Sage Journals: Discover world-class research

Abstract

Background

Several studies have shown that robotic devices can effectively improve motor function in stroke patients through limb activation. However, the effects of robot-assisted therapy on perceptual deficits after stroke is unclear.

Objective

This review aimed to evaluate the effectiveness of robotic limb activation in patients with unilateral spatial neglect (USN) after stroke.

Methods

In this systematic review, a literature search was performed using MEDLINE, EMBASE, CENTRAL, CINAHL, and LILACS databases without language restrictions. Randomized controlled trials (RCTs) and quasi-RCTs of robot-assisted therapy for USN after stroke were selected. Two reviewers independently assessed the risk of bias and certainty of the evidence of the included studies.

Results

A total of 630 studies were identified, including five studies for qualitative synthesis and four meta-analyses. The results of RCTs comparing robotic limb activation with a control group suggested an improvement in the degree of USN measured by the line bisection test (standardized mean difference [SMD], −0.64; 95% confidence interval [CI], −1.13 to −0.15; P = .01). There were no differences between the groups in the motor-free visual perception test 3rd edition (SMD, 0.27; 95% CI, −0.25–0.79; P = .31), star cancellation test (SMD, 0.26; 95% CI, −0.42−0.94; P = .54), Albert’s test (SMD, −0.67; 95% CI, −2.01−0.66; P = .32), and Catherine Bergego Scale (SMD, −0.81; 95% CI, −2.07−0.45; P = .21).

Conclusion

The study demonstrated that limb activation through robotic therapy can improve midline perception. However, there was no impact on tasks assessing visual scanning, functionality, or activities of daily living.

Keywords

stroke unilateral spatial neglect robot-assisted therapy limb activation

Introduction

The nonrecognition of space, usually on one side of the body, or any stimulus applied to it can be called hemineglect syndrome or unilateral spatial neglect (USN).^1,2 USN is characterized by the inability to report or respond to people or objects presenting contralaterally to the injured side of the brain, which cannot be explained by primary motor or sensory deficits.³ This syndrome occurs in approximately 50% of individuals after the stroke of the right hemisphere, and up to 75% may persist with some symptoms in the chronic phase.⁴ Moreover, USN has important clinical implications, as it leads to prolonged periods of hospitalization, impaired functional recovery, and worsening quality of life.^5,6

A tendency to explore and accentuate orientation to the ipsilesional side of the patients and nonrecognition of items on the contralesional side is provoked by USN.^2,7-14 Main treatments for patients with USN after stroke are the “top-down” approaches,¹⁵ which aim to implant voluntary attention, and “bottom-up” techniques, which aim to automatically capture attention to stimuli or events in the environment.^16-22 There is evidence of the effectiveness of several “bottom-up” approaches to improve USN performance, among which limb activation has been highlighted as a technique to reduce visuospatial deficits in patients with USN after stroke.^23-27

Given the high incidence of upper limb paresis associated with USN after stroke, the positive effect of passive or active movement of the left upper or lower limb can be an excellent strategy in the field of USN rehabilitation.²⁷ Furthermore, Frassinetti et al. observed a reduction in USN during passive and active movements of the left hand.²⁸ Some authors have reported that somatosensory activation on the contralateral side of limb activation facilitates the neural circuits underlying the representation of space, which could explain the clinical efficacy of activation of the left hand in improving USN.^29,30

With the development of robotic devices, there is a growing interest in using them in post-stroke rehabilitation, as they can provide a consistent, intensive, and accurate rehabilitation process for the paretic limb.³¹ In fact, several robotic devices have been developed in the last decade and have been shown to be effective in improving motor function in stroke patients using passive, assisted, and active training modes.³² However, unlike motor function in the upper limb, few studies have reported the effects of robot-assisted therapy in patients with USN after stroke. In the last decade, there have been some clinical trials using limb activation through robotic therapy for patients with USN; however, no systematic review has assessed the effectiveness of this training in guiding clinical practice. Therefore, this review aimed to identify the effects of robot-assisted training on patients with USN after stroke.

Materials and Methods

This review was based on the methods described in the Cochrane Handbook for Intervention Reviews³³ and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).³⁴ Randomized clinical trials (RCTs) and quasi-RCTs conducted up to February 2022 were included in this evaluation. This review was registered in the International Prospective Register of Systematic Reviews (CRD42022303587).

Eligibility Criteria

Study Design

RCTs and quasi-RCTs.

Participants

Adults with an objective diagnosis of USN after stroke (ischemic or hemorrhagic), over 18 years of age, regardless of sex, duration of illness, or severity of initial impairment.

Interventions

Therapies that use robot-assisted therapy including an exoskeleton (based on torque actuators that control each joint of the affected limb) or end-effector (guides only the distal part of the affected limb).

