Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Review ArticleNoninvasive Brain Stimulations for Unilateral SpatialNeglect after Stroke: A Systematic Review and Meta-Analysis ofRandomized and Nonrandomized Controlled Trials
Flávio Taira Kashiwagi,1 Regina El Dib,2 Huda Gomaa,3 Nermeen Gawish,3
Erica Aranha Suzumura,4 Taís Regina da Silva,1 Fernanda Cristina Winckler,1
Juli Thomaz de Souza,2 Adriana Bastos Conforto ,5 Gustavo José Luvizutto ,6
and Rodrigo Bazan1
1Neurology Department, Botucatu Medical School, Universidade Estadual Paulista (UNESP), Botucatu, SP, Brazil2Science and Technology Institute, Universidade Estadual Paulista (UNESP), São José dos Campos, SP, Brazil3Department of Pharmacy, Tanta Chest Hospital, Tanta, Egypt4Research Institute, Hospital do Coração (HCor), São Paulo, SP, Brazil5Neurostimulation Laboratory, University of São Paulo (USP), São Paulo, SP, Brazil6Department of Applied Physical Therapy, Federal University of Triângulo Mineiro (UFTM), Uberaba, MG, Brazil
Correspondence should be addressed to Gustavo José Luvizutto; [email protected]
Received 7 September 2017; Accepted 15 April 2018; Published 28 June 2018
Academic Editor: Michele Fornaro
Copyright © 2018 Flávio Taira Kashiwagi et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work isproperly cited.
Background. Unilateral spatial neglect (USN) is the most frequent perceptual disorder after stroke. Noninvasive brain stimulation(NIBS) is a tool that has been used in the rehabilitation process to modify cortical excitability and improve perception andfunctional capacity. Objective. To assess the impact of NIBS on USN after stroke. Methods. An extensive search was conductedup to July 2016. Studies were selected if they were controlled and noncontrolled trials examining transcranial direct currentstimulation (tDCS), repetitive transcranial magnetic stimulation (rTMS), and theta burst stimulation (TBS) in USN after stroke,with outcomes measured by standardized USN and functional tests. Results. Twelve RCTs (273 participants) and 4 non-RCTs(94 participants) proved eligible. We observed a benefit in overall USN measured by the line bisection test with NIBS incomparison to sham (SMD −2.35, 95% CI −3.72, −0.98; p = 0 0001); the rTMS yielded results that were consistent with theoverall meta-analysis (SMD −2.82, 95% CI −3.66, −1.98; p = 0 09). The rTMS compared with sham also suggested a benefit inoverall USN measured by Motor-Free Visual Perception Test at both 1Hz (SMD 1.46, 95% CI 0.73, 2.20; p < 0 0001) and 10Hz(SMD 1.19, 95% CI 0.48, 1.89; p = 0 54). There was also a benefit in overall USN measured by Albert’s test and the line crossingtest with 1Hz rTMS compared to sham (SMD 2.04, 95% CI 1.14, 2.95; p < 0 0001). Conclusions. The results suggest a benefit ofNIBS on overall USN, and we conclude that rTMS is more efficacious compared to sham for USN after stroke.
1. Background
Stroke is the second leading cause of death worldwide and theprimary cause of chronic disability in adults [1]. In theUnited States, it is the fourth leading cause of death overall[2]. Among people who survive a stroke, unilateral spatial
neglect (USN) is the most frequent disorder for right hemi-sphere lesions [3].
The incidence of USN varies widely from 10% to 82%[4, 5]. USN is characterized by the inability to report orrespond to people or objects presented on the side contralat-eral to the lesioned side of the brain and has been associated
HindawiNeural PlasticityVolume 2018, Article ID 1638763, 25 pageshttps://doi.org/10.1155/2018/1638763
http://orcid.org/0000-0001-7869-3490http://orcid.org/0000-0002-6914-7225https://doi.org/10.1155/2018/1638763
with poor functional outcomes and long stays in hospitalsand rehabilitation centers [6].
Pharmacological interventions such as dopamine andnoradrenergic agonists or procholinergic treatment havebeen used in people affected by USN after stroke, butthe evidence derived from a Cochrane systematic reviewthat included only two available RCTs was very low andinconclusive [7].
Other nonpharmacological rehabilitation techniqueshave been explored for USN with the aim to facilitate therecovery of perception and behavior, which include righthalf-field eye-patching [8], mirror therapy [9], prism adapta-tion [10], left-hand somatosensory stimulation with visualscanning training [11], contralateral transcutaneous electri-cal nerve stimulation and optokinetic stimulation [12], trunkrotation [13], repetitive transcranial magnetic stimulation[14], galvanic vestibular stimulation [15], and dressing prac-tice [16]. However, their results do not support the use ofthese techniques in isolation for improvement of secondaryoutcomes such as performance and sensorimotor functions,activities of daily living (ADLs), or quality of life [9, 14, 17].
Noninvasive brain stimulations (transcranial direct cur-rent stimulation (tDCS) and repetitive transcranial magneticstimulation (rTMS)) have already shown their ability tomodify cortical excitability [18]. tDCS is a noninvasivemethod used to modulate cortical excitability by applying adirect current to the brain that is less expensive than repeti-tive transcranial magnetic stimulation (rTMS). The latter isan electric current that creates magnetic fields that penetratethe brain and can modulate cortical excitability by decreasingor increasing it and potentially improve perceptual and cog-nitive abilities [19, 20].
A previous Cochrane systematic review summarizedresults about the effects of tDCS versus control (sham/anyother intervention) on activities of daily living (ADLs)among stroke survivors. The authors included 32 random-ized controlled trials (RCTs) and concluded that tDCS mightenhance ADLs, but upper and lower limb function, musclestrength, and cognitive abilities should be further explored[21]. Another Cochrane systematic review assessed the effi-cacy of repetitive transcranial magnetic stimulation (rTMS)compared to sham therapy or no therapy for improving func-tion in people with stroke. The 19 included trials showed thatrTMS was not associated with a significant increase in ADLsor in motor function; therefore, the authors do not supportthe use of rTMS for the treatment of stroke, and they planto complete further trials to confirm their findings [22].
Previous reviews were, however, limited in that they didnot include non-RCT studies nor did they evaluate the new-est noninvasive brain stimulation—theta burst. We thereforeconducted a systematic review of RCT and non-RCT studiesthat assessed the impact of tDCS, rTMS, and TBS for unilat-eral spatial neglect after stroke.
2. Methods
We adhered to methods described in the Cochrane Hand-book for Intervention Reviews [23]. Our reporting alsoadheres to the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) [24] and Meta-Analysis of Observational Studies in Epidemiology (MOOSE)statements [25].
2.1. Eligibility Criteria. The eligibility criteria are as follows:
(1) Study designs: RCTs, quasi-RCTs, and non-RCTs
(2) Participants: adults over 18 years of age, regardless ofgender and the duration of illness or severity of theinitial impairment, with USN after any type of strokediagnosis (ischemic or intracranial hemorrhage)measured by clinical examination or radiographicallyby computed tomography (CT) or magnetic reso-nance imaging (MRI), regardless ofwhether theywereincluded after evaluation by standardized USN tests.
(3) Interventions: any noninvasive brain stimulationssuch as tDCS, rTMS, and including theta burst (con-tinuous TBS (cTBS) or intermittent theta burst(iTBS)) (we considered evaluating both the differenttypes of stimulations (i.e., cathodal tDCS versusanodal tDCS versus dual tDCS) and types of fre-quency (i.e., high-frequency versus low frequency))
(4) Comparators: interventions were to be comparedagainst sham stimulation or any conventional strokerehabilitation (e.g., pharmacological therapy or non-pharmacological therapy such as right half-field eye-patching, mirror therapy, prism adaptation, left-handsomatosensory stimulation, andvisual scanning train-ing or other conventional treatment)
We also considered noninvasive brain stimulations as anadjunct to any type of conventional stroke rehabilitation.
(5) Outcomes:
(i) Overall USN measured by any paper-and-pencil tests, such as the line cancellation task[26], the line bisection test [27], or the starcancellation test [28], and by any validatedspecific instrument, such as the CatherineBergego Scale [29], and the Behavioral Inatten-tion Test [30]
(ii) Disability in neurological and functional abili-ties as measured by any validated specificinstrument, such as the National Institutes ofHealth Stroke Scale and the Modified RankinScale [31], the box and block test [32], or theFugl-Meyer Assessment [33] after treatmentand over the long term
(iii) Daily life functions as measured by any vali-dated measurement scale, such as the Barthelindex [31]
(iv) Number of reported falls as measured by dia-ries of falls, by the Morse Fall Scale [34], or bythe Hendrich II Fall Risk Model [35] after treat-ment and over the long term
2 Neural Plasticity
(v) Balance as measured by the Berg Balance Scale,the balance subscale of the Fugl-Meyer test, andthe Postural Assessment Scale for StrokePatients [36] after treatment and over thelong term
(vi) Depression or anxiety as measured by the BeckDepression Inventory, the Hospital Anxietyand Depression Scale, Symptom Checklist-90(SCL-90), and the Hamilton Depression RatingScale [37] after treatment and over the longterm
(vii) Evaluation of poststroke fatigue by the FatigueSeverity Scale [38] after treatment and overthe long term
(viii) Quality of life (however defined by the studyauthors) after treatment and over the long term
(ix) Adverse events (e.g., euphoria, hallucinations,orthostatic hypotension, nausea, insomnia, diz-ziness, and syncope) after treatment and overthe long term
(x) Death
2.2. Data Source and Searches. We searched MEDLINE(OvidSP) (1966 to July 2016), EMBASE (OvidSP) (1980 toJuly 2017), the Cochrane Central Register of Controlled Tri-als (CENTRAL) (The Cochrane Library, 2017, issue 7),CINAHL (1961 to July 2017), and Latin-American andCaribbean Center on Health Sciences Information (LILACS)(from 1982 to July 2017) without language restrictions. Thedate of the most recent search was 26 July 2017. All searcheswere conducted with the assistance of a trained medicallibrarian. We also searched the reference lists of relevant arti-cles and conference proceedings and contacted the authors ofincluded trials.
