting is fully possible because the thin metal wire can be bent with a weak force without having to be axially stretched. This suggests that our new warp-knitted actuator has the potential for mass production as well as customized products.

The findings of the functional evaluation indicate that glove operation in a hemiplegic patient has a significant effect in terms of increasing the finger joint ROM in hand extension motions. Unlike other joints, the MCP joint motion tended to decrease by −34.390% during flexion. The patient had spasticity in her fingers, which caused stiffness and unintentional resistance when applying bending force. The muscle tone increase after the exercise might also produce this effect.³⁵ This indicates that simple repetitive motions of flexion and extension of the hands of hemiplegic patients may reduce the joint ROMs of fingers. Patients with cerebral lesions temporarily experience increased muscle tone and spasticity, which is indicative of partial contraction of passive muscles.³⁶ In particular, there is a risk of increasing side effects, such as increased muscle tone, when patients attempt to perform upper extremity rehabilitation exercises independently.³⁷ Additionally, the joint ROM increases were largest after the actuation of the SMA during both flexion and extension exercises. These findings indicate that using the rehabilitation glove with SMA actuators, which induces thermal effects and generates actuation forces, can be a practical approach for improving joint ROM in individuals with hemiplegia. The joint ROM of fingers after extension exercises had the most significant increase after SMA operation in both PIP and MCP joints. However, in the PIP joint after the flexion exercise, the ROM increase tended to be significant when thermal stimulation was applied compared with the ROM induced after SMA actuation. We can infer that this is attributed to the fact that only thermal stimulation can cause muscle relaxation and increase the joint ROM, particularly during rehabilitation in flexion motion, which causes increased stiffness.^38,39

Experiments demonstrated that the grip strength increased during the actuation of the developed glove compared with heat stimulation and non-actuation. These results indicate that using the rehabilitation glove can positively affect grip strength during rehabilitation exercises. A study that proposed the use of hand splints to assist in hand extension exercises also proved that extension exercises increased grip strength.⁴⁰ These results support the findings of another study that suggested a close relationship between muscle tone and strength³⁸ and those of a study in which rehabilitation exercises were conducted using a soft robot hand that demonstrated a significant increase in voluntary grip strength in both mild and severe finger flexor stiffness cases.⁴¹ Because an increase in grip strength means functional recovery,⁴² it can be shown that performing rehabilitation exercises using the warp-knitted rehabilitation glove developed can lead to the functional recovery of the hemiplegic side’s hand. To date, soft robotic gloves using SMA wire have been developed based on tendon-driven,^7,8,43 and hard components still have to be added to increase the driving force. A rehabilitation assistive glove using an actuator made of a new flexible and soft textile is proposed for soft robotic applications based on the findings of this study. This supports the findings of a prior study that proposed design guidelines for rehabilitation gloves using a textile actuator with weft-knitted SMA fibers on the basis of anthropometric data from hemiplegic patients with specific spasticity grades.²⁸ In addition, some stroke patients can induce joint movement with less force in case they are flaccid without stiffness, which will be helpful in rehabilitation to restore hand function.

Furthermore, after glove actuation, the temperature outside the glove increased to 42.300°C, and the skin temperature of the patient’s hand inside the glove increased to 34.333°C and 34.800°C on the dorsal hand side and fingers, respectively. The patient’s skin temperature did not exceed 41.900°C owing to the insulating fabric inserted into the glove, and hence, prevented deep dermal burns during rehabilitation exercises.⁴⁴ Because hemiplegic patients have a lower skin temperature than healthy adults, many previous studies demonstrated that providing a thermal effect helps relieve spasticity.^45,46 Thus, the Joule heating generated during the operation of the developed glove increased the skin temperature of the patient, thereby justifying the functional capability of the rehabilitation glove in providing thermal stimulation capable of relieving muscle tone. Because hemiplegia patients have different symptoms, customized rehabilitation gloves must be designed for each patient even if they have the same degree of spasticity. Although this is a case study of a hemiplegic patient with specific spasticity, it is significant as it demonstrated for the first time the possibility of a novel actuator design based on the warp-knitted structure of an SMA wire.

However, this study has some limitations. While this study is meaningful in that we fabricated and prototyped a glove with integrated warp-knitted SMA and evaluated its feasibility with an actual stroke patient, additional validation with a larger sample size is desirable. Although we found a strong relationship between the muscle tone of the hemiplegic hand and the recovery of muscle strength and function through the use of the prototyped glove, the same results may not be obtained when there is severe spasticity.³⁸ That is, because the stiffnesses of the patient’s fingers vary, there may be cases where the effects of rehabilitation exercise are not noticeable in fingers and joints with severe stiffness characteristics; therefore, additional evaluations are required. This study presented the results of the evaluation based on short-term rehabilitation exercises. In future studies, it will be necessary to obtain the evaluation results based on long-term, repetitive rehabilitation exercises by using the developed gloves with hemiplegia patients having various degrees of spasticity. In addition, the characteristics of temperature changes versus the generated Joule heating time (when actuating the SMA actuator) should be analyzed because hemiplegic patients have a higher risk of thermal injury due to heating than healthy people owing to sensory loss on the hemiplegic side.⁴⁴ Nevertheless, this study proposed a new warp-knitted SMA actuator with a higher contraction ratio than the previously used weft-knitted SMA actuator. This actuator can be used in rehabilitation gloves to help hemiplegic patients during rehabilitation exercises.