Comparators

The interventions were compared with any conventional rehabilitation for USN after stroke (e.g., pharmacological or non-pharmacological therapies, such as “top-down” techniques, which include noninvasive brain stimulation and mental practice, and “bottom-up” techniques, which include prismatic adaptation, visual scanning training, optokinetic stimulation, virtual reality, combined neck vibration, voluntary trunk rotation, vestibular rehabilitation, and paretic limb activation).^15-24

Outcomes

(1) USN degree measured by any paper and pencil test,³⁵ such as the line cancellation task, line bisection test (LBT), and star cancellation test, and any validated specific instrument, such as the Catherine Bergego Scale³⁶ or the behavioral inattention test (BIT)³⁷ after treatment and in the long-term; (2) functional disability measured by the modified Rankin scale³⁸ or the Fugl–Meyer scale³⁹ after treatment and in the long-term; (3) autonomy in activities of daily living measured by any scale validated in this population, such as the Barthel index³⁸ after treatment and in the long-term; (4) balance measured using the Berg Balance scale⁴⁰ or mini-best test⁴¹ after treatment and in the long-term; (5) depression or anxiety measured using the Beck Depression Inventory,⁴² hospital anxiety and depression scale,⁴³ and Hamilton Depression Rating Scale⁴⁴ after treatment and in the long-term; and (6) quality of life, measured by the stroke impact scale,⁴⁵ EUROQOL,⁴⁶ and SF-36⁴⁷ after treatment and in the long-term.

Search Strategy and Databases

The Cochrane Central Register of Controlled Trials (CENTRAL) (1966 to February 2022), EMBASE (OvidSP) (1980 to February 2022), Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, 2022), CINAHL (1961 to February 2022), and Latin American and Caribbean Center on Health Sciences Information (LILACS) (1982 to February 2022) were used without language restrictions. The most recent search date was February 26, 2022. All searches were performed by a librarian trained in systematic reviews. The reference lists of relevant studies and conference proceedings were also searched, and the authors of the included studies were contacted. The review search strategy for each database is shown in Supplemental File 1.

Selection of Studies

Two pairs of reviewers independently examined all titles and abstracts identified in the literature search, obtained full articles from all potentially eligible studies, and assessed them for eligibility. Disagreements were resolved by discussion or, if necessary, with the help of a third reviewer. Studies that were presented at conferences were also included.

Data Extraction

The reviewers performed calibration exercises and worked in pairs to independently extract data from the included studies. The following data were extracted using a pre-tested data extraction form: study design, participants, interventions, comparators, assessed outcomes, and relevant statistical data.

Risk of Bias Assessment

Reviewers working in pairs independently assessed the risk of bias in the included studies using the modified version of the Cochrane Collaboration instrument (http://distillercer.com/resources/).^48,49 For incomplete outcome data in individual studies, we set a low risk of bias for loss of follow-up at <10% and a difference of less than 5% in missing data between the intervention and control groups. When information on the risk of bias or other aspects of the methods or results was not available, we attempted to contact the authors for additional information.

The Physiotherapy Evidence Database (PEDRO) scale was also used to assess the risk of bias. It is based on the Delphi list and consists of verifying eligibility criteria, randomization and allocation concealment, the similarity between groups and subjects, verifying whether therapists and evaluators were blinded, 85% of subjects remaining to initiate the study, intent-to-treat outcome analysis, intergroup comparison, and measurement of precision and variability in at least one outcome. The total score is 10 points; the higher the score, the lower the risk of bias.⁵⁰

Certainty of Evidence

The GRADE (Recommendations Assessment, Development, and Evaluation) methodology was used to classify the certainty of the evidence for each outcome as high, moderate, low, or very low.⁵¹ In the GRADE approach, RCTs start with high certainty and non-randomized trials with moderate certainty. We planned to assess publication bias by visually inspecting funnel plots for each outcome in which we identified 10 or more eligible studies; however, this was not possible because there was an insufficient number of studies to allow this assessment.

Data Synthesis and Statistical Analysis

A meta-analysis was performed by using similar studies with homogeneous outcomes. Standardized mean differences in combined inverse variance (SMD) and 95% confidence intervals (CIs) between groups were calculated using random-effects models. The variability in the results between the studies was analyzed using the I² statistic and P-value obtained from the Cochran chi-square test. To determine the mean difference between the groups (experimental—physical exercise and control – any comparison), the Z statistic was used, with P < .05 considered statistically significant. Review Manager (RevMan) (version 5.3; Nordic Cochrane Center, Cochrane)⁵² was used for analysis.

Subgroup and Sensitivity Analysis

Possible subgroup analyses have been planned according to the following characteristics:

(i) Participants (stroke type: ischemic stroke vs. intracranial hemorrhage)

(ii) Interventions (type of robotic assistance: exoskeleton versus end-effector; frequency: high-frequency versus low-frequency training)

(iii) Comparator (type of control intervention: pharmacological therapy versus non-pharmacological therapy; associated use of noninvasive brain stimulation versus sham group).

Subgroup analyses were performed only when five or more studies were available, with at least two in each subgroup.

Results

A total of 630 studies were identified, and after reading the titles and abstracts, 23 studies were selected for eligibility analysis, 18 of which were excluded because they were prospective studies (n = 2), case series (n = 3), proof of concept (n = 1), did not use any type of robot (n = 4), did not include USN patients (n = 5), used robot without limb activation (n = 1), and were clinical trial protocols (n = 2). In total, two RCTs,^53,54 two pilot RCTs,^55,56 and one quasi-RCT⁵⁷ were included in the qualitative analysis. A flowchart of the review steps is shown in Figures 1 and 2.

Figure 1.

Flowchart of the review steps.

Figure 2.

Meta-analysis of the unilateral spatial neglect degree measured by the motor-free visual perception test third edition.