The search strategy was: (tDCS OR TDCS OR CathodalStimulation Transcranial Direct Current Stimulation ORCathodal Stimulation tDCSs OR Cathodal Stimulation tDCSOR Transcranial Random Noise Stimulation OR Transcra-nial Alternating Current Stimulation OR Transcranial Elec-trical Stimulation OR dual transcranial direct currentstimulation OR Transcranial Electrical Stimulations ORAnodal Stimulation Transcranial Direct Current StimulationOR Anodal Stimulation Tdcs OR Anodal Tdcs OR AnodalStimulation TDCSs OR Repetitive Transcranial ElectricalStimulation OR repetitive transcranial magnetic stimulationOR RTMS OR rTMS OR High-frequency rTMS OR Trasn-cranial Magnetic Stimulation OR Transcranial MagneticStimulations OR Low-frequency transcranial magnetic stim-ulation OR Stimulation Transcranial Magnetic OR Stimula-tions Transcranial Magnetic OR Single Pulse TranscranialMagnetic Stimulation OR Paired Pulse Transcranial Mag-netic Stimulation OR Repetitive Transcranial MagneticStimulation OR theta burst OR theta burst stimulationOR theta-burst OR theta-burst stimulation OR burst stimu-lation OR continuous theta burst stimulation OR continuous
TBS OR TBS) AND (cerebrovascular disorders OR basalganglia cerebrovascular disease OR hemispatial neglect ORhemispatial neglect OR spatial attentional asymmetries ORbrain ischemia OR carotid artery diseases OR intracranialarterial diseases OR intracranial embolism and thrombosisOR intracranial hemorrhages OR stroke OR brain infarctionOR vertebral artery dissection OR post-stroke OR poststrokeOR hemineglect OR hemi-neglect OR unilateral visuospatialneglect OR visuospatial neglect OR visual spatial neglect ORspatial neglect OR unilateral neglect of acute stroke patientsOR unilateral spatial neglect OR patients with stroke ORstroke patients with spatial neglect OR right hemispherestrokes OR rehabilitation after stroke OR chronic spatialneglect after stroke OR unilateral neglect OR spatial neglectOR hemispatial neglect OR visual neglect OR inattentionOR hemi-inattention OR space perception OR visual percep-tion OR perceptual disorders OR perceptual disorder ORextinction OR functional laterality).
2.3. Selection of Studies. Two pairs of reviewers indepen-dently screened all titles and abstracts identified by the liter-ature search, obtained full-text articles of all potentiallyeligible studies, and evaluated them for eligibility. Reviewersresolved disagreement by discussion or, if necessary, withthird party adjudication. We also considered studies reportedonly as conference abstracts.
2.4. Data Extraction. Reviewers underwent calibrationexercises and worked in pairs to independently extractdata from included studies. They resolved disagreementby discussion or, if necessary, with third party adjudication.They abstracted the following data using a pretested dataextraction form: study design, participants, interventions,comparators, outcome assessed, and relevant statistical data.
2.5. Risk of Bias Assessment. Reviewers, working in pairs,independently assessed the risk of bias of included RCTsusing a modified version of the Cochrane Collaboration’sinstrument (http:/distillercer.com/resources/) [39]. That ver-sion includes nine domains: adequacy of sequence gener-ation, allocation sequence concealment, blinding of participantsand caregivers, blinding of data collectors, blinding foroutcome assessment, blinding of data analysts, incom-plete outcome data, selective outcome reporting, and thepresence of other potential sources of bias not accountedfor in the previously cited domains [40]. For incompleteoutcome data in individual studies, we stipulated as lowrisk of bias for loss to follow-up as less than 10% and adifference of less than 5% in missing data between inter-vention/exposure and control groups.
When information regarding risk of bias or other aspectsof methods or results was unavailable, we attempted to con-tact study authors for additional information.
2.6. Certainty of Evidence. We summarized the evidence andassessed its certainty separately for bodies of evidence fromRCT and non-RCT studies. We used the Grading of Recom-mendations Assessment, Development and Evaluation(GRADE) methodology to rate certainty of the evidence for
3Neural Plasticity
http://distillercer.com/resources
each outcome as high, moderate, low, or very low [41]. In theGRADE approach, RCTs begin as high certainty and non-RCT studies begin as moderate certainty. Detailed GRADEguidance was used to assess overall risk of bias [42], impreci-sion [43], inconsistency [44], indirectness [45], and publica-tion bias [46] and to summarize the results in an evidenceprofile (Table 1).
We planned to assess publication bias through visualinspection of funnel plots for each outcome in which weidentified 10 or more eligible studies; however, we were notable to do so because there were an insufficient number ofstudies to allow for this assessment.
2.7. Data Synthesis and Statistical Analysis. We calculatedpooled inverse variance standardized mean difference(SMD) and associated 95% CIs using random-effects models.We addressed variability in results across studies by using I2
statistic and the P value obtained from the Cochran chisquare test. Our primary analyses were based on eligiblepatients who had reported outcomes for each study (com-plete case analysis). We used Review Manager (RevMan)(version 5.3; Nordic Cochrane Centre, Cochrane) for all anal-yses [47].
2.8. Subgroup and Sensitivity Analyses. We planned possiblesubgroup analyses according to the following characteristics:
(i) Participants (stroke type: ischemic stroke versusintracranial hemorrhage)
(ii) Interventions (type of stimulation: cathodal versusanodal and position of electrodes; type of frequency:high frequency versus low frequency)
(iii) Comparator (type of control intervention: phar-macological therapy versus nonpharmacologicaltherapy)
(iv) Different tests for overall USN (star cancellation testversus line bisection test)
We planned to conduct subgroup analyses only when fiveor more studies were available, with at least two in each sub-group. We planned to synthesize the evidence separately forbodies of evidence from RCT and non-RCT studies by a sen-sitivity analysis.
3. Results
3.1. Study Selection. We identified a total of 4129 citationsthrough database searches and a further four studies fromthe reference lists of the Cochrane reviews [22, 48, 49](see Figure 1 for search results). After screening by title andthen by abstract, we obtained full-paper copies for 30 cita-tions that were potentially eligible for inclusion in the review.We excluded 15 studies for the following reasons: case report,case series, self-controlled study, review, and off-topic. Theremaining 12 RCTs [14, 50–60] with a total of 273 partic-ipants and four non-RCTs [61–64] with a total of 94 par-ticipants met the minimum requirements, and we includedthem in this review.
3.2. Study Characteristics. Table 2 describes study charac-teristics related to design of study, setting, number of par-ticipants, mean age, gender, inclusion and exclusioncriteria, and follow-up. Eight studies [14, 54, 56, 59, 61–64]were conducted largely in Europe and eight in Asia [50–53,55, 57, 58, 60]. Randomized trials’ sample sizes rangedfrom 10 [56] to 38 [55], and non-RCT studies rangedfrom 12 [63] to 36 [62]. Typical participants were malesin their 40s, 50s, and 60s. Studies followed participantsimmediately after treatment [50, 57, 58, 62] to one month[51, 52, 54–56].
Table 3 describes study characteristics related to inter-vention and comparators and assessed outcomes. Of the 16included studies, nine trials [14, 50, 52, 54, 55, 59–62] evalu-ated TBS:
(1) Eight trials compared cTBS versus
(i) sham cTBS (both groups received conventionalrehabilitation training [52, 60]);
(ii) 1Hz rTMS, 10Hz rTMS, and sham rTMS (allgroups with addition of routine rehabilitation[55]);
(iii) sham TBS [14];
(iv) sham cTBS [54, 59, 61, 62].
(2) One trial compared iTBS with 80% resting motorthreshold (RMT) versus iTBS 40% RMT [50].
Of the remaining seven studies, four trials [56–58, 64]evaluated tDCS:
(1) Two trials compared tDCS over the left (cathodal)and right (anodal) posterior parietal cortex, one ver-sus placebo at an intensity of 2mA [56] and the otherversus sham tDCS [64].
(2) One trial [57] compared tDCS dual versus eithertDCS single or tDCS sham.
(3) One trial [58] compared tDCS versus sham tDCS.
Three further trials [51, 53, 63] evaluated rTMS:
(1) One trial [51] compared rTMS with sham rTMS,both plus conventional rehabilitation therapy (neu-rodevelopmental facilitation techniques).
(2) Two trials compared 1Hz rTMS, one versus 10HzrTMS and sham rTMS [53] (both groups receivedconventional rehabilitation), and the other trial com-pared 1Hz rTMS versus sham rTMS [63].
None of the included studies evaluated noninvasive brainstimulations as an adjunct to any type of conventionalstroke rehabilitation.
3.3. Risk of Bias. Figure 2 describes the risk of bias assessmentfor the RCTs and non-RCTs, respectively. The major issue
4 Neural Plasticity
Table1:GRADEevidence
profi
leforRCTs:no
ninvasivebrainstim
ulations
forun
ilateralspatialneglectafterstroke.
Qualityassessment
Illustrative
comparative
risks(95%
CI)
Certainty
inestimates
orqu
alityof
evidence
Assum
edrisk
Correspon
ding
risk
Num
berof
participants
(studies)
Range
follow-up
Tim
ein
weeks
Riskof
bias
Inconsistency
Indirectness
Imprecision
Pub
licationbias
Sham
Non
invasive
brain
stim
ulations
OverallUSN
measuredby
star
cancellation
test
116(6)
Immediately
postintervention
4weeks
Seriou
slim
itation1
Seriou
slim
itation2
Noseriou
slim
itation3
Seriou
slim
itation4
Und
etectable
The
meanin
change
inUSN
measuredby
star
cancellation
testwas
45.29(SD
5.94)∗
The
std.
meanin
changes
inUSN
measuredby
star
cancellation
testin
the
intervention
grou
pwas
onaverage0.51
fewer
(1.89fewer
to0.88
more)
Verylow
OverallUSN
measuredby
linebisectiontest
107(5)
Immediately
postintervention
1mon
th
Seriou
slim
itation1
Seriou
slim
itation2
Noseriou
slim
itation3
Noseriou
slim
itation
Und
etectable
The
meanin
change
inUSN
measuredby
line
bisectiontestwas
35.79(SD
18.65)
∗
The
std.
meanin
changes
inUSN
measuredby
line
bisectiontestin
the
intervention
grou
pwas
onaverage2.33
fewer
(3.54fewer
to1.12
fewer)
Low
OverallUSN
measuredby
Motor-FreeVisualP
erceptionTest
38(2)2–4weeks
Seriou
slim
itation1
Noseriou
slim
itation
Noseriou
slim
itation
Noseriou
slim
itation
Und
etectable
The
meanin
change
inUSN
measuredby
Motor-FreeVisual
PerceptionTestwas
16.9(SD2.1)
∗∗
The
std.
meanin
changes
inUSN
measuredby
Motor-FreeVisual
PerceptionTestin
the
intervention
grou
pwas
onaverage1.46
more
(0.73moreto
2.20
more)
Mod
erate
OverallUSN
measuredby
Alberttestandlin
ecrossing
test
50(2)4weeks
Seriou
slim
itation1
Seriou
slim
itation2
Noseriou
slim
itation
Noseriou
slim
itation
Und
etectable
The
meanin
change
inUSN
measuredby
Alberttestandlin
ecrossing
testwas
27.33(SD4.55)∗
∗
The
std.
meanin
changes
inUSN
measuredby
Alberttestandlin
ecrossing
testin
the
intervention
grou
pwas
onaverage1.01
more
(1fewer
to3.02
more)
Low
SD=standard
error;std.=standardized.∗Baselinerisk
estimates
foroverallU
SNcomefrom
controlarm
ofFu
etal.’s
[52]
stud
y(low
estriskof
bias
trialinthemeta-analysis).