Conclusions

This study demonstrated the positive effects of a developed rehabilitation glove with integrated warp-knitted SMA actuators on the ROMs of the finger joints, grip strength, and skin temperature in a hemiplegia patient. These findings suggest that the warp-knitted SMA, a soft actuator in a novel knit structure, is a potential material for wearable assistive products for rehabilitation. Specifically, the rehabilitation glove developed using a highly contractile, warp-knitted, SMA-based actuator demonstrated an increase in the joint ROMs of fingers and the grip strength of a hemiplegic patient after flexion and extension exercises. The safety of the SMA actuator was verified when used with the rehabilitation glove. Additionally, the heat generated provided positive thermal stimulation effects as in previous research. These findings show that the developed rehabilitation glove can potentially help the functional recovery of hemiplegic patients. Therefore, the warp-knitted SMA actuators, developed for the first time in this study, contributed to the increase in the applicability of lightweight, flexible, and highly contractile actuators for rehabilitation assistance devices. Additional research is required to optimize the design and function of the device for various patient populations with different grades of paralysis.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Seoul National University Research Grant in 2021; the National Research Foundation of Korea (NRF) grant funded by the Korean Government (MSIT) [RS-2023-00208052]; the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education [2020R1A6A3A01099046].

ORCID iDs

Soo-Min Lee

Juyeon Park

References

Wayman

. Shape memory alloys. MRS Bull 1993; 18: 49–56.

Jani

Leary

Subic

, et al. A review of shape memory alloy research, applications and opportunities. Mater Des 2014; 56: 1078–1113.

Farber

Zhu

Popovich

, et al. A review of NiTi shape memory alloy as a smart material produced by additive manufacturing. Mater Today Proc 2020; 30: 761–767.

Srivastava

Alsamhi

Murray

, et al. Shape memory alloy-based wearables: a review, and conceptual frameworks on HCI and HRI in Industry 4.0. Sensors 2022; 22: 6802.

Copaci

Cano

Moreno

, et al. New design of a soft robotics wearable elbow exoskeleton based on shape memory alloy wire actuators. Appl Bionics Biomech 2017; 2017: 1605101.

Hope

McDaid

. Development of wearable wrist and forearm exoskeleton with shape memory alloy actuators. J Intell Rob Syst 2017; 86: 397–417.

Hadi

Alipour

Kazeminasab

, et al. ASR glove: a wearable glove for hand assistance and rehabilitation using shape memory alloys. J Intell Mater Syst Struct 2018; 29: 1575–1585.

Chernyshov

Tag

Caremel

, et al. Shape memory alloy wire actuators for soft, wearable haptic devices. In: Proceedings of the 2018 ACM international symposium on wearable computers, 2018, pp. 112–119.

Kim

Heo

Rodrigue

, et al. Shape Memory Alloy (SMA) Actuators: the role of material, form, and scaling effects. Adv Mater 2023; 35: 2208517.

10.

Park

. Suit-type wearable robot powered by shape-memory-alloy-based fabric muscle. Sci Rep 2019; 9: 9157.

11.

Kim

Lee

, et al. Shape memory alloy actuator-embedded smart clothes for ankle assistance. Smart Mater Struct 2020; 29: 055003.

12.

Han

Kim

Ahn

. Shape memory textile composites with multi-mode actuations for soft morphing skins. Compos B Eng 2020; 198: 108170.

13.

Eschen

Granberry

Abel

. Guidelines on the design, characterization, and operation of shape memory alloy knitted actuators. Smart Mater Struct 2020; 29: 035036.

14.

Han

Ahn

. Blooming knit flowers: loop-linked soft morphing structures for soft robotics. Adv Mater 2017; 29: 1–6.

15.

Granberry

Eschen

Holschuh

, et al. Functionally graded knitted actuators with NiTi-based shape memory alloys for topographically self-fitting wearables. Adv Mater Technol 2019; 4: 1–11.

16.

Lee

Jung

Park

, et al. Development of a 4D hand gripping aid using a knitted shape memory alloy and evaluation of finger-bending angles in elderly women. Fashion Text 2022; 9: 1–16.

17.

Jung

Lee

Ahn

, et al. Development and assessment of a knitted shape memory alloy-based multifunctional elbow brace. J Ind Text 2022; 51(2_suppl): 1989S–2009S.

18.

Ray

(ed) Fundamentals and advances in knitting technology. New York, NY: CRC Press, 2012.

19.

Leong

Ramakrishna

Huang

, et al. The potential of knitting for engineering composites—a review. Composer A 2000; 31: 197–220.

20.

David

. Knitting technology: a comprehensive handbook and practical guide. Boca Raton, FL: CRC Press, 2001.

21.

Sankaran

Rittner

Hahn

, et al. Development of multiaxial warp knitting technology for production of three-dimensional near net shape shell preforms. Textil Res J 2017; 87: 1226–1241.