The relevant characteristics of each study included in the qualitative analysis are presented in Tables 1 and 2. There were 73 and 69 participants in the intervention and control groups, respectively. The mean age of the individuals in the experimental and control groups was 60.65 (±8.88) years and 63.41 (±9.13) years, respectively. Three studies included patients in the subacute phase,^53,54,56 one in the chronic phase,⁵⁵ and one in the acute phase⁵⁷ of stroke. In all studies, the individuals had ischemic and/or hemorrhagic strokes. The type of robotic therapy was the end-effector type in four studies^53-55,57 and the exoskeleton type in one study.⁵⁶ In all studies, the outcome assessment was performed after the intervention protocol. No long-term follow-up was performed for any of the included studies.

Table 1.

Characteristics of Included Studies.

Author/year	Study design	N/Age	Stroke phase	Stroke type	Inclusion criteria	Exclusion criteria
Choi et al. 2016⁵³	RCT	46 individuals allocated in two groups (GI: n = 23; 60.75 ± 12.54 years; CG: n = 23; 63.17 ± 11.12 years)	Subacute	Ischemic and hemorrhagic	Patients with stroke in the right hemisphere and USN. USN was defined as a mean deviation of ≥5 mm to the right side in LBT.	Prior stroke, severe cognitive impairment, left upper extremity muscle strength <2 on manual muscle testing, any visual field defect, and severe impairment of sitting balance.
Kim et al. 2018⁵⁴	RCT	39 participants allocated in three groups (rTMS: n = 13; 70.3 ± 9.6 years; Robotics: n = 13; 66.6 ± 12.2 years; Combined: n = 13; 62.5 ± 16.5 years)	Subacute	Ischemic and hemorrhagic	Patients with stroke in the right hemisphere and USN. USN was defined as a mean deviation of ≥5 mm to the right side in LBT.	Prior stroke, severe cognitive impairment, left upper extremity muscle strength <2 on manual muscle testing, any visual field defect, and severe impairment of sitting balance.
Park 2021⁵⁵	Pilot RCT	24 individuals allocated in two groups (GI: n = 12; 69.08 ± 4.71 years; CG: n = 12; 71.58 ± 3.17 years)	Chronic	Ischemic and hemorrhagic	Age >65 years; right hemisphere stroke; first stroke; preserved cognitive function; time since stroke onset ≥6 months, and USN diagnosed by LBT and MVPT-3.	Any additional treatment for USN, left upper limb sensory deficit, visual deficit, modified Ashworth ≥2, left upper limb muscle strength <2 on manual muscle testing, orthopedic conditions, and apraxia.
Chen et al. 2021⁵⁶	Pilot RCT	20 individuals assigned in two groups (GI: n = 10; 46.20 ± 7.02 years; CG: n = 10; 48.60 ± 9.95 years)	Subacute	Ischemic and hemorrhagic	Age 18−80 years, right hemisphere stroke (stroke onset 2 weeks to 6 months), UE-FMA score 8–47, and USN defined by BIT-C.	Prior stroke, visual impairment, fixed contracture, shoulder subluxation, aphasia, or cognitive impairment.
Chege 2021⁵⁷	Quasi-RCT (conference)	23 individuals were assigned to two groups (GI: n = 15; CG: n = 11)	Acute	Not reported	Stroke patients exhibit signs of USN, as determined by AT, SCT, and CBS.	Not reported

RCT, randomized controlled trial; IG, robotics intervention group; CG, control group; LBT, line bisection test; AT, Albert’s test; SCT, star cancellation test; CBS, Catherine Bergego Scale; UE-FMA, upper extremity Fugl-Meyer; BIT-C, Behavioral Inattention Test Conventional; MVPT-3, motor-free visual perception test, 3rd edition.

Table 2.

Outcomes and Treatment of Included Studies.