∗∗Baselinerisk
estimates
foroverall
USN
comefrom
controlarm
ofCha
etal.’s[51]
stud
y(new
esttrialinthemeta-analysis).
1 The
majorityofthestud
ieswererank
edas
high
risk
ofbiasforboth
allocation
sequ
ence
andallocation
concealm
ent.2 There
was
asubstantialh
eterogeneity
(I2 >
70%).
3 There
was
nosubstantiald
ifference
relatedto
themeanageandeligibility
criteriathrougho
utthesixinclud
edstud
ies.
4 95%
CIforabsoluteeffectsinclud
esclinically
impo
rtantbenefitandno
benefit.
5Neural Plasticity
regarding riskofbias in theRCTsandnon-RCTswasproblemsof random sequence generation [14, 50, 51, 54–59, 61–64] andconcealment of randomization [14, 50, 53–59, 61–64]. Anadditional problem was blinding of the statistician in allincluded studies.
3.4. Outcomes
3.4.1. Synthesized Results from Randomized Controlled Trials
(1) Overall USN Measured by the Star Cancellation Test. Theresults from six RCTs [51–55, 57] comparing noninvasivebrain stimulations with sham failed to show a benefit in over-all USN measured by the star cancellation test (SMD −0.51,95% CI −1.87, 0.85; p = 0 46; I2 = 90%) (Figure 3). Theresults were consistent regardless of the type of noninvasivebrain stimulations (TBS in three RCTs [52, 54, 55] (SMD−1.61, 95% CI −4.28, 1.06; p = 0 24; I2 = 93%); dual-tDCSin one RCT [57] (SMD −0.12, 95% CI −0.99, 0.76; p = 0 79 ;I2 =not applicable); and 1Hz rTMS in two RCTs [51, 53](SMD 0.57, 95% CI −2.95, 4.10; p = 0 75; I2 = 95%))(Figure 3). Certainty in evidence was rated as very lowbecause of imprecision, inconsistency, and risk of bias due tothe studies that were ranked as high risk of bias for bothallocation sequence and allocation concealment (Figure 2).
A sensitivity analysis from the same RCTs using TBS[52, 54, 55], single-tDCS [57], and 10Hz rTMS [51, 53]yielded results that were also consistent with the primaryanalysis and failed to show a difference in the effects ofnoninvasive brain stimulations compared to sham (SMD−0.62, 95% CI −1.89, 0.65; p = 0 34; I2 = 88%) (Figure 4).
(2) Overall USN Measured by the Line Bisection Test. Resultsfrom five RCTs [51–53, 55, 57] comparing noninvasivebrain stimulations with sham suggested a benefit in overallUSN measured by the line bisection test (SMD −2.33, 95%CI −3.54, −1.12; p = 0 0002; I2 = 81%) (Figure 5). The resultswere inconsistent when the data were analyzed by type ofnoninvasive brain stimulations: TBS in two RCTs [52, 55](SMD −3.08, 95% CI −6.54, 0.38; p = 0 08; I2 = 90%) anddual-tDCS in one RCT [57] (SMD −0.66, 95% CI −1.56,0.25; p = 0 15; I2 = not applicable) except by 1Hz rTMSin two RCTs [51, 53] that yielded results that were con-sistent with the overall meta-analysis (SMD −2.33, 95%CI −3.54, −1.12; p < 0 0002; I2 = 81%) (Figure 5). Certaintyin evidence was rated as low because of inconsistency andrisk of bias due to the studies that were ranked as high riskof bias for both allocation sequence and allocation conceal-ment (Figure 2).
A sensitivity analysis from the same RCTs using TBS[52, 55], tDCS [57], and 10Hz rTMS [53] yielded resultsthat were also consistent with the primary analysis and sug-gested a difference in the effects of noninvasive brain stimula-tions compared to sham (SMD −2.35, 95% CI −3.72, −0.98;p = 0 002; I2 = 85%) (Figure 6).
(3) Overall USN Measured by Motor-Free Visual PerceptionTest. The results from two RCTs [51, 53] comparing nonin-vasive brain stimulations with sham suggested a benefit
Number of recordsidentifiedthroughdatabase
searching: 4129
Number of additionalrecords
identifiedthrough other
sources: 4
Reference listfrom Cochrane
reviews: 4
Number of records after duplicatesremoved: 3327
Number of recordsscreened: 3327
Number of recordsexcluded: 3297
Number of full-textarticles
excluded, withreasons: 15
Number of full-textarticles assessedfor eligibility: 30
Number of studiesincluded inqualitative
synthesis: 16
Number of studiesincluded inquantitative
sybthesis(meta-analysis): 6
Case series: 11
Case report: 1
Self-controlled: 1
Review: 1
Off-topic 1
RCTs: 12
Non-RCTs: 4
RCTs: 6
Non-RCTs: 0
MEDLINE: 1903EMBASE: 1377CENTRAL: 830
CINAHL: 13LILACS: 6
Figure 1: Flow diagram of the systematic review.
6 Neural Plasticity
Table2:Stud
ycharacteristicsrelatedto
design
ofstud
y,setting,nu
mberof
participants,m
eanage,gend
er,and
inclusionandexclusioncriteria.
Autho
r,year
Designof
stud
yStatus
ofpu
blication
Location
No.
∗of
participants
Meanage
No.of
males
(%)
Inclusioncriteria
Exclusion
criteria
Rando
mized
controlledtrials
Cao
etal.,
2016
[50]
Parallel
RCT
Fulltext
Asia
I:7
C:6
I:55.0
C:62.0
I:85.7
C:83.3
Right-handedpatientswho
had
afirst-ever
stroke
intheright
hemisph
ereandvisuospatial
neglectwithno
rmalor
corrected-to-normalvision
NR
Cha
andKim
,2016
[51]
Parallel
RCT
Fulltext
Asia
I:15
C:15
I:64.07
C:63.33
I:64
C:60.0
Had
afirstrighthemisph
ere
stroke
(cerebralinfarction
orhemorrhage)
morethan
2weeks
before
thestud
y,which
hadbeen
confi
rmed
bycompu
tedtomograph
yor
magneticresonance
imaging(M
RI);h
adVSN
determ
ined
bylin
ebisection
tests(rightwardbias>12%)
orstar
cancelationtest
(omission
ofanynu
mber
ofstars);h
adaGlasgow
comascalescore<15;
18–80yearsold;
right-hand
ed;
norm
alvision
orno
rmal
correctedvision
;and
had
theability
toun
derstand
thestud
yandsigned
aninform
edconsentform
Allpatientsdidno
thave
brain
tumorsor
otherbrainpathology.
Excludedwerepatientswith
hemiano
pia;subarachno
idhemorrhage,veno
ussinu
sthrombosis,transientischem
icattack,reversibleischem
ia,or
acond
itionexacerbatedby
anewinfarction
orhemorrhage
site;a
medicalhistoryor
family
historyof
seizure;or
with
metaldevicesor
claustroph
obia
preventing
MRI
Fuetal.,
2015
[52]
Parallel
RCT
Fulltext
Asia
I:11
C:11
I:55.1β
C:59.5β
I:80.0
C:80.0
Right-handedpatients
withrighthemisph
ere
stroke
(hem
orrhagicor
ischem
iclesion
)confi
rmed
bycompu
tedtomograph
yor
magneticresonance
imaging>2weeks
before
thebeginn
ingof
thestud
yanddiagno
sisof
visuospatial
neglectbasedon
clinician
judgem
entandon
deficits
inat
leaston
eou
tof
two
paper-penciltests
Age
<30
yearsor
>80
years,
historyof
epilepsy,previous
head
trauma,drug
andalcoho
labuseandpsychiatricdisorders,
recurrentstroke,obvious
aphasiaandcommun
ication
obstacles,family
historyof
seizures,everuseof
tricyclic
antidepressantsor
antipsycho
tic
drugs,diam
agneticmetal
implantssuch
ascardiac
pacemakers,andvisual
fielddefects
7Neural Plasticity
Table2:Con
tinu
ed.
Autho
r,year
Designof
stud
yStatus
ofpu
blication
Location
No.