22.

LaBat

Ryan

. Human body. A wearable product designer’s guide. 1st ed. Boca Raton (BR): CRC Press, 2019.

23.

Hunter

Crome

. Hand function and stroke. Rev Clin Gerontol 2002; 12: 68–81.

24.

Carmeli

Peleg

Bartur

, et al. HandTutor^TM enhanced hand rehabilitation after stroke—a pilot study. Physiother Res Int 2011; 16: 191–200.

25.

Sale

Mazzoleni

Lombardi

, et al. Recovery of hand function with robot-assisted therapy in acute stroke patients: a randomized-controlled trial. Int J Rehabil Res 2014; 37: 236–242.

26.

Yurkewich

Hebert

Wang

, et al. Hand extension robot orthosis (HERO) glove: development and testing with stroke survivors with severe hand impairment. IEEE Trans Neural Syst Rehabil Eng 2019; 27: 916–926.

27.

Khalid

Alnajjar

Gochoo

, et al. Robotic assistive and rehabilitation devices leading to motor recovery in upper limb: a systematic review. Disabil Rehabil Assist Technol 2021; 18: 1–15.

28.

Lee

Jung

Park

. Mapping actuating performance of knitted shape memory alloys and proposing design guidelines for assistive wearable applications. J Ind Text 2023; 53: 1–25.

29.

Gajjar

. Advances in warp knitted fabric production. In: Advances in knitting technology. Woodhead Publishing, 2011, pp. 110–135.

30.

Renkens

Kyosev

. Geometry modelling of warp knitted fabrics with 3D form. Textil Res J 2011; 81: 437–443.

31.

Eschen

Abel

. Performance and prediction of large deformation contractile shape memory alloy knitted actuators. Smart Mater Struct 2019; 28: 025014.

32.

Lee

Kim

Park

. An ergonomic glove pattern drafting method for hand assistive devices. In: Proceedings of the international textile and apparel association annual conference 2022, Denver, CO, USA, Iowa State University Digital Press, p. 77.

33.

Cha

Kim

Lee

, et al. Prevalence rate of spasticity at 3 months after stroke in Korea: the Korean stroke cohort for functioning and rehabilitation (KOSCO) study. Brain Neurorehabil 2016; 9: e6.

34.

Feng

. Development of the warp knitted spacer fabrics for cushion applications. J Ind Text 2008; 37: 213–223.

35.

Kurbak

. Models for basic warp knitted fabrics Part II: single guide bar fabrics (closed-lap and open-lap). Textil Res J 2019; 89: 1886–1916.

36.

O'Sullivan

Schmitz

Fulk

. Physical rehabilitation. 6th ed. Philadelphia, PA: F. A. Davis, 2019.

37.

Ranzani

Chiriatti

Schwarz

, et al. An online method to monitor hand muscle tone during robot-assisted rehabilitation. Front Robot AI 2023; 10: 1093124.

38.

Brown

Van Rensburg

Lakie

GWM

, et al. A neurological study of hand function of hemiplegic children. Dev Med Child Neurol 1987; 29: 287–304.

39.

Gracies

. Physical modalities other than stretch in spastic hypertonia. Phys Med Rehabil Clin 2001; 12: 769–792.

40.

Chang

Lai

. New design of home-based dynamic hand splint for hemiplegic hands: a preliminary study. J Phys Ther Sci 2015; 27: 829–831.

41.

Shi

Heung

Tang

, et al. Effects of a soft robotic hand for hand rehabilitation in chronic stroke survivors. J Stroke Cerebrovasc Dis 2021; 30: 105812.

42.

Park

. Grip strength in post-stroke hemiplegia. J Phys Ther Sci 2016; 28: 677–679.

43.

Wang

Zheng

Pang

, et al. Design and experiment of a hand movement device driven by shape memory alloy wires. J Robot 2021; 2021: 1–13.

44.

Suzuki

Hirayama

Aihara

, et al. Experimental studies of moderate temperature burns. Burns 1991; 17: 443–451.

45.

Alfieri

Massaro

Filippo

, et al. Evaluation of body temperature in individuals with stroke. NeuroRehabilitation 2017; 40: 119–128.

46.

Toki

Yamai

Fukai

. Skin temperature changes during a footbath in patients who had had a stroke with consequent sensory impairment. Jpn J Nurs Sci 2015; 12: 276–286.

Feasibility of a warp-knitted shape memory alloy rehabilitation glove

Abstract

Keywords

Introduction

Methods

Materials

Design and fabrication of rehabilitation assistive glove

Knitted Shape memory alloy

Fabrication of rehabilitation assistive glove

Evaluation

Participants

Experimental process

Sensor attachment

Functional evaluation of joint range of motion

Rehabilitation exercise assessment

Data analysis

Results

Differences in contraction ratios according to knit structure

Rehabilitation assistive glove

Changes in joint ROM after actuation

Changes in joint ROM after rehabilitation exercise

Changes in grip force after rehabilitation exercise

Changes in skin temperature after glove operation

Discussion

Conclusions

Footnotes

Declaration of conflicting interests

Funding

ORCID iDs

References