Author/year	Intervention	Control	Robotic type	Outcomes	Follow-up
Choi et al. 2016⁵³	Use of a planar robotic arm to perform shoulder abduction-adduction and elbow flexion-extension. The protocol was implemented through games that induce passive, active-assisted, and active exercises. Protocol: 30 min per day, 5 days a week for 3 weeks.	Conventional treatment (visual scanning and range-of-motion exercises) Protocol: 30 min per day, 5 days a week for 3 weeks.	End-effector (Neuro-X®)	MVPT-3, LBT, SCT, AT, CBS, MMSE, K-MBI	After intervention
Kim et al. 2018⁵⁴	Use of a planar robotic arm to perform shoulder abduction-adduction and elbow flexion-extension. The protocol was implemented through games that induce passive, active-assisted, and active exercises. The protocol was added to repetitive magnetic stimulation (rTMS) over the left posterior parietal cortex (900 stimuli at an intensity of 95% of motor thresholds and a frequency of 0.9 Hz). Protocol: 20 min per day, 5 days a week for 2 weeks.	In this study, there were two control groups: a) isolated rTMS stimulation and b) isolated robotic stimulation Protocol: 20 min per day, 5 days a week for 2 weeks.	End-effector (Neuro-X®)	MVPT-3, LBT, SCT, AT, CBS, MMSE, K-MBI	After intervention
Park 2021⁵⁵	Robot training with 5 degrees of freedom and interface between the hand and the machine made through elastic bands and the wrist stabilized by a Velcro strap. The exercises were: (1) catch and release training for 15 min; and (2) counting practice (counting a sequence of numbers from one to five) for 15 min. The participant’s hand movement was assisted by the robot and adjusted to the individual’s level of function through the assistive therapy mode. During training, participants received visual feedback on their hand movements through a video animation presented on a monitor. Protocol: 30 min each session, 20 sessions.	Visual scanning exercises using prism adaptation, vibration in the left neck extensor, and the middle part of the left forearm. Additionally, CG participants learned a compensatory approach to improving USN symptoms. Protocol: 30 min each session, 20 sessions.	End-effector (Amadeo Robotic Device)	LBT, AT, CBS	After intervention
Chen et al. 2021⁵⁶	The exoskeleton allowed movement in 7 degrees of freedom involving the shoulder (flexion/extension, abduction/adduction and internal/external rotation), elbow (flexion/extension, forearm supination/pronation), and wrist (flexion/extension and ulnar deviation/radial deviation). Each session consisted of a passive mode (15 min) and an assistance mode (30 min). During the passive mode, the exoskeleton manipulated the upper extremity with a three-dimensional trajectory. In the assisted mode, patients practiced ADL games and training programs that emphasize the left side. Protocol: 45 min daily, 5 days/week for 4 weeks.	Cognitive and occupational rehabilitation for USN, consisting of visual scanning therapy, range of motion, and perceptual training integrated with task-specific activities. Protocol: 45 min daily, 5 days/week for 4 weeks.	Exoskeleton (Armule®)	BIT-C, CBS, FMA-UE, MBI, WHODAS 2.0	After intervention
Chege, 2021⁵⁷	The treatment group received conventional therapy and an additional 30 min of robotic training. Robotic treatment consisted of computer games with targets acquired through the interactive movement of the robot with visual feedback. Protocol: 10 days in the first two weeks of rehabilitation.	Conventional therapy involves training in activities of daily living, muscle activation, and midline guidance. Protocol: 10 days in the first two weeks of rehabilitation.	End-effector (InMotion 2)	CBS, FIM, SCT, e AT	After intervention

MVPT-3, motor-free visual perception test 3rd edition; LBT, line bisection test; SCT, star cancellation test; AT, Albert’s test; CBS, Catherine Bergego Scale; MMSE, Mini-Mental State Examination; K-MBI, Korean version of the modified Barthel index; FIM, Functional Independence Measure.

The risk of bias in the included studies is presented in Table 3. The PEDRO score was eight in two studies,^55,56 five in one study,⁵⁴ and four in one study.⁵³ The detailed guidance of GRADE in the evidence profile is presented in Table 4.

Table 3.

Risk of Bias in the Studies Included in the Qualitative Analysis.

Study, year	Random sequence generation	Allocation concealment	Blind of participants	Blind of outcome assessment	Incomplete outcome data	Selective reporting	Other bias	PEDRO score
Choi, 2016	Uncertain	Uncertain	High	Uncertain	High	Low	High	4
Kin, 2018	Low	High	High	High	High	Low	High	5
Chen, 2021	Low	Low	High	Low	Low	Low	Uncertain	8
Park, 2021	Low	Low	High	Low	Low	Low	Uncertain	8

Table 4.

GRADE Scale to Rate the Certainty of the Evidence for Motor-Free Visual Perception Test Third Edition Outcome.

Motor-free visual perception test 3rd edition
No. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	QUALITY
CHOI, 2016
1	RCT	Serious	Serious	Not serious	Serious	⊕○○○ VERY LOW
KIM, 2018
1	RCT	Serious	Not serious	Not serious	Not serious	⊕⊕⊕○ MODERATE
Line bisection test
CHOI, 2016
1	RCT	Not serious	Serious	Not serious	Serious	⊕⊕○○ LOW
KIM, 2018
1	RCT	Not serious	Not serious	Not serious	Not serious	⊕⊕○○ LOW
PARK, 2021
1	RCT	Not serious	Not serious	Not serious	Not serious	⊕⊕⊕⊕ HIGH
Star cancellation test
CHOI, 2016
1	RCT	Not serious	Serious	Not serious	Serious	⊕⊕○○ LOW
KIM, 2018
1	RCT	Not serious	Serious	Not serious	Serious	⊕⊕○○ LOW
CHOI, 2016
1	RCT	Serious	Serious	Not serious	Serious	⊕○○○ VERY LOW
Albert’s Test
KIM, 2018
1	RCT	Serious	Serious	Not serious	Serious	⊕○○○ VERY LOW
PARK, 2021
1	RCT	Not serious	Not serious	Not serious	Not serious	⊕⊕⊕⊕ HIGH
Catherine Bergego Scale
CHOI, 2016
1	RCT	Serious	Serious	Not serious	Serious	⊕○○○ VERY LOW
KIM, 2018
1	RCT	Serious	Serious	Not serious	Serious	⊕○○○ VERY LOW
PARK, 2021
1	RCT	Not serious	Not serious	Not serious	Not serious	⊕⊕⊕⊕ HIGH
CHEN, 2021
1	RCT	Not serious	Serious	Not serious	Serious	⊕⊕○○ LOW

Four studies^53-56 were included in the meta-analysis because of the presence of homogeneous outcomes. The study by Chege⁵⁷ was not reported in the meta-analysis because of a lack of data.