∗of
participants
Meanage
No.of
males
(%)
Inclusioncriteria
Exclusion
criteria
Fuetal.,
2017
[60]
Parallel
RCT
Fulltext
Asia
I:6
C:6
I:60.17
C:62
I:75
C:75
Had
afirstrighthemisph
ere
stroke
(cerebralinfarction
orhemorrhage)
morethan
2weeks
before
thestud
y,which
hadbeen
confi
rmed
bycompu
tedtomograph
yor
magneticresonance
imaging(M
RI);h
adVSN
determ
ined
bylin
ebisection
tests(rightwardbias>12%)
orstar
cancelationtest
(omission
ofanynu
mber
ofstars);h
adaGlasgow
comascalescore<15;
18–80yearsold;
right-hand
ed;
norm
alvision
orno
rmal
correctedvision
;and
hadthe
ability
toun
derstand
thestud
yandsign
aninform
edconsent
form
;allpatientsdidno
thave
braintumorsor
otherbrain
pathology
Patientswithhemiano
pia;
subarachno
idhemorrhage,
veno
ussinu
sthrombosis,
transientischem
icattack,
reversibleischem
ia,ora
cond
itionexacerbatedby
anewinfarction
orhemorrhagesite;a
medical
historyor
family
history
ofseizure;or
withmetal
devicesor
claustroph
obia
preventing
MRI
Smitetal.,
2015
[56]
RCT
cross-over
stud
yFu
lltext
Europ
eI:5€
C:5
€I:64.8€
C:64.8€
I:60.0€
C:60.0€
Patientswithlefthemispatial
neglectafterright-hemisph
eric
lesion
,right-handed,
olderthan
theageof
18,m
orethan
four
mon
thsafterstroke
Patientswithsevere
language
andcommun
icationdisorders,
bilateralcorticald
amage,
psychiatricdisorders,alcoho
land/or
drug
addiction,
epilepsy,
eczemaor
damages
onthescalp,
metalor
otherforeignparts
inthehead
Yangetal.,
2015
[55]
Parallel
RCT
Fulltext
Asia
I:9
I2:10
I3:9
C:10
I:46.7
I2:48.0
I3:49.4
C:47.7
I:66.6
I2:40.0
I3:55.6
C:30.0
Age
between18
and80;
firststroke
patients(cerebral
infarction
orhemorrhage)
andin
recovery
timewithin
60–180
days;U
SNconfi
rmed
bylin
ebisectiontest,star
cancellation
test,orclinical
exam
ination;
nometallic
implantof
diam
agnetic
substance;signed
the
inform
edconsent
Subarachno
idhemorrhage,
veno
ussinu
sthrombosis,and
reversibleor
transientischem
icattacks;worsening
cond
ition
andnew-onsetinfarction
orhemorrhage;GCSscore<15;
obviou
saphasiaandsevere
cognitive-commun
ication
disorders;family
historyof
epilepsy;im
paired
organ
function
orfailu
rein
the
heart,lung,liver,kidney,or
8 Neural Plasticity
Table2:Con
tinu
ed.
Autho
r,year
Designof
stud
yStatus
ofpu
blication
Location
No.
∗of
participants
Meanage
No.of
males
(%)
Inclusioncriteria
Exclusion
criteria
othervitalo
rgansandlife
expectancy
<6mon
ths;history
ofclaustroph
obiaand
uncoop
erativedu
ring
exam
ination;
andhemiano
psia
Kim
etal.[53]
Parallel
RCT
Fulltext
Asia
I:9
I2:9
C:9
I:68.6
I2:64.1
C:68.3
I:55.6
I2:44.4
C:66.7
Patientswithrightcerebral
ischem
icor
hemorrhagic
withvisuospatialneglect
(con
firm
edusingthelin
ebisectiontest);allp
atients
wereright-hand
ed
Severe
cognitiveim
pairment
makingthem
unableto
understand
theinstructions;
contraindication
sforTMS,
such
asahistoryof
epileptic
seizure,major
head
trauma,
andpresence
ofmetalin
the
skullo
rpacemaker;orun
stable
medicalor
neurologic
cond
itions
Sunw
ooetal.,
2013
[57]
RCT
cross-over
stud
yFu
lltext
Asia
I:10
I2:10
C:10
62.6¢
40.0β
Stroke
patientswithlesion
intherighthemisph
ereinvolving
theparietalcortex,and
left
USN
diagno
sedby
clinical
observationandconfi
rmed
byalin
ebisectiontest;allpatients
werepreviouslyright-hand
ed
Patientswho
hadmetallic
implantsin
thecranial
cavity,a
skulld
efect,history
ofseizure,un
controlled
medicalproblems,and
severe
cognitiveim
pairment
Cazzolietal.,
2012
[14]
Parallel
RCT
Fulltext
Europ
e24
£58.0¢
70.8¢
Ischem
icor
hemorrhagiclesion
totherighthemisph
ereand
left-sided
spatialn
eglect
determ
ined
onthebasisof
deficitsin
atleasttwoou
tof
threeclassesof
paper-
penciltestsandon
clinical
judgem
ent;allp
atientshad
tohave
norm
alor
corrected-to-
norm
alvisualacuity
History
ofepilepsy,prior
head
trauma,drug
and
alcoho
labu
se,and
major
psychiatricdisorders
Koetal.,
2008
[58]
RCT
cross-over
stud
yFu
lltext
Asia
I:15
€
C:15€
I:62.1€
C:62.1€
I:66.6€
C:66.6€
Patientswithsubacutestroke
withneglect
Patientswho
hadmetalin
thecranialcavityor
calvarium,skinlesion
sin
thearea
ofelectrod
e,un
controlledmedical
cond
itions,and
severe
cognitiveim
pairments
9Neural Plasticity
Table2:Con
tinu
ed.
Autho
r,year
Designof
stud
yStatus
ofpu
blication
Location
No.
∗of
participants
Meanage
No.of
males
(%)
Inclusioncriteria
Exclusion
criteria
Kochetal.,
2012
[54]
Parallel
RCT
Fulltext
Europ
eI:10
C:10
I:61.4#
C:71.9#
I:55.5#
C:55.5#
Right-handedpatients,w
ith
righthemisph
eresubacute
ischem
icstroke
affectedby
hemispatialneglect,confi
rmed
byradiologic(CTor
MRI)and
clinicalexam
ination
NR
Bon
nìetal.,
2011
[59]
Parallel
RCT
Con
ference
abstract
Europ
eNR
NR
NR
Subacutestroke
patients
withneglect
NR
Non
-RCTs
Cazzolietal.,
2015
[61]
Non
-RCT
cross-over
stud
y§Fu
lltext
Europ
eI:8¥
C:8
¥I:52.6and54.2α
C:53.0
NR
Patientswithleft-sided,
hemispatialneglectafter
asubacuteright-hemisph
eric
stroke;allpatientshadno
rmal
orcorrected-to-normal
visualacuity
Not
clearlyrepo
rted,
however,autho
rshave
assessed
patientsby
means
ofinternationally
accepted
safety
guidelines
forthe
applicationof
TMS,which
includ
edscreeningfora
historyof
epilepsy,prior
head
trauma,drug
and
alcoho
labu
se,and
major
psychiatricdisorders
Hop
fner
etal.,
2015
[62]
Non
-RCT
cross-over
Fulltext
Europ
eI:18
€
C:18€
I:64.5€
C:64.5€
I:50.0€
C:50.0€
Left-sided
neglect,basedon
clinicaljudgem
entand
neurop
sychologicaltesting,
aftersubacuteright-
hemisph
ericstroke;allsubjects
hadno
rmalor
corrected-to-
norm
alvisualacuity
NR
Làdavasetal.,
2015
[64]
Quasi-
RCT
Fulltext
Europ
eI:8
I2:11
C:11
I:72.0
I2:66.0
C:67.0
I:50.0
I2:54.5
C:54.5
Patientswithrighthemisph
ere
stroke
withhemispatial
neglectandperformance
ontheBehavioralInattention
Testbatterywithscores≤129
Presenceof
widespread
mentald
eterioration
(Mini-
MentalStateExamination
score<20),psychiatric
disorders,ahistoryof
prior
stroke
orhemorrhage,any
severe
internalmedicaldisease,
epilepsy,andaddition
alfactors
influencingtherisk
ofepilepsy
10 Neural Plasticity
Table2:Con
tinu
ed.
Autho
r,year
Designof
stud
yStatus
ofpu
blication
Location
No.
∗of
participants
Meanage
No.of
males
(%)
Inclusioncriteria
Exclusion
criteria
Agostaetal.,
2014
[63]
Non
-RCT
cross-over
stud
yFu
lltext
Europ
eI:6€
C:6
€I:67.83€
C:67.83
€I:66.6€
C:66.6€
Patientswithright
hemisph
ereun
ilaterallesions
dueto
acerebrovascular
stroke,con
firm
edby
radiologicalexam
ination
(CTor
MR),in
theirchronic
stageafterthestroke
(atleastsixmon
thspo
ston
set);
besides,participantswere
right-hand
ed,n
ativeItalian
speakers,and
hadno
rmalor
corrected-to-normalvisualacuity
History
orevidence
ofdegenerative
diseaseor
psychiatricdisorder
C:con
trolgrou
p;CT:com
putedtomograph
y;GCS:Glasgow
comascale;I:intervention
;MR:m
agneticresonanceim
aging;No.:num
ber;RCT:rando
mized
controlledtrial;TMS:transcranialmagneticstim
ulation;
USN
:unilateralspatialneglect.€ Participantsoftheexperimentalgroup
also
served
ascontrols.¥Five
patientswererand
omized
inparalleldesign,andthreefurtherpatientsinclud
edinboth
grou
ps.£The
authorsdid
notspecify
thesamplesize
perstud
iedgrou
p.αDatacomprises
threepatientsthat
received
both
experimentalandcontrolintervention
s.βDatawas
calculated
from
10patients(one
patientwas
exclud
edafter
rand
omization).¢ D
ataarefrom
thewho
lesample,as
theauthorsdidno
tspecify
itperstud
iedgrou
p.# D
ataarefrom
9patients
ineach
grou
p.§ The
stud
ywas
across-over
foron
lythreepatients,forthe
remaining
tenpatientsthestud
ywas
aRCT.
11Neural Plasticity
Table3:Stud
ycharacteristicsrelatedto
intervention
andcontrolgroup
s,assessed
outcom
es,and
follow-up.
Autho
r,year
Description
ofintervention
sDescription
ofcontrolgroup
sMeasuredou
tcom
esFo
llow-up
Rando
mized
controlledtrials
Cao
etal.,
2016
[50]
iTBS80%
RMTin
therTMSgrou
p:stim
ulationwas
appliedusingan
87mm
butterflycoilconn
ectedto
aMagstim
Rapid2(M
agstim
Co.,W
hitland
,UK),
withpeak
intensityof
2.0Tandamaxim
umpu
lselength
of250μs.Pulses(theta
bursttype)
weredelivered
totheleftdo
rsallateralp
refron
tal
cortex,the
F5labelo
fthelefthemisph
ere,which
isbetweentheF3
andF7,at80%
ofrestingmotor
threshold.
Twosessions
wereappliedwitha
15min
intervalon
each
day.Eachsessioninclud
ed20
stim
ulationtrains
consisting
ofthreepu
lses
delivered
atafrequencyof
50Hzin
every200msfor
2s(total10
bursts,30pu
lses)withan
intervalof
8s.