Motor-Free Visual Perception Test Third Edition (MVPT-3)

The results of the two RCTs (Choi et al.⁵³ and Kim et al.⁵⁴), comparing robotic limb activation with a control group, suggested that there was no improvement in the degree of USN measured by MVPT-3 (SMD, 0.27; 95% CI, −0.25−0.79; P = .31; I² = 0%) (Figure 3). The quality of this outcome was rated very low by one study⁵³ because of the high risk of bias in sample blinding and incomplete data, while the other study⁵⁴ was rated moderate quality due to the low risk of bias in the randomization of the allocation sequence, consistency, and accuracy in the data.

Figure 3.

Meta-analysis of the unilateral spatial neglect degree measured by the line bisection test.

LBT

The results of the three RCTs (Choi et al⁵³ Kim et al,⁵⁴ and Park⁵⁵), comparing robotic limb activation with a control group, suggested an improvement in the degree of USN measured by the LBT (SMD, −0.64; 95% CI, −1.13 to −0.15; P = .01; I² = 92%) (Figure 3). The quality of this outcome was rated as low in one study⁵³ because of the high risk of bias in sample blinding and incomplete data. However, one study⁵⁴ was rated moderate quality because of the low risk of bias in sequence randomization, allocation, consistency, precision, and direct data, while another study⁵⁵ was rated high quality evidence because of the low risk of bias in blinding and randomization of the allocation sequence, blinding of evaluators, and incomplete data.

Star cancellation test (SCT)

The results of the two RCTs (Choi et al⁵³ and Kim et al⁵⁴), comparing robotic limb activation with a control group, suggested that there was no improvement in the degree of USN measured by SCT (SMD, 0.26; 95% CI, −0.42−0.94; P = .54; I² = 38%) (Figure 4). The quality of this outcome was rated as low in two studies^53,54 because of the high risk of bias in sample and rater blinding, incomplete data, and allocation concealment.

Figure 4.

Meta-analysis of the unilateral spatial neglect degree measured by the star cancellation test.

Albert’s Test (AT)

The results of the three RCTs (Choi et al,⁵³ Kim et al,⁵⁴ and Park⁵⁵), comparing robotic limb activation with a control group, suggested that there was no improvement in the USN degree measured by AT (SMD, −0.67; 95% CI, −2.01−0.66; P = .32; I² = 87%) (Figure 5). The quality of evidence for this outcome was rated very low in two studies^53,54 because of imprecision, inconsistency, and high risk of bias for patient blinding and incomplete data. However, one study⁵⁵ was rated high quality evidence because of a low risk of bias for allocation concealment and allocation sequence, incomplete data, and blinding of the rater.

Figure 5.

Meta-analysis of the unilateral spatial neglect degree measured by Albert’s test.

Catherine Bergego Scale (CBS)

The results of the four RCTs (Choi et al,⁵³ Kim et al,⁵⁴ Park,⁵⁵ and Chen et al.⁵⁶), comparing robotic limb activation with a control group, suggested that there was no improvement in the functional impact of USN as measured by CBS (SMD, −0.81; 95% CI, −2.07−0.45; P = .21; I² = 87%) (Figure 6). The quality of evidence for this outcome was rated high in one study⁵⁵; very low in two studies^53,54 because of the high risk of bias in sequence blinding and allocation, sample and rater blinding, and incomplete data; and low in one study⁵⁶ because of the high risk of bias in sequence blinding and allocation and sample and rater blinding.

Figure 6.

Meta-analysis of the functional impairment degree measured by the Catherine Bergego Scale.

Discussion

The main objective of this review is to investigate the effects of limb activation using a robotic device in individuals with USN following stroke. The results of this review suggest that robot-assisted training of the left hand may improve LBT-midline perception, indicating that robot-assisted limb activation training may be useful in improving USN symptoms in stroke patients. However, no impact was observed in the other visual-perceptual neglect tests (star cancellation test, motor-free visual perception test, and Albert’s test), showing that robotic therapy did not affect tasks assessing visual scanning. In addition, we did not observe positive effects on functionality and activities of daily living.

Previous studies have reported that passive and active movements in the left limb can reduce USN in stroke patients.^23,30,55 The positive effects of robot-assisted training can be attributed to increased activation of the right hemisphere motor circuits and recruitment of the associated attention mechanism during the task, improving attention to the left side of space.⁷ Takahashi et al observed that robot-assisted manual training increased the size of the representational map of the primary sensorimotor cortex.²⁶ In addition, continuous visual feedback during robotic training can activate participants’ extra-personal space when viewing the animation of their hand movements on the monitor.

USN is an attentional neuropsychological disorder characterized by the inability to explore, orient, or respond to the contralesional side of space. Another possible explanation for these findings is that robotic training focused on personal and peripersonal space, increasing performance on pencil and paper tests.⁵⁸ Furthermore, in all robot-assisted therapy intervention protocols, the therapist provided continuous feedback to participants to increase attention and explore the space contralateral to the brain injury, which can serve as a visual scanning training for visual search and exploring objects.²² Another theory that can be added to explain the effects of robotic therapy in USN is that increased activation of the right hemisphere by activation of the left upper limb increases inhibition of the ipsilesional hemisphere and reduces hemispheric inhibition via the corpus callosum to the affected hemisphere.⁵⁹ In addition, passive movements for multiple joints with the exoskeleton may play a role in limb activation therapy, arousing contralateral attention.⁶⁰

In this review, we observed that only one study (Park et al)⁵⁵ had a significant effect on line bisection improvement. In contrast to Choi et al⁵³ and Kim et al,⁵⁴ Park et al⁵⁵ investigated the transfer effects of robot-assisted limb activation on a participant’s daily function. The study by Park et al. used a different robot training system than other studies, and the exercises were focused on grasping and release training with visual feedback of their hand movements via video animation.⁵⁵ Choi et al and Kim et al used a robot treatment program through games that induced passive and active assistive exercises of the upper limb in general.^53,54 In addition, the study by Park et al⁵⁵ had a higher frequency and number of total sessions than other studies.^53,54 The higher frequency of stimulation and type of exercise may have been favorable for improving the outcome associated with midline perception.

The lack of a significant effect of robotic therapy in the other visual-perceptual neglect tests (star cancellation test, motor-free visual perception test, Albert’s test) showed that robotic therapy did not affect tasks assessing visual scanning. Line bisection and cancellation tasks have different components of neglect syndrome. Line bisection necessitates the correct perception of the size of a single stimulus (“midline perception”), and cancellation tasks depend on the ocular exploration of a display containing many visual stimuli that have to be scanned systematically from left to right.⁶¹

The line bisection test is a screening test that is valid and reliable³⁵; however, the minimal clinically important difference (MCID) of this test as a tool to measure responsiveness to intervention is not known. As only improvement in midline perception by LBT was observed in this review, we should be cautious in stating that robotic therapy can improve the neglect syndrome, and the use of the gold standard test with known responsiveness should be encouraged in other clinical trials.

In all studies, there was another potential stimulus in addition to robotic assistance, directing attention to the affected hand. Choi et al⁵³ used game programs to prompt participants’ attention through sound. Kim et al⁵⁴ added repetitive magnetic stimulation (rTMS) to the left posterior parietal cortex during robotic training and auditory feedback. Park⁵⁵ performed training with visual feedback on hand movements via video animation. Chen et al⁵⁶ used three-dimensional animation and voice prompt from the exoskeleton. Potentially, these may be critical active ingredients as the passive-active movements induced by the robot because visual and auditory feedback can improve perception during upper limb exercises.

Robot-assisted therapy can improve the degree of USN at the level of structure and function (midline perception); however, in the analyzed studies, we did not observe any impact on functionality at the level of activity and participation. Chen⁵⁶ was the first to investigate the effects of exoskeleton-driven robotic training according to the International Classification of Functioning, Disability, and Health (ICF) framework. The authors noted that robotic training using an exoskeleton was feasible and safe for patients with USN after stroke. The results demonstrated that RAT could be effective for USN symptoms, arm motor impairment, and social participation, except for the performance of daily living, compared with the control group.

Although this study provided evidence to support the therapeutic concept of robot-assisted therapy in USN after stroke, there are still limitations to this review. First, the relatively small sample size of the included studies prevented the generalization of the training effects and reduced the effect size. Second, the long-term effects of robot-assisted limb activation were not confirmed by the included studies, as none of them performed follow-ups. Furthermore, the inclusion criteria for some studies were not multidimensional and did not use the gold standard test (BIT-conventional) for objective diagnosis of USN after stroke.

Therefore, future studies should provide robotic limb activation interventions in the acute and subacute stroke phases, with larger sample size, using exercises oriented to specific tasks of the patient’s daily life, and using more sensitive tests to include patients in future clinical trials.

Conclusion

An improvement in degree of USN (“midline perception”) was observed after robot-assisted limb activation therapy in stroke patients; however, there was no impact on tasks assessing visual scanning, functionality, and activities of daily living. The quality of the evidence from the available RCTs for statistically significant outcomes was high in one study and low in two studies. The applicability of these findings may be compromised, as most of the results described in this review have been obtained from studies with very small sample sizes. However, this review provides preliminary support for the use of upper limb activation by robotics in future clinical trials for USN.

Supplemental Material

sj-docx-1-nnr-10.1177_15459683221110894 – Supplemental material for Effect of Robot-Assisted Training on Unilateral Spatial Neglect After Stroke: Systematic Review and Meta-Analysis of Randomized Controlled Trials

Supplemental material, sj-docx-1-nnr-10.1177_15459683221110894 for Effect of Robot-Assisted Training on Unilateral Spatial Neglect After Stroke: Systematic Review and Meta-Analysis of Randomized Controlled Trials by Rodrigo Bazan, Bruno Henrique de Souza Fonseca, Jessica Mariana de Aquino Miranda, Hélio Rubens de Carvalho Nunes, Silméia Garcia Zanati Bazan and Gustavo José Luvizutto in Neurorehabilitation and Neural Repair

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) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Gustavo José Luvizutto

Supplementary material for this article is available on the Neurorehabilitation & Neural Repair along with the online version of this article.

References

Malhotra

PA.

Spatial neglect. Pract Neurol. 2015;15(5): 333-339.

Plummer

Morris

Dunai

Assessment of unilateral neglect. Phys Ther. 2003;83(8):732-740.

Cubelli

Definition: Spatial neglect. Cortex. 2017;92:320-321.

Swan

Unilateral spatial neglect. Phys Ther. 2001;81(9): 1572-1580.

Luvizutto

Moliga

Rizzatti

GRS

, et al. Unilateral spatial neglect in the acute phase of ischemic stroke can predict long-term disability and functional capacity. Clinics (Sao Paulo). 2018;73:e131.