Sameas
intervention
grou
p;ho
wever,p
ulseswere
delivered
at40%
ofRMT
Line
bisectionand
star
cancellation
tests
After
intervention
Cha
andKim
,2016
[51]
RepetitiverT
MS+convention
alrehabilitation
therapy(neurodevelopm
entalfacilitation
techniqu
es)
foratotalo
f40
minutes
(rTMS:10
min;rehabilitation
:30
min)perday,witha10-m
inuterestperiod
halfw
aythroughthesession,
for4weeks,5
days
perweek:
stim
ulationwas
delivered
usingfigure-of-eightcoil
withadiam
eter
of80
mm
conn
ectedto
Magstim
Rapid2(M
agstim
Co.Ltd.,W
ales,U
K).Stim
ulation
was
appliedin
therightpo
steriorparietal(P3andP4
areas)basedon
theelectroencephalogram
10/20
system
atafrequencyof
1Hzfor5minutes
with
90%
ofthemotor
thresholddu
ring
rest.
Sham
rTMSandconvention
alrehabilitationtherapyusing
thesameprotocol
than
the
experimentalgroup
Motor-FreeVisual
PerceptionTest;lin
ebisectiontest;A
lberttest;
star
cancellation
test
4weeks
Fuetal.,
2015
[52]
Leftpo
steriorparietalcortex
cTBS+convention
alrehabilitationtraining:cTBSwas
setover
P5,three-
pulsebu
rstwas
delivered
at30
Hzandrepeated
every
200msfor40
swithintensitywas
80%
oftheresting
motor
threshold.
cTBSwas
delivered
usingaSuper
Rapid
2magneticstim
ulator
(Magstim
,Whitland
,UK)
with2.0-Teslamaxim
umfieldstrength,con
nected
with
afigure-of-eightcoil(diameter
ofou
tsideloop
,87
mm).Patientsreceived
4trains
daily,w
ithan
intervalof
15min,for
14consecutivedays.
Sham
cTBS+convention
alrehabilitationtraining
Star
cancellation
test;
linebisectiontest
4weeks
12 Neural Plasticity
Table3:Con
tinu
ed.
Autho
r,year
Description
ofintervention
sDescription
ofcontrolgroup
sMeasuredou
tcom
esFo
llow-up
Fuetal.,
2017
[60]
The
cTBSgrou
preceived
continuo
usTBSwiththecoil
placed
tangentiallyto
thescalpat
P3over
theleft
posteriorparietalcortex
(according
tothe10–20
electrod
epo
sition
system
oftheAmerican
Electroenceph
alograph
icAssociation
28).The
magnitude
ofthepu
lses
was
maintainedat
80%
restingmotor
threshold.
Oneach
dayfor10
consecutivedays,4
sessions
ofstim
ulation
weredelivered,w
ithan
intervalof
15min
between
every2sessions.E
achsessionlasted
40sandcontained
600pu
lses
delivered
in200bu
rstsat
5Hz(theta
rhythm
).Eachbu
rstinclud
ed3pu
lses
delivered
at30
Hz.
The
active
controlgroup
received
stim
ulations
with
thesamefeatures
atthe
samepo
sition
asthe
cTBSgrou
p,bu
twith
thecoilplaced
perpendicular
tothescalpsurfaceandthe
amplitud
eof
thestim
ulation
pulses
redu
cedto
40%
restingmotor
threshold
Star
cancellation
test;
linebisectiontest
10days
Smitetal.,
2015
[56]
tDCSwas
appliedfor20
minutes
over
theleft(catho
dal)
andright(ano
dal)po
steriorparietalcortex
onfive
consecutivedays
withabattery-driven
directcurrent
stim
ulator
(NeuroCon
nDC-Stimulator;serialnum
ber
0096).Stim
ulationparametersweresetat
acurrentof
2000
mA,and
aresistance
of
Table3:Con
tinu
ed.
Autho
r,year
Description
ofintervention
sDescription
ofcontrolgroup
sMeasuredou
tcom
esFo
llow-up
Group
I3:cTBStwotimes
adayfor2
weeks
+routinerehabilitation:
stim
ulationwas
administeredusingarapidmagneticstim
ulator
(Magstim
Com
pany)withafigure-of-eightcoil,
peak
intensityof
stim
ulationat
2T,and
pulsedu
ration
of250s,at
thecontralateral
hemisph
ere(P3),intensity
80%
ofRMT,
801pu
lses,inbu
rstsof
3pu
lses
at30
Hz,repeated
every100ms.
Kim
etal.,
2013
[53]
Group
A:10sessions
oflow-frequ
ency
(1Hz)
rTMS
over
theno
nlesionedleftpo
steriorparietalcortex
(P3)
ata90%
motor
thresholdin
4trains
of5-minute
duration
,eachseparatedby
1minute(resulting
ina
totalstimulationperiod
of20
minutes).rTMSwas
delivered
usingaMagstim
SuperRapid
Magnetic
Stim
ulator
witha70-m
illim
eter,air-cooled8-shaped
coil.rTMSwas
performed
5times
perweekfor2weeks.
Patientsalso
received
convention
alrehabilitation
treatm
ent(including
physical,occup
ational,and
cognitivetherapies).
Group
B:10sessions
ofhigh-frequ
ency
(10Hz)
rTMS
over
thelesion
edrightpo
steriorparietalcortex
(P4)
ata
90%
motor
thresholdin
4trains
of5-minutedu
ration
,each
separatedby
55second
s(resulting
inatotal
stim
ulationperiod
of20
minutes).The
remaining
oftheprotocol
followed
thesameinstructions
asgrou
pA.
Sham
rTMS+
convention
alrehabilitation
Motor-FreeVisual
PerceptionTest;
linebisectiontest;
cancellation
test;
Catherine
Bergego
scale;
Korean-mod
ified
Barthelindex
2weeks
Sunw
ooetal.,
2013
[57]
Group
A:d
ual-mod
e(tDCSdu
al)directcurrentwas
delivered
bytwosetsof
battery-po
wered
devices
(Pho
resorIIAutoMod
-elPM850,IO
MED,U
SA)
usingtwopairsof
surfacesalin
e-soaked
spon
geelectrod
es(5cm
×5cm
).Ano
daltDCSof
thefirstcircuit
over
therightPPC(P4)
was
accompanied
bycathod
altD
CSof
thesecond
circuitover
theleftPPC
(P3).T
herefore,inthefirsttD
CScircuit,theanod
ewas
placed
over
P4andthecathod
ewas
placed
over
theleft
supraorbitalarea.Inthesecond
tDCScircuit,the
anod
ewas
placed
over
therightsupraorbitalarea
and
thecathod
ewas
placed
over
theP3.Aconstant
current
of1mAwas
delivered
for20
min.
Sham
mod
e(tDCSsham
)in
thefirstandsecond
tDCS
circuits.T
hestim
ulator
was
turned
onandthecurrent
intensitywas
gradually
increasedfor5s,andwas
then
taperedoff
over
5s
Line
bisectiontest;
star
cancelationtest
Immediately
aftertreatm
ent
14 Neural Plasticity
Table3:Con
tinu
ed.
Autho
r,year
Description
ofintervention
sDescription
ofcontrolgroup
sMeasuredou
tcom
esFo
llow-up
Group
B:Single-mod
e(tDCSsingle)
directcurrentwas
delivered
bytwosetsof
battery-po
wered
devices
(Pho
resorIIAutoMod
-elPM850,IO
MED,U
SA)using
twopairsof
surfacesalin
e-soaked
spon
geelectrod
es(5cm
×5cm
).The
anod
ewas
placed
over
P4andthe
cathod
eover
theleftsupraorbitalarea
(the
firsttD
CS
circuit),and
realstim
ulationwas
provided,w
hereas
the
second
tDCScircuitreceived
sham
stim
ulation.
Forthe
realstim
ulation,
aconstant
currentof
1mAwas
delivered
for20
min.F
orthesham
stim
ulation,
the
stim
ulator
was
turned
onandthecurrentintensitywas
gradually
increasedfor5s,andwas
then
taperedoff
over
5s.
Cazzolietal.,
2012
[14]
cTBSfor2days
onweek1andsham
TBSfor2days
onweek2.cT
BSwas
appliedby
means
ofaMagPro
X100stim
ulator
(Medtron
icFu
nction
alDiagnostics)
conn
ectedto
aroun
dcoilwith60
mm
outer
radius
(MagneticCoilT
ransdu
cerMC-125).cT
BS
protocol
comprised
801pu
lses,d
elivered
ina
continuo
ustrainandconsisting
of267bu
rsts,
each
onecontainedthreepu
lses
at30
Hz,repeated
at6Hz(totaldu
ration
ofon
esingle,cTBStrainwas
44s),
andeightcT
BStrains
wereappliedover
2days.cTBS
was
appliedover
P3.Besides,p
atientsreceived
neurorehabilitation
therapyinclud
ing1h
neurop
sychologicaltraining,1
hof
occupation
altherapy,and1hof
physiotherapyperday.
Con
trol
A:sham
TBSfor2
days
onweek1andcT
BSfor2
days
onweek2.cT
BSprotocol
was
thesamedescribedfor
intervention
A.B
esides,p
atients
received
neurorehabilitation
therapyinclud
ing1h
neurop
sychologicaltraining,1
hof
occupation
altherapy,and1h
ofph
ysiotherapyperday
Con
trol
B:sham
TBSfor2days
onweek1andsham
TBSfor2
days
onweek2.Besides,
patientsreceived
neurorehabilitation
therapy
includ
ing1h
neurop
sychologicaltraining,
1hof
occupation
altherapy,and
1hof
physiotherapyperday
Catherine
Bergego
scale;Vienn
aTest
System
;rando
mshapecancelation
test
2weeks
Koetal.,
2008
[58]
tDCSto
therightpo
steriorparietalcortex
for20
min
(2mAanod
alDCbrainpo
larization
)deliveryby
abattery-po
wered
device
(Pho
resorIIAutomod
elPM850,IO
MED,U
SA),usingapairof
salin
e-soaked
surfacespon
geelectrod
es(5cm
×5cm
).The
anod
ewas
placed
over
P4,andcathod
ewas
placed
over
left
supraorbitalarea.
Sham
tDCS(current
was
delivered
for10
sandthen
turned
off)
Line
bisectiontest;
shape-un
structured
cancellation
test;
letter-structured
cancellation
test
Immediatelypo
stintervention
15Neural Plasticity
Table3:Con
tinu
ed.
Autho
r,year
Description
ofintervention
sDescription
ofcontrolgroup
sMeasuredou
tcom
esFo
llow-up
Kochetal.,
2012
[54]
cTBSwas
delivered
usingaMagStim
SuperRapid
magneticstim
ulator
(Magstim
Com
pany,W
hitland
,Wales,U
K),conn
ectedwithafigure-of-eightcoilwitha
diam
eter
of70
mm.Ineach
session,
3-pu
lsebu
rstsat
50Hzrepeated
every200msfor40
sweredelivered
at80%
oftheactive
motor
thresholdover
theleftPPC(600
pulses).Every
day,2sessions
ofleftPPCcT
BSwere
appliedwithan
intervalof
15minutes
andlasted
for
10days
(5days
perweek,Mon
dayto
Friday).Patients
also
received
rehabilitationprogram
consistedof
20sessions
of45
minutes
each,h
eld5days
perweek(based
oncompu
terizedvisuospatialscanning
training)and
motor
rehabilitationwhennecessary.
Sham
cTBSwas
delivered
withthecoilangled
at90
° ,withon
lytheedge
ofthecoil
restingon
thescalp
Stim
ulus
intensity,expressed
asapercentage
ofthemaxim
umstim
ulator
output,w
assetat
80%
oftheactive
motor
thresholdindu
cing
thesame
acou
sticsensationas
forrealTBS
Patientsalso
received
rehabilitationprogram
Line
crossing
test;
letter
cancellation
test;
star
cancellation
test;
figure
andshape
copyingtest;
representative
draw
ingtest
1mon
th
Bon
nìetal.,
2011
[59]
cTBSover
theleftPPC,for
twoweeks.
Sham
cTBS
Standardized
behaviou
ral
inattentiontest;
excitabilityof
the
parieto-fron
tal
function
alconn
ection
s
NR
Non
-RCTs
Cazzolietal.,
2015
[61]
cTBSover
theleft,con
tralesionalP
PC(P3),w
asapplied
usingaMagPro
X100stim
ulator,con
nected
toeither
aroun
dcoil(M
C-125
Medtron
icFu
nction
alDiagnostics).
The
cTBSprotocol
consistedof
801pu
lses
delivered
inacontinuo
ustrain.
The
trainwas
comprised
of267bu
rsts,w
here
each
containedthreesinglepu
lses
at30
Hz,repeated
at6Hz,andhadatotald
urationof
44s.
App
licationconsistedon
twotrains
separatedby
a15
min
interval.
Sham
cTBSover
theleft,
contralesion
alPPC,w
asappliedusingasham
coil
(MC-P-B70
Medtron
icFu
nction
alDiagnostics)
Com
puterised
balloon
testwith
eyemovem
ent
recording;paper-pencil
cancellation
tasks
8ho
urs
Hop
fner
etal.,
2015
[62]
cTBScomprised
801pu
lses,d
elivered
inacontinuo
ustrainof
267bu
rsts(eachinclud
ing3pu
lses
at30
Hz,
repeated
at6Hz).T
hetotald
urationof
asingle
cTBStrainwas
44s.TwocT
BStrains
wereapplied
overP3,withan
intertrain
intervalof
15min.A
MagPro
X100stim
ulator
(Medtron
icFu
nction
alDiagnostics,F
arum
,Denmark),con
nected
toarou
ndcoil(M
agneticCoilT
ransdu
cerMC-125)was
used
todeliver
biph
asic,repetitivemagneticpu
lses.B
esides,12
(from
18)patientsalso
received
smooth
pursuiteye
movem
enttraining.
Sham
cTBSconn
ectedto
aplacebocoil(M
agneticCoil
Transdu
cerMC-P-B70)
Besides,12(from
18)
patientsalso
received
Smooth
pursuiteye
movem
enttraining
Centerof
cancellation
score;x-po
sition
ofleftmostcancelled
target;n
umberof
cancelledtargets
Right
after
treatm
ent
16 Neural Plasticity
Table3:Con
tinu
ed.
Autho
r,year
Description
ofintervention
sDescription
ofcontrolgroup
sMeasuredou
tcom
esFo
llow-up
Làdavasetal.,
2015
[64]
Group
A:2-w
eekrehabilitationprogram
consistedof
10sessions
ofcathod
altD
CSlasting30
minutes
each
and
held
5days
perweek.tD
CSwas
appliedusingabattery-
driven
Eldith(neuroCon
nGmbH
,Ilm
enau,G
ermany)
Program
mableDirectCurrent
Stim
ulator
withapair
ofsurfacesalin
e-soaked
spon
geelectrod
es.Ineach
session,
aconstant
currentof
2mAintensity(current
density:0.57
mA/cm2)
was
delivered
lasting20
minutes
ofcathod
altD
CSof
theleft,intactPPC(overP5).
Group
B:2-w
eekrehabilitationprogram
consistedof
10sessions
ofan
odaltD
CSlasting30
minutes
each
and
held
5days
perweek.The
anod
altD
CSwas
placed
over
thePPCof
thedamaged
hemisph
ere(P6).T
heremaining
protocol
was
thesameused
ingrou
pA.
Sham
tDCS(m
ontage
used
inthesham
grou
pmim
ickedthat
used
inthetwoactive
grou
ps)
BehavioralInattention
Test
Finalfollow-up
withinthefirstweek
afterthelastsession
Agostaetal.,
2014
[63]
A10-m
inutetrainof
repetitive
low-frequ
ency
(1Hz)
rTMSover
theleftparietallobe
(P3site)identified
using
the10/20EEGmeasurementsystem
.The
stim
ulus
was
delivered
usinga70
mm
figure-of-eightcoil
conn
ectedto
aMagstim
Rapid2(M
agstim
Co.,U
K).
Stim
ulationstrength
was
setto
90%
ofthethresholdto
evokemotor
respon
sesat
rest.
Sham
rTMSover
the
intactleftparietalcortex
Visualtrackingtask;
unilateraland
bilateral
tasks
30minutes
C:con
trol
grou
p;cT
BS:continuo
usthetabu
rststim
ulation;
I:intervention
;iTBS:interm
ittent
thetabu
rst;PPC:p
osterior
parietalcortex;R
MT:resting
motor
threshold;
rTMS:repetitive
transcranialmagnetic
stim
ulation;
tDCS:transcranialdirectcurrentstim
ulation;
TBS:thetabu
rststim
ulation;
USN
:unilateralspatialneglect.
17Neural Plasticity
in overallUSNmeasured by theMotor-FreeVisual PerceptionTest at 1Hz (SMD1.46, 95%CI0.73, 2.20;p < 0 0001; I2= 0%),and the difference was not observed using 10Hz (SMD 0.97,95% CI −0.02, 1.96; p = 0 06; I2 =not applicable) (Figure 7).
Certainty in evidence was rated as moderate because ofrisk of bias due to the studies that were ranked as highrisk of bias for both allocation sequence and allocationconcealment (Figure 2).
Agosta 2014
Bonní 2011
Cao 2016
Cazzoli 2012
Cazzoli 2015
Cha 2016
Fu 2015
Fu 2017
Hopfner 2015
Kim 2013
Ko 2008
Koch 2012
Làdavas 2015
Smit 2015
Sunwoo 2013
Yang 2015
Rand
om se
quen
ce g
ener
atio
n (s
elec
tion
bias
)
Allo
catio
n co
ncea
lmen
t (se
lect
ion
bias
)
Blin
ding
of p
atie
nts
Blin
ding
of c
areg
iver
s
Blin
ding
of d
ata c
olle
ctor
s
Blin
ding
of s
tatis
ticia
n
Blin
ding
of o
utco
me a
sses
sors
Inco
mpl
ete o
utco
me d
ata (
attr
ition
bia
s)
Sele
ctiv
e rep
ortin
g (r
epor
ting
bias
)
Oth
er b
ias
− − + + + +
+ + + +
+ +
+ +
+
+
++ + + +
+
+
+
+ + + +
+
+
+
++ +
+
++
+
+
+
+
+
+
+
+
+
+
++
+ +
+
+ +
+ + +
+ + +
+ +
+
+
+ + +
+
+
+
+
+ + ++
− −
− − −
− −
− −
−
−
−
−
−
−
−
−
− −
− −
− − − −
− −
−
− − −
−
−
−
− − − − −
− −
− −
−
−−−
− −
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
− − − −
− − − −
−
−
−
−
− − −
Figure 2: Risk of bias assessment for RCTs and non-RCTs.
18 Neural Plasticity
(4) Overall USN Measured by Albert’s Test and the LineCrossing Test. The results from two RCTs [51, 54] com-paring noninvasive brain stimulations with sham failedto show a benefit in overall USN measured by Albert’s testand the line crossing test (SMD 1.01, 95% CI −1.0, 3.02;p = 0 32; I2 = 90.2%) (Figure 8). However, in the subgroupanalysis with the use of 1Hz rTMS, we found a statisti-cally significant difference compared to sham (SMD 2.04,95% CI 1.14, 2.95; p < 0 00001; I2 =not applicable). Regard-ing the use of TBS, there was no benefit compared to sham(SMD −0.01, 95% CI −0.89, 0.87; p = 0 98; I2 =not applica-ble). Certainty in evidence was rated as low because of incon-sistency and risk of bias due to the studies that were ranked ashigh risk of bias for both allocation sequence and allocationconcealment (Figure 2).
(5) Other Outcomes: Daily Life Functions and AdverseEvents. Only Kim et al. [53] reported on daily life func-tions with a higher mean in the 10Hz rTMS group thanin the sham and 1Hz rMTS groups; however, there wasonly a statistically significant difference favoring the10Hz rTMS group compared to the sham group (SMD1.83, 95% CI 0.68, 2.97; p = 0 002; I2 = not applicable).Làdavas et al.’s study [64] was the only study that reportedon adverse events; no significant adverse effect of tDCS wasreported, except only a few cases of minimal irritation ofthe skin beneath the electrodes.
None of the included studies reported on the followingoutcomes: neurological and functional disabilities, loss ofbalance, depression or anxiety and evaluation of poststrokefatigue, quality of life, and death.
Noninvasive brain stimulations Sham Std. mean differenceIV, random, 95% CI
Std. mean differenceIV, random, 95% CIStudy or subgroup Mean SD Total Mean SD Total Weight
1.4.1 TBS
Subtotal (95% CI)
Fu 2015Koch 2012Yang 2015
6.2546.8918.46
5.947.964.91
45.2948
14.79
5.945.724.59
12.2%17.8%17.6%47.6%
−6.29 (−8.64, −3.95)−0.15 (−1.03, 0.72)0.74 (−0.23, 1.70)
−1.61 (−4.28, 1.06)
10109
29
10109
29
Subtotal (95% CI)
Subtotal (95% CI)
16.316.4
5.35.4
159
24
159
24
18.1%16.4%34.5%
−1.19 (−1.97, −0.40)2.41 (1.13, 3.69)
0.57 (−2.95, 4.10)
23.43.6
6.34.7
Total (95% CI)
1.4.2 tDCSSunwoo 2013 16.61 36.31 21.37 41.52 17.8%
17.8%−0.12 (−0.98, 0.76)−0.12 (−0.99, 0.76)
1010
1010
Heterogeneity: Tau2 = 5.01; Chi2 = 29.54, df = 2 (P = 0.00001); I2 = 93%
Heterogeneity: Tau2 = 6.17; Chi2 = 21.95, df = 1 (P = 0.00001); I2 = 95%
Heterogeneity: Tau2 = 2.50; Chi2 = 51.53, df = 5 (P = 0.00001); I2 = 90%
Heterogeneity: not applicableTest for overall effect: Z = 0.26 (P = 0.79)
Test for overall effect: Z = 0.32 (P = 0.75)
Test for overall effect: Z = 0.73 (P = 0.46)
1.1.3 rTMSCha 2016Kim 2013
6363 100.0% −0.51 (−1.87, 0.85)
Test for subgroup differences: Chi2 = 1.29, df = 2 (P = 0.52), I2 = 0%
Test for overall effect: Z = 1.18 (P = 0.24)
0 10−10 −5 5Favours noninvasive brain stimulations Favours sham
Figure 3: Meta-analysis of overall USN measured by the star cancellation test.
Noninvasive brain stimulations Sham Std. mean differenceIV, random, 95% CI
Std. mean differenceIV, random, 95% CI
Favours noninvasive brain stimulations−10 −5 0 5 10
Study or subgroup Mean SD Total Mean SD Total Weight
1.2.1 TBS
1.2.2 tDCS
1.2.3 rTMSCha 2016Kim 2013
16.310.4
5.33.6
109
19
58 58 100.0% −0.62 (−1.89, 0.65)
23.43.6
6.34.7
109
19
17.6%17.2%34.8%
−1.17 (−2.13, −0.20)1.55 (0.46, 2.63)
0.18 (−2.48, 2.84)
Sunwoo 2013 12.2 31.37 1010
1010
17.9%17.9%
−0.24 (−1.12, 0.64)−0.24 (−1.12, 0.64)
21.37 41.52
Fu 2015Hoch 2012Yang 2015Subtotal (95% CI)
Subtotal (95% CI)
Subtotal (95% CI)
Heterogeneity: Tau2 = 5.01; Chi2 = 29.54, df = 2 (P < 0.00001); I2 = 93%
Heterogeneity: Tau2 = 3.41; Chi2 = 13.41, df = 1 (P < 0.00003); I2 = 93%
Heterogeneity: Tau2 = 2.14; Chi2 = 43.27, df= 5 (P < 0.00001); I2 = 88%
Heterogeneity: not applicable
Test for overall effect: Z =1.18 (P = 0.24)
Test for overall effect: Z = 0.53 (P = 0.60)
Test for overall effect: Z = 0.96 (P = 0.34)Test for subqroup differences: Chi2 = 1.06, df = 2 (P = 0.59), I2 = 0%
Test for overall effect: Z = 0.13 (P = 0.90)
Total (96% CI)
6.2546.8918.46
5.947.964.91
10109
29
10109
29
11.7%17.9%17.6%47.3%
−6.29 (−8.64, −3.95)−0.15 (−1.03, 0.72)0.74 (−0.23,1.70)
−1.61(−4.28, 1.06)
45.2948
14.79
5.945.724.59
Favours sham
Figure 4: Sensitivity analysis of overall USN measured by the star cancellation test using TBS, single-mode tDCS, and 10Hz rTMS.
19Neural Plasticity
3.4.2. Synthesized Results from Non-RCTs. The non-RCTsdid not report data in a usable way to allow for any statis-tical analysis.
4. Discussion
4.1. Main Findings. Based on pooled data from six random-ized trials with 116 participants, we found evidence for a ben-efit in overall USN with noninvasive brain stimulation,especially with the use of rTMS in comparison to the sham(Figures 5, 7, and 8). The evidence is from moderate-quality evidence because of risk of bias due to the studies thatwere ranked as high risk of bias for both allocation sequenceand allocation concealment (Figure 2). Non-RCT studiesprovided no evidence, suggesting that future trials should
adhere to CONSORT guidelines to ensure clarity and repro-ducibility in the reporting of methods.
We presented the results of overall USN in a forest plot,which showed a statistically significant difference betweenthe noninvasive brain stimulations and sham in the followingtests: line bisection test, Motor-Free Visual Perception Test,and Albert’s test and line crossing test. Nevertheless, thestudy also showed a nonsignificant difference between thenoninvasive brain stimulations and sham on the star cancel-lation test.
Several noninvasive brain stimulations have beenexplored to determine whether some of these techniquesmight be useful in promoting recovery from USN afterstroke. The lesion of the right parietal cortex after strokecauses disinhibition of the left hemisphere and thus a
Noninvasive brain stimulations Sham Std. mean differenceIV, random, 95% CI
Std. mean differenceIV, random, 95% CIStudy or subgroup Mean SD Total Mean SD Total Weight
1.4.1 TBS
Subtotal (95% CI)
Subtotal (95% CI)
Fu 2015Yang 2015
11.1722.36
14.275.31
−1.42 (−2.42, −0.42)−4.96 (−6.94, −2.97)−3.08 (−6.54, 0.38)
109
19
101020
21.8%15.2%36.9%
35.7954.12
18.656.76
Subtotal (95% CI)
19.33−30
6.879.2
159
24
34.6−8.3
44.2
159
24
21.7%19.0%40.7%
−2.64 (−3.66, −1.63)
−2.73 (−3.55, −1.90)−2.89 (−4.30, −1.48)
Total (95% CI)
Heterogeneity: Tau2 = 5.61; Chi2 = 9.70, df = 1 (P = 0.002); I2 = 90%
Heterogeneity: Tau2 = 0.00; Chi2 = 0.08, df = 1 (P = 0.78); I2 = 0%
Heterogeneity: Tau2 = 1.49; Chi2 = 21.15, df = 4 (P = 0.0003); I2 = 81%
Test for overall effect: Z = 1.75 (P = 0.08)
1.3.2 tDCSSunwoo 2013 7.05 7.45 10
101010
14.27 12.85 22.4%22.4%
−0.66 (−1.56, 0.25)−0.66 (−1.56, 0.25)
Test for overall effect: Z = 6.49 (P < 0.00001)
Test for overall effect: Z = 3.77 (P = 0.0002)
1.3.3 rTMSCha 2016Kim 2013
53 54 100.0% −2.33 (−3.54, −1.12)
Test for subgroup differences: Chi2 = 11.49, df = 2 (P = 0.003), I2 = 82.6%
Heterogeneity: not applicableTest for overall effect: Z = 1.42 (P = 0.15)
0 10−10 −5 5Favours noninvasive brain stimulations Favours sham
Figure 5: Meta-analysis of overall USN measured by the line bisection test.
Noninvasive brain stimulations Sham Std. mean differenceIV, random, 95% CI
Std. mean differenceIV, random, 95% CIStudy or subgroup Mean SD Total Mean SD Total Weight
1.4.1 TBSFu 2015Yang 2015
11.1722.36
14.275.31
109
19
35.7954.12
18.656.76
101020
21.5%16.1%37,6%
−1.42 (−2.42, −0.42)−4.96 (−6.94, −2.97)−3.08 (−6.54, 0.38)Subtotal (95% CI)
Subtotal (95% CI)
Subtotal (95% CI)
Total (95% CI)
1.4.2 tDCS
1.4.3 rTMSCha 2016Kim 2013
19.33−36,9
6.8711.2
159
24
34.6−8.3
44.2
159
24
21.5%18.8%40.3%
−2.64 (−3.66, −1.63)−3.22 (−4.73, −1.71)−2.82 (−3.66, −1.98)
Sunwoo 2013 10.4 9.64 14.27 12.85 22.1%22.1%
−0.33 (−1.21, 0.56)−0.33 (−1.21, 0.56)
1010
1010
Heterogeneity: Tau2 = 5.61; Chi2 = 9.70, df = 1 (P = 0.002); I2 = 90%
Heterogeneity: not applicable
Heterogeneity: Tau2 = 0.00; Chi2 = 0.39, df = 1 (P = 0.53); I2 = 0%
Heterogeneity: Tau2 = 2.02; Chi2 = 27.01, df = 4 (P < 0.0001); I2 = 85%
Test for overall effect: Z =1.75 (P = 0.08)
Test for overall effect: Z = 0.72 (P = 0.47)
Test for overall effect: Z = 3.35 (P < 0.0008)Test for subqroup differences: Chi2 = 16.74, df = 2 (P = 0.0002), I2 = 88.1% Favours noninvasive brain stimulations
58 58 100.0% −2.35 (−3.72, −0.98)
-10 -5 0 5 10
Test for overall effect: Z = 6,58 (P < 0.00001)
Favours sham
Figure 6: Sensitivity analysis of overall USN measured by the line bisection test using TBS, single-mode tDCS, and 10Hz rTMS.
20 Neural Plasticity
pathological overactivation of the latter. This overactivationin the left depresses the neural activity by an increased inhi-bition on the right hemisphere, aggravating the perception.The rTMS can generate currents capable of depolarizing cor-tical neurons, and tDCS changes cortical activity by means ofsmall electric currents and does not evoke action potentials.The tDCS has the advantage that the device is inexpensive,portable, and easy to use, but rTMS presented more activa-tion of the neural network and induced a neuroplasticresponse for a long-term potentiation [65].
In three of four meta-analyses, rTMS was responsiblefor the improvement of overall USN, revealing that anelectric current is an effective strategy for generating last-ing promising effects in the brain. Unfortunately, we didnot find any significant TBS or tDCS effects compared tosham procedures.
4.2. Strengths and Limitation. Strengths of our review includea comprehensive search; assessment of eligibility, risk of bias,and data abstraction independently and in duplicate; assess-ment of risk of bias that included a sensitivity analysisaddressing loss to follow-up; and use of the GRADEapproach for rating the certainty of evidence for each out-come (Table 3). Furthermore, there were no language
restrictions, and translations of non-English trials wereobtained whenever possible.
The primary limitation of our review is the low certaintyconsequent to study limitations. We identified a smallnumber of RCTs with a modest number of participantsresulting in wide confidence intervals. The total numberof participants was relatively very low (RCTs n = 278,non-RCTs n = 94) due to the small sample sizes of individualtrials, which led to downgrading the quality of evidence insome instances because underpowered trials are likely tohave a greater degree of imprecision.
Moreover, selection bias and unblinding were substantial.Another limitation of this review was having an insufficientnumber of included studies to allow for the complete statisti-cal analysis that we had planned. We were not able to assesspublication bias because there were fewer than 10 eligiblestudies addressing the same outcome in a meta-analysis. Wealso planned to perform subgroup analyses according to thecharacteristics of stroke type, type of stimulation, type of fre-quency, and comparators (type of control intervention, i.e.,pharmacological therapy versus nonpharmacological). How-ever, we also were not able to conduct these analyses becausethey did not meet our minimal criteria, which was at least fivestudies available with at least two in each subgroup.
Noninvasive brain stimulations Sham Std. mean differenceIV, random, 95% CI
Std. mean differenceIV, random, 95% CIStudy or subgroup Mean SD Total Mean SD Total Weight
Subtotal (95% CI)
Heterogeneity: Tau2 = 0.00; Chi2 = 0.02, df = 1 (P = 0.88); I2 = 0% Test for overall effect: Z = 3.89 (P = 0.0001)
1.5.1 1Hz rTMS
Cha 2016Kim 2013
21.65.4
41.8
109
19
16.93.3
2.12.3
109
19
49.4%50.6%
100.0%
1.41 (0.41, 2.41)0.97 (−0.02, 1.96)1.19 (0.48, 1.89)
1.52 10Hz rTMS
Cha 2016Kim 2013Subtotal (95% CI)
21.66.6
41.8
109
19
16.93.3
2.12.3
109
19
53.8%46.2%
100.0%
1.41 (0.41, 2.41)1.52 (0.44, 2.60)1.46 (0.73, 2.20)
Heterogeneity: Tau2 = 0.00; Chi2 = 0.38, df = 1 (P = 0.54); I2 = 0% Test for overall effect: Z = 3.30 (P = 0.0010)
Test for subgroup differences: Chi2 = 0.28, df = 1 (P = 0.60), I2 = 0%0 10−10 −5 5
Favours sham Favours noninvasive brain stimulation
Figure 7: Meta-analysis of overall USN measured by the Motor-Free Visual Perception Test.
Noninvasive brain stimulations Sham Std. mean differenceIV, random, 95% CI
Std. mean differenceIV, random, 95% CIStudy or subgroup Mean SD Total Mean SD Total Weight
1.6.1 TBSKoch 2012 35.44
35.33 2.9 1515
25 25 100.0% 1.01 (−1.00, 3.02)
27.33 4.55 1515
49.8% 2.04 (1.14, 2.95)2.04 (1.14, 2.95)49.8%
1.01 1010
1010
50.2%50.2%
−0.01 (−0.89, 0.87)−0.01 (−0.89, 0.87)
35.45 0.65Subtotal (95% CI) Heterogeneity: not applicableTest for overall effect: Z = 0.03 (P = 0.98)
1.6.2 rTMSCha 2016Subtotal (95% CI)
Total (95% CI)Heterogeneity: Tau2 = 1.90; Chi2 = 10.19, df = 1 (P = 0.001); I2 = 90% Test for overall effect: Z = 0.99 (P = 0.32)
Heterogeneity: not applicableTest for overall effect: Z = 4.42 (P < 0.00001)
Test for subgroup differences: Chi2 = 10.19, df = 1 (P = 0.001), I2 = 90.2%0 10−10 −5 5
Favours sham Favours noninvasive brain stimulation
Figure 8: Meta-analysis of overall USN measured by Albert’s test and the line crossing test.
21Neural Plasticity
Although this review presents several limitations, theissue is whether one should dismiss these results entirely orconsider them bearing in mind the limitations. The latterrepresents our view of the matter.
4.3. Relation to Prior Work. The research question we inves-tigated in our review has been addressed before from differ-ent perspectives using our population of interest but with adifferent intervention (i.e., pharmacological intervention)[7] or investigating either the intervention or the controlarms explored in this review but with a different population(e.g., idiopathic Parkinson’s disease (IPD) [48], panic disor-der in adults [66], or amyotrophic lateral sclerosis or motorneuron disease) [13].
Two Cochrane reviews [21, 49] evaluated the effect oftDCS in people after stroke but not in comparison withrTMS; instead, the authors compared tDCS with placebo,sham tDCS, no intervention, or conventional motor rehabil-itation. The first review’s [49] authors found evidence ofeffect regarding activities of daily living performance at theend of the intervention period and at the end of follow-up.However, the results did not persist in a sensitivity analysisincluding only trials of good methodological quality. In thesecond review [21], the authors found that there were nostudies examining the effect of tDCS on cognition in strokepatients with aphasia.
Another Cochrane review [22] that addressed the use ofrTMS compared to sham treatment or other conventionaltreatment for improving function after stroke revealed thatrTMS treatment was not associated with improved activitiesof daily living, nor did it have a statistically significant effecton motor function.
Three additional Cochrane reviews also discussed theeffects of both tDCS [48] and rTMS [13, 66] but in differentpopulations—in Parkinsonism [48], in patients with amyo-trophic lateral sclerosis or motor neuron disease [13], and inadults with panic disorder [66]. All reviews suffered frompoormethodological quality, imprecision, and hence low confi-dence in the estimate of the true effect to draw a consistentconclusion on the effects of noninvasive brain stimulations.
4.4. Implications. Moderate-quality evidence shows thatrTMS, at 1Hz, is more efficacious than sham for unilateralspatial neglect after stroke measured by Motor-Free VisualPerception Test. Furthermore, low-quality evidence also sug-gests a benefit of noninvasive brain stimulation, particularlywith the use of rTMS, for overall USN measured by the linebisection test, Albert’s test, and the line crossing test. Futuretrials should adhere to CONSORT guidelines to ensure clarityand reproducibility in the reporting of methods [67].
Disclosure
The funding agencies played no role in conducting theresearch or preparing the manuscript.
Conflicts of Interest
The authors have no conflicts of interest.
Authors’ Contributions
Flávio Taira Kashiwagi, Regina El Dib, Adriana Bastos Con-forto, Gustavo José Luvizutto, and Rodrigo Bazan conceivedthe review. Erica Aranha Suzumura, Taís Regina da Silva,and Fernanda Cristina Winckler undertake searches. HudaGomaa, Juli Thomaz de Souza, and Nermeen Gawishscreened search results. Flávio Taira Kashiwagi, Regina ElDib, and Gustavo José Luvizutto organized the retrieval ofpapers. Flávio Taira Kashiwagi, Huda Gomaa, NermeenGawish, Erica Aranha Suzumura, Taís Regina da Silva, andRodrigo Bazan helped in screening retrieved papers againstinclusion criteria. Flávio Taira Kashiwagi, Huda Gomaa,Nermeen Gawish, Erica Aranha Suzumura, Taís Regina daSilva, Fernanda Cristina Winckler, Juli Thomaz de Souza,and Rodrigo Bazan appraised the quality of papers. FlávioTaira Kashiwagi, Erica Aranha Suzumura, Taís Regina daSilva, and Fernanda Cristina Winckler extracted data fromthe papers. Flávio Taira Kashiwagi, Huda Gomaa, NermeenGawish, and Erica Aranha Suzumura helped in writing tothe authors of papers for additional information. Flávio TairaKashiwagi, Gustavo José Luvizutto, Juli Thomaz de Souza,and Nermeen Gawish provided additional data about thepapers. Flávio Taira Kashiwagi, Erica Aranha Suzumura, TaísRegina da Silva, Nermeen Gawish, and Huda Gomaaobtained and screened the data of unpublished studies. FlávioTaira Kashiwagi, Regina El Dib, Gustavo José Luvizutto, andRodrigo Bazan managed the data for the review. Flávio TairaKashiwagi, Regina El Dib, Erica Aranha Suzumura, GustavoJosé Luvizutto, and Rodrigo Bazan entered the data intoReview Manager (RevMan). Regina El Dib, Gustavo JoséLuvizutto, Adriana Bastos Conforto, and Rodrigo Bazan ana-lyzed the RevMan statistical data. Flávio Taira Kashiwagi,Huda Gomaa, Nermeen Gawish, Erica Aranha Suzumura,Adriana Bastos Conforto, Gustavo José Luvizutto, Regina ElDib, and Rodrigo Bazan interpreted the data. Flávio TairaKashiwagi, Regina El Dib, Huda Gomaa, Nermeen Gawish,Erica Aranha Suzumura, Gustavo José Luvizutto, andRodrigo Bazan made statistical inferences. Flávio TairaKashiwagi, Gustavo José Luvizutto, Regina El Dib, Erica Ara-nha Suzumura, and Rodrigo Bazan wrote the review. FlávioTaira Kashiwagi, Huda Gomaa, Nermeen Gawish, EricaAranha Suzumura, Gustavo José Luvizutto, Adriana BastosConforto, and Rodrigo Bazan take the responsibility forreading and checking the review before submission.
Acknowledgments
The authors would like to acknowledge São Paulo ResearchFoundation (FAPESP) (2015/14231-0) and National Councilfor Scientific and Technological Development (CNPq)(423924/2016-8) for the grants and financial support.
References
[1] R. Bonita, “Epidemiology of stroke,” Lancet, vol. 339, no. 8789,pp. 342–344, 1992.
[2] E. C. Jauch, J. L. Saver, Adams HP Jr et al., “Guidelines for theearly management of patients with acute ischemic stroke: a
22 Neural Plasticity
guideline for healthcare professionals from the AmericanHeart Association/American Stroke Association,” Stroke,vol. 44, no. 3, pp. 870–947, 2013.
[3] N. Gorgoraptis, Y. H. Mah, B. Machner et al., “The effects ofthe dopamine agonist rotigotine on hemispatial neglect follow-ing stroke,” Brain, vol. 135, no. 8, pp. 2478–2491, 2012.
[4] P. Chen, K. Hreha, P. Fortis, K. M. Goedert, and A. M. Barrett,“Functional assessment of spatial neglect: a re