Sobrinho

KRF

Santini

ACM

Marques

CLS

, et al. Impact of unilateral spatial neglect on chronic patient’s post-stroke quality of life. Somatosens Mot Res. 2018;35(3-4):199-203.

Tanaka

Ifukube

Sugihara

Izumi

A case study of new assessment and training of unilateral spatial neglect in stroke patients: Effect of visual image transformation and visual stimulation by using a Head Mounted Display system (HMD). J Neuroeng Rehabil. 2010;7:20.

Becker

Karnath

HO.

Neuroimaging of eye position reveals spatial neglect. Brain. 2010;133(Pt 3):909-914.

Bazan

Fernandes

Braga

Luvizutto

Resende

A new clinical sign probably associated to left hemiplegia with left hemineglect syndrome: the crossed legs. Arq Neuropsiquiatr. 2014;72(6):418-421.

10.

Luvizutto

Neto

Resende

LAL

Nunes

HRC

Betting

LEGG

Bazan

Crossed leg sign is associated with severity of unilateral spatial neglect after stroke. Front Neurol. 2018; 9:256.

11.

Ostrow

Llinás

RH.

Eastchester clapping sign: A novel test of parietal neglect. Ann Neurol. 2009; 66:114.

12.

Cermak

Hausser

The behavioral inattention test for unilateral visual neglect: a critical review. Phys Occup Ther Geriatr. 1989;7(3):43-53.

13.

Azouvi

Olivier

de Montety

Samuel

Louis-Dreyfus

Tesio

Behavioral assessment of unilateral neglect: study of the psychometric properties of the Catherine Bergego Scale. Arch Phys Med Rehabil. 2003;84(1):51-57.

14.

Marques

CLS

de Souza

Gonçalves

, et al. Validation of the Catherine Bergego Scale in patients with unilateral spatial neglect after stroke. Dement Neuropsychol. 2019;13(1):82-88.

15.

Azouvi

Jacquin-Courtois

Luauté

. Rehabilitation of unilateral neglect: evidence-based medicine. Ann Phys Rehabil Med. 2017;60(3):191-197.

16.

Bowen

Hazelton

Pollock

Lincoln

. Cognitive rehabilitation for spatial neglect following stroke. Cochrane Database Syst Rev. 2013;2013(7):CD003586.

17.

Dintén-Fernández

Fernández-González

Koutsou

Alguacil-Diego

Laguarta-Val

Molina-Rueda

Top-down and bottom-up approaches for the treatment of unilateral spatial neglect in stroke patients: a systematic review. Rehabilitacion (Madr). 2019;53(2):93-103.

18.

Ogourtsova

Souza Silva

Archambault

Lamontagne

Virtual reality treatment and assessments for post-stroke unilateral spatial neglect: a systematic literature review. Neuropsychol Rehabil. 2017;27(3):409-454.

19.

Pedroli

Serino

Cipresso

Pallavicini

Riva

Assessment and rehabilitation of neglect using virtual reality: a systematic review. Front Behav Neurosci. 2015;9:226.

20.

Luaut.

Halligan

Rode

Rossetti

Boisson

. Visuo-spatial neglect: a systematic review of current interventions and their effectiveness. Neurosci Biobehav Rev. 2006;30:961-982.

21.

Yang

Zhou

Chung

Li-Tsang

Fong

KN.

Rehabilitation interventions for unilateral neglect after stroke: a systematic review from 1997 through 2012. Front Hum Neurosci. 2013;7:187.

22.

Liu

KPY

Hanly

Fahey

Fong

SSM

Bye

. A systematic review and meta-analysis of rehabilitative interventions for unilateral spatial neglect and hemianopia poststroke from 2006 through 2016. Arch Phys Med Rehabil. 2019;100(5):956-979.

23.

Reinhart

Schmidt

Kuhn

, et al. Limb activation ameliorates body-related deficits in spatial neglect. Front Hum Neurosci. 2012;6:88.

24.

Machner

Dorr

Sprenger

, et al. Impact of dynamic bottom-up features and top-down control on the visual exploration of moving real-world scenes in hemispatial neglect. Neuropsychologia. 2012;50(10):2415-2425.

25.

Robertson

North

Spatio-motor cueing in unilateral left neglect: the role of hemispace, hand and motor activation. Neuropsychologia. 1992;30:553-563.

26.

Vallar

Rusconi

Barozzi

, et al. Improvement of left visuo-spatial hemineglect by left-sided transcutaneous electrical stimulation. Neuropsychologia. 1995;33:73-82.

27.

Park

JH.

Effect of robot-assisted left hand training on unilateral neglect in patients with stroke. J Kor Soc Occup Ther. 2015;23:117-127.

28.

Frassinetti

Rossi

Ladavas

Passive limb movements improve visual neglect. Neuropsychologia. 2001;39:725-733.

29.

Formisano

Barbanti

Catarci

De Vuono

Calisse

Razzano

Prolonged muscular flaccidity: frequency and association with unilateral spatial neglect after stroke. Acta Neurol Scand. 1993;88(5):313-315.

30.

Eskes

Butler

Using limb movements to improve spatial neglect: the role of functional electrical stimulation. Restor Neurol Neurosci. 2006;24:385-398.

31.

Veerbeek

Langbroek-Amersfoort

Van Wegen

, et al. Effects of robot-assisted therapy for the upper limb after stroke: a systematic review and meta-analysis. Neurorehabil Neural Repair. 2017;31:107-121.

32.

Sale

Franceschini

Mazzoleni

, et al. Effects of upper limb robot-assisted therapy on motor recovery in subacute stroke patients. J Neuroeng Rehabil. 2014;11:104.

33.

Higgins

JPT

. Cochrane handbook for systematic reviews of interventions. Version 5.1.0. The Cochrane Collaboration 2011. http://handbook-5-1.cochrane.org/.

34.

Moher

Liberati

Tetzlaff

Altman

; for the PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. BMJ. 2009;339:article b2535.

35.

Luvizutto

Fogaroli

Theotonio

Moura

Nunes

HRC

Bazan

Norm scores of cancelation and bisection tests for unilateral spatial neglect: data from a Brazilian population. Clinics (Sao Paulo). 2020;75:e1468.

36.

Marques

CLS

de Souza

Gonçalves

, et al. Validation of the Catherine Bergego Scale in patients with unilateral spatial neglect after stroke. Dement Neuropsychol. 2019;13(1): 82-88.

37.

Cermak

Hausser

The behavioral inattention test for unilateral visual neglect: A critical review. Phys Occup Ther Geriatr. 1989;7:43-53.

38.

Cincura

Pontes-Neto

Neville

, et al. Validation of the National Institutes of Health Stroke Scale, modified Rankin Scale and Barthel Index in Brazil: The role of cultural adaptation and structured interviewing. Cerebrovasc Dis. 2009;27(2):119-122.

39.

Fugl-Meyer

Jaasko

Leyman

Olsson

Steglind

The post-stroke hemiplegic patient. Scand J Rehabil Med. 1975;7:13-31.

40.

Alghadir

Al-Eisa

Anwer

Sarkar

Reliability, validity, and responsiveness of three scales for measuring balance in patients with chronic stroke. BMC Neurol. 2018;18(1):141.

41.

Mao

Hsueh

Tang

Sheu

Hsieh

CL.

Analysis and comparison of the psychometric properties of three balance measures for stroke patients. Stroke. 2002;33(4):1022-1027.

42.

Aben

Verhey

Lousberg

Lodder

Honig

Validity of the Beck depression inventory, hospital anxiety and depression scale, SCL-90, and hamilton depression rating scale as screening instruments for depression in stroke patients. Psychosomatics. 2002;43(5):386-393.

43.

Zigmond

Snaith

RP.

The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983;67(6):361-370.

44.

Hamilton

A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23(1):56-62.

45.

Mulder

Nijland

Stroke Impact Scale. J Physiother. 2016;62(2):117.

46.

Pinto

Maso

Vilela

Santos

Oliveira-Filho

Validation of the EuroQol quality of life questionnaire on stroke victims. Arq Neuro Psiquiatr. 2011;69:320-323

47.

Bugge

Hagen

Alexander

Measuring stroke patients’ health status in the early post-stroke phase using the SF36. Int J Nurs Stud. 2001;38(3):319-327.

48.

Higgins

JPT

Altman

Gotzsche

, et al. The Cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343(oct182):d5928.

49.

Guyatt

Busse

. Modification of Cochrane tool to assess risk of bias in randomized trials. http://distillercer.com/resources/.

50.

Maher

Sherrington

Herbert

Moseley

Elkins

Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther. 2003;83(8):713-721

51.

Guyatt

Oxman

Vist

, et al. GRADE guidelines: 4. Rating the quality of evidence—study limitations (risk of bias). J Clin Epidemiol. 2011;64(4):407-415.

52.

The Cochrane Collaboration, Review Manager (RevMan) 5.3. The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark, 2011.

53.

Choi

Lee

Kim

Park

Lee

SJ.

The effect of an upper limb rehabilitation robot on hemispatial neglect in stroke patients. Ann Rehabil Med. 2016;40(4):611-619.

54.

Kim

Lee

, et al. Effect of combined therapy of robot and low-frequency repetitive transcranial magnetic stimulation on hemi-spatial neglect in stroke patients. Ann Rehabil Med. 2018;42:788-797.

55.

Park

JH.

The effects of robot-assisted left-hand training on hemispatial neglect in older patients with chronic stroke: A pilot and randomized controlled trial. Medicine (Baltimore). 2021;100(9):e24781.

56.

Chen

Xiong

Huang

XL.

Robot-assisted arm training in stroke individuals with unilateral spatial neglect: A pilot study. Front Neurol. 2021;12:691444.

57.

Chege

Robotic rehabilitation effect on unilateral spatial neglect in patients with acute stroke. Arch Phys Med Rehabil. 2021;102(10):e104

58.

Spaccavento

Cellamare

Falcone

Loverre

Nardulli

Effect of subtypes of neglect on functional outcome in stroke patients. Ann Phys Rehabil Med. 2017;60:37-381.

59.

Brandt

Glasauer

Strupp

Dieterich

Spatial neglect: hypothetical mechanisms of disturbed interhemispheric crosstalk for orientation. Ann N Y Acad Sci. 2009;1164:216-221.

60.

Bowen

Hazelton

Pollock

Lincoln

NB.

Cognitive rehabilitation for spatial neglect following stroke. Cochrane Database Syst Rev. 2013;2013:Cd003586.

61.

Ferber

Karnath

HO.

How to assess spatial neglect—line bisection or cancellation tasks?

J Clin Exp Neuropsychol. 2001;23(5):599-607.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB