Altered cerebral activities and functional connectivity in depression: a systematic review of fMRI studies

Abstract

Background
As one of the leading causes of global disability, major depressive disorder (MDD) places a noticeable burden on individuals and society. Despite the great expectation on finding accurate biomarkers and effective treatment targets of MDD, studies in applying functional magnetic resonance imaging (fMRI) are still faced with challenges, including the representational ambiguity, small sample size, low statistical power, relatively high false positive rates, etc . Thus, reviewing studies with solid methodology may help achieve a consensus on the pathology of MDD.

Methods
In this systematic review, we screened fMRI studies on MDD through strict criteria to focus on reliable studies with sufficient sample size, adequate control of head motion, and a proper multiple comparison control strategy.

Results
We found consistent evidence regarding the dysfunction within and among the default mode network (DMN), the frontoparietal network (FPN), and other brain regions. However, controversy remains, probably due to the heterogeneity of participants and data processing strategies.

Conclusion
Future studies are recommended to apply a comprehensive set of neuro‐behavioral measurements, consider the heterogeneity of MDD patients and other potentially confounding factors, apply surface‐based neuroscientific network fMRI approaches, and advance research transparency and open science by applying state‐of‐the‐art pipelines along with open data sharing.

Full text

RESEARCH ARTICLE
Altered cerebral activities and functional
connectivity in depression: a systematic
review of fMRI studies
Xue-Ying Li1,2,3,4, Xiao Chen1,5,6,7,*, Chao-Gan Yan1,2,3,5,6,7,*
1CASKeyLaboratoryofBehavioralScience,InstituteofPsychology,Beijing100101,China
2Sino-DanishCollege,UniversityofChineseAcademyofSciences,Beijing101400,China
3Sino-DanishCenterforEducationandResearch,Beijing101400,China
4CFINandPETCenter,AarhusUniversity,Nørrebrogade44,8000Aarhus,Denmark
5MagneticResonanceImagingResearchCenter,InstituteofPsychology,ChineseAcademyofSciences,Beijing100101,
China
6InternationalBig-DataCenterforDepressionResearch,ChineseAcademyofSciences,Beijing100101,China
7DepartmentofPsychology,UniversityofChineseAcademyofSciences,Beijing100049,China
*Correspondence:yancg@psych.ac.cn;chenxiao@psych.ac.cn
ReceivedJanuary29,2021;RevisedMay27,2021;AcceptedMay27,2021

Background: As one of the leading causes of global disability, major depressive disorder (MDD) places a noticeable
burden on individuals and society. Despite the great expectation on finding accurate biomarkers and effective
treatment targets of MDD, studies in applying functional magnetic resonance imaging (fMRI) are still faced with
challenges, including the representational ambiguity, small sample size, low statistical power, relatively high false
positive rates, etc. Thus, reviewing studies with solid methodology may help achieve a consensus on the pathology of
MDD.
Methods: In this systematic review, we screened fMRI studies on MDD through strict criteria to focus on reliable
studies with sufficient sample size, adequate control of head motion, and a proper multiple comparison control
strategy.
Results: We found consistent evidence regarding the dysfunction within and among the default mode network
(DMN), the frontoparietal network (FPN), and other brain regions. However, controversy remains, probably due to
the heterogeneity of participants and data processing strategies.
Conclusion: Future studies are recommended to apply a comprehensive set of neuro-behavioral measurements,
consider the heterogeneity of MDD patients and other potentially confounding factors, apply surface-based
neuroscientific network fMRI approaches, and advance research transparency and open science by applying state-of-
the-art pipelines along with open data sharing.
Keywords: depression;resting-statefMRI;task-basedfMRI;defaultmodenetwork;frontoparietalnetwork
Author summary: Major depressive disorder (MDD) is a prevalent disorder and places noticeable societal burdens,
however,findingsof objective biomarkersinprevious researches wereinconsistent.In this systematicreview,to detect
more reproducible MDD-specific brain circuits, fMRI studies on MDD are screened by strict criteria that carefully
minimize the effect of fMRI methodology issues. Though dysfunction in the default mode network (DMN) and the
frontoparietal network (FPN) was repeatedly reported, heterogeneity remains in the included studies. Based on these
findings,we highlightthenecessity ofconsideringsome specificpotentiallyconfounding factorsinfuture fMRIstudies
onMDD.

QuantitativeBiology2022,10(4):366–380
https://doi.org/10.15302/J-QB-021-0270

366 ©TheAuthors(2022).PublishedbyHigherEducationPress

INTRODUCTION
Major depressive disorder (MDD) is a prevalent dis-
orderandplacesnoticeablesocietalburdens[1].Forthe
recent 20 years, it has been stably listed in the leading
20 causes of global disability across ages and genders,
whichaccountfor 37% ofthetotaldisability causedby
mental disorders [2]. Despite the burgeoning metho-
dological advances of neuroimaging, the current
diagnosis approach of MDD is mainly based on the
clinical interview or patient rating scales, which is
associated with high rates of misdiagnosis [3]. Recent
advances in neuroimaging techniques have made it
possible to leverage the brain’s functional architecture
towards the objective biomarker of MDD. With
advantagesofsimpleness, non-invasiveness, safety,and
relatively high spatial and temporal resolutions [4],
functionalmagneticresonance imaging (fMRI)method,
especially resting-state fMRI (R-fMRI) method, has
enabled to depict of dynamic maps of the brain and
carry great expectations on finding biomarkers and
effectivetreatmenttargetsofMDD.
Unfortunately,thougha largenumberofstudies have
beenconducted,few consensuseshavebeen reachedon
theneural mechanism ofdepression.Several challenges
for fMRI studies have been noted, including between-
subject and within-subject variability, the representa-
tional ambiguity, small sample size, low statistical
power, relatively high false positive rates, and lack of
consistency in data preprocessing procedures and
statistical analysis [5–8]. For instance, the challenge of
small sample size has been proposed to dampen the
reliability of the fMRI results directly. Moreover, the
between-participantvariabilitythatsignificantlyimpacts
the reliability of fMRI studies is limited by the
measurementand the experimentaldesign,especially in
task-basedfMRIstudies[8].
To deal with the current challenges in fMRI studies,
researchers have highlighted a couple of recommenda-
tions on the designs of study, preprocessing pipelines,
and strategies to correct multiple comparisons. These
recommendations included performing adequate con-
foundregressionstrategies[9–11]andmultiplecompari-
son correction, enlarging the sample size, and openly
datasharing[12–14].Chenandhiscolleaguesassessed
the test-retest reliability of fMRI studies and found
studieswithsmallsamplesizes(<40pergroup)arenot
wellreliable[14].HeadmotionsofsubjectsinanMRI
scanner can produce artifacts, leading to results not
caused by the “real” intrinsic brain functions [15,16].
Thus, adequate head motion corrections should be
included to get reliable results. Eklund et al. reported
thattheliberalthresholdingstrategiescommonlyusedin
thefieldofneuroimagingcouldcausehighfalsepositive
rates [17], which emphasized the significance of the
effective multiple comparison correction, such as
permutation test with threshold-free cluster enhance-
ment (TFCE) and false discovery rate (FDR) based
correction [14,18,19]. Moreover, within-subjects de-
signs, paired repetitive measurements, and meta-
analyses have also been noted as effective ways to re-
ducethefalsepositivityinbrainimagingstudies[12,20].
Therefore, it is clear that adequate sample size,
confound regression strategies, and strict enough
threshold for the multiple comparison correction are
important and necessary. In the present systematic
review, we screened studies according to the criteria
such as sufficient sample size (i.e., more than 40
participants per group according to the findings of our
previous study [14]), proper head motion correction,
appropriate correction strategies for multiple compari-
sons (e.g., FWE correction with voxel-wise P < 0.001
and cluster-wise P < 0.05 or permutation test with
TFCE, FDR correction with q < 0.05). Through this
strictscreeningforstudies,weintend to focus on more
reliable studies in this review and to get more
reproducible and reliable results with less false
positivitytosomeextent.Weaimto reach a consensus
on the key brain circuits in MDD. Finally, we also
intended to raise some suggestions and directions for
furtherfMRIstudiesonMDD.
RESULTS
General information
In the first step for article identification in PubMed,
1,012 papers were obtained and fed into the next
screening procedures. After the first screening of titles
and abstracts, most review and comment articles, the
structural MRI studies, and the studies not focusing on
MDDwereexcluded,with462lefts.Then,thefulltexts
ofthesefMRI studies on depression were read through
and screened by the relatively strict criteria (see
materialsandmethods)fortheirqualities,leavingafinal
sample of 39 studies (Table1), of which 25 were R-
fMRI studies, 17 were task-based fMRI (included 9
meta-analysis studies [49–51,53,55–59] and 3 multi-
modality studies with both resting-state and task-based
fMRI [36,46,56]). Note that some multi-modality
studies only reported significant results regarding one
modality(i.e.,resting-stateortask-basedfMRI),sothey
werelistedinonlyoneresulttable.
Most of the early MRI studies on depression were
structural MRI studies, in which the earliest one found
inthissearchingwaspublishedin1993[60].Therecent
10 years have witnessed the rise of fMRI studies and
multi-modalitystudies(theearliestfMRIstudyincluded

SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 367


Table 1 A summary of all the included articles
Study Group Samplesize Realignment Strategytocontrolmultiple
comparisons
Demenescu
et al.2011[21]
HCs 56 SPM5realignment q_FDR<0.05
MDD 59
Anxietydisorders 57
Depression-anxietyco-morbidity 66
vanTolet al.
2011[22]
OutpatientswithMDD 65 SPM5realignment;excluded
whenmovement>3mm
P_FWE<0.05
MDDwithcomorbidanxiety 82
AnxietydisorderswithoutMDD 64
HCs 63
Bermingham
et al.2012[23]
MDD 44 SPM8realignment;excluded
whenmovement>4.8mm
(oneslicethickness)
P_FWE_wholebrain<0.05
HCs 44
Yanget al.
2015[24]
MDD 50 Excludedwhenmovement>
2mminx,y,orzand2°of
angularmotion
P_Alphasim_Monte_Carlo<
0.001
HCs 50
Gollier-Briant
et al.2016[25]
Healthyadolescents 685(368girls) SPM8realignment P_FWE<0.05
Posneret al.
2016[26]
Highfamilyrisk 57 SPM8realignment q_FDR<0.05
Lowfamilyrisk 47
Casement et al.
2016[27]
Longitudinalstudyfromage9‒13 123 SPM8realignment P_Alphasim<0.05
Hermesdorf
et al.2016[28]
MDD 368 DPARSF2.3realignment P_Alphasim_Monte_Carlo<
0.05(p<0.01forsinglevoxel)
HCs 461
Daveyet al.
2017[29]
MDD 71 SPM12realignment;excluded
movement>2mmor2°
P_FWE_whole_brain<0.05
HCs 88
Yükselet al.
2017[30]
HealthysubjectswithMDDriskscores 107 SPM8realignment P_Monte_Carlo_whole_brain<
0.05(clusterlevel)
Yeet al.
2017[31]
First-episodeanduntreatedMDDpatients 69 DPARSFrealignment P_AlphaSim<0.001withmore
than6voxelsofclustersize
HCs 81
Panet al.
2017[32]
NoMDDatfollow-up 529 Yes(usedAFNI,version
2011_12_21_1014,andthe
FMRIBSoftwareLibrary,
version5.0)
P_Bonferroni<0.05/55=
0.00091
MDDatfollow-up 56
Admonet al.
2017[33]
Unmedicateddepressedparticipants 46 SPM12realignment P_FWE_whole_brain<0.05
HCs 43
Lopez et al.
2018[34]
MDD-Hx 58 Sixheadrealignment
parameters
q_FDR<0.05(foragivenseed)
NoMDD-Hx 85
Mehta et al.
2018[35]
MDDpatients 48 Yes P_AFNI_3dClustsim<0.05
Qiet al.
2018[36]
MDDpatients 81 SPM8INRIalign q_FDR<0.05
HCs 123
Tokuda et al.
2018[37]
MDDpatients 67 SPM8Realignment P_Bonferroni<0.05
HCs 67
Tu et al.
2018[38]
MDDoutpatient 76 SPM8Realignment q_FDR<0.05
Wanget al.
2019[39]
MDDpatients 55 DPABIrealignment;excluded
movement>2mmor2°
P_GRF_voxel<0.001;
P_GRF_cluster<0.05
HCs 40
Fitzgeraldet al.
2019[40]
GADpatients 47 SPM8Realignment P_FWE<0.05
SADpatients 78
MDDpatients 49

Xue-YingLietal.

368 ©TheAuthors(2022).PublishedbyHigherEducationPress

(continued)
Study Group Samplesize Realignment Strategytocontrolmultiple
comparisons
Xiaet al.
2019[41]
MDDpatients 709 SPM12realignment q_FDR<0.05;P_Bonferroni<
0.05;
HCs 725
Yaoet al.
2019[42]
MDDpatients 55 SPM8realignment q_FDR<0.05
HCs 71
ChinFattet al.
2020[43]
Sertralinearmofdepression 139 SPM8realignment P_MultipleComparison<0.05
Placeboarmofdepression 140
Zhuet al.
2020[44]
MDDwithNSE 42 SPM12realignment P_FWE<0.05(clusterlevel)
MDDwithLSE 54
Hillandet al.
2020[45]
Previousdepressionwithplacebo 70 FMRIBSoftwareLibrary
version(FSLversion6.00)
TFCEwith5000permutations;
andFSLFEATcorrectionwith
p(cluster)<0.05
PreviousdepressionwithABMtraining 64
Korgaonkar
et al.2020[46]
MDDpatients 163 Yes q_FDR<0.05
HCs 62
Rupprechter
et al.2020[47]
MDDpatients 130 SPM12realignment P_whole_brain_corrected<
0.001
HCs 345
Yang et al.
2020[48]
MDDwithNSE 42 DPABIrealignment P_FWE<0.05
MDDwithLSE 54
Grahamet al.
2013a[49]
MDDpatients 566 / q_FDR<0.05
HCs 599
Groenewold
et al.2013a[50]
MDDpatients 795 / q_FDR<0.05orP_uncorrected
<0.001
HCs 792
Zhanget al.
2013a[51]
MDDpatients 341 / q_FDR<0.05
HCs 367
Iwabuchiet al.
2015a[52]
MDDpatients 225 / P<0.005
HCs 230
Wanget al.
2015a[53]
MDDpatients 160 / P<0.005
HCs 203
Zhonget al.
2016a[54]
MDDpatients 457 / P<0.001
HCs 451
Wanget al.
2017a[55]
First-episodedrug-naïveMDDpatients VBM:471;ALFF:
261
/ P<0.005
HCs VBM:521;ALFF:
278
Kambeitz
et al.2017a[56]
MDDpatients 912 / P<0.005
HCs 894
Zhouet al.
2017a[57]
MDDpatients 438 / P<0.001
HCs 421
Kerenet al.
2018a[58]
MDDpatientsornon-depressedsubjects
at-riskofMDD
653 / P<0.005
Depressiononcontinuum 503
HCs 828
Shaet al.
2018a[59]
Patientsacross11braindisorders 6683(817depressive
disorderpatients)
/ P_FDR/GRF<0.05
HCs 6692
aMeta-analysis(athresholdofuncorrected P< 0.005wasalsoacceptedfor meta-analyticstudies). Abbreviations:MDD(majordepressive
disorder),HCs(healthycontrolsubjects),MDD-Hx(historyofMDD),GAD (generalizedanxietydisorder),SAD(socialanxietydisorder),NSE
(normalsleepefficiency),LSE(lowsleepefficiency);VBM(voxel-basedmorphometry),ALFF(amplitudeoflow-frequencyfluctuations),FWE
(family-wiseerror),FDR(falsediscoveryrate),TFCE(thresholdfreeclusterenhancement).


SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 369

inthis reviewwaspublishedin2011[22],seeTable1).
Amongstudiesreviewedhere,ameta-analysisexamined
multiple neuroimaging methods and reported the lower
sensitivity and specificity of structural MRI and task-
basedfMRI methodsthanR-fMRIinthedifferentiation
of MDD patients from healthy control subjects (HCs)
[56], which may occur due to the introduction of the
potentialcomplexity fromtaskdesign andmanipulation
inmeasuringdynamicbrainfunctionsandwhichimplies
the advantages of R-fMRI in identifying neuroimaging
markersforMDD[61].
Table1showsthatSPM[62],DPABI [63],FSL [64],
and AFNI [65], realignment tools are most commonly
used in the literature. All the included studies have
appliedhead motion corrections, but some omittedkey
detailsregardingthenuisancecovariatesregressionthey
applied. For example, simply “realignment” or “app-
lyingheadmotioncorrection”wasdeclaredinthemeth-
ods section without further description. Moreover, a
large proportion of studies were excluded due to the
small sample size and inadequate multiple comparison
strategies.
Results from R-fMRI studies
AlteredspontaneousfunctionalactivitiesinMDD
Nine R-fMRI studies reported the abnormal functional
metric values in MDD patients vs. HCs, including the
regional homogeneity (ReHo), the amplitude of low-
frequency fluctuations (ALFF), and the fractional
amplitude of low-frequency fluctuations (fALFF)
(Table2).
Four studies reported abnormally increased ReHo
values in the left precuneus, the inferior frontal gyrus
(IFG),andthemedialprefrontalcortex(MPFC),aswell
as abnormally decreased ReHo values in the left
putamen, the right postcentral gyrus (poCG), the right
poCGandthelingualgyrus(LG) inpatients withMDD
[24,41,48,52]. Three studies reported abnormally
increased ALFF in the IFG, the supplementary motor
area(SMA), theleftparahippocampalgyrus(PHG),the
left anterior cingulate cortex (ACC), and the left
superior temporal gyrus (STG) and abnormally decrea-
sed activation in the orbitofrontal cortices (OFC), the
left cerebellum and the left middle temporal gyrus
(MTG) [41,55,57]. Three studies reported abnormally
increased fALFF in the visual cortex (VC), as well as
abnormally decreased fALFF in the cuneus, the
thalamus,MTG,thehippocampus, PHG, the amygdala,
the dorsolateral prefrontal cortex (dlPFC), the insula,
ACC, the superior frontal gyrus (SFG) and the inferior
parietallobule(IPL)[36,44,57].
Among these findings, abnormally decreased ALFF
andfALFFinMTG[44,57]andreducedReHoinpoCG
[41,48] were reported convergently. Furthermore, alte-
red spontaneous activities in PHG and ACC were
reported,albeitinoppositedirections[36,55,57].
Alteredresting-statefunctionalconnectivityinMDD
Fourteen studies reported abnormal functional connec-
tivity in the resting state (Table3), in which most of
them performed the seed-based analysis except one
applied the voxel-mirrored homotopic connectivity

Table 2 Altered spontaneous functional activities in local brain regions of MDD reported by R-fMRI studies
Study Metric Principalfindings
Findingofalteredincreasingactivity Findingofaltereddecreasingactivity
Yanget al.2015[24]ReHo Leftprecuneus Leftputamen
Qiet al.2018[36] fALFF VC Hippocampus,PHG,amygdala,dlPFC,insula,ACC
andIPL
Xiaet al.2019[41] ALFF,ReHo IFG(ALFF) RightpoCG(ReHo)
Yanget al.2020[48] ReHo LeftandrightLG,rightpoCG
Zhuet al.2020[44] fALFF Rightcuneus,thalamus,andMTG(inLSE)
Iwabuchiet al.2015a[52]ReHo MPFC
Zhonget al.2016a[54] ReHo,ALFF,fALFF Putamenandanteriorprecuneus MTG,STG,dlPFC,LG,PCC,posteriorprecuneus,
fusiformandoccipitalareas
Wanget al.2017a[55]ALFF BilateralSMAandleftPHG BilateralOFC
Zhouet al.2017a[57] ALFF,fALFF LeftACC(ALFF),leftSTG(ALFF) Leftcerebellum(ALFF),leftMTG(ALFF),right
SFG(fALFF)
aMeta-analysis.Abbreviations:ReHo(regionalhomogeneity),ALFF(amplitudeoflow-frequencyfluctuations), fALFF(fractionalamplitudeof
low-frequencyfluctuations),VC(visualcortex),PHG(parahippocampalgyrus),dlPFC (dorsolateralprefrontalcortex),ACC(anteriorcingulate
cortex),IPL(inferiorparietallobule),IFG(inferiorfrontalgyrus),poCG(postcentralgyrus),LG(lingualgyrus),MTG(middletemporalgyrus),
MPFC(medialprefrontalcortex),PCC(posteriorcingulatecortex),SMA(supplementarymotorarea),OFC(orbitofrontalcortex),STG(superior
temporalgyrus),SFG(superiorfrontalgyrus),LSE(lowsleepefficacygroup).


Xue-YingLietal.

370 ©TheAuthors(2022).PublishedbyHigherEducationPress

(VMHC) [28], and two performed the independent
componentanalysis(ICA)[26,38].
DecreasedVMHCwas found inSTG,theinsula,and
the precuneus [28]. In ICA studies, abnormally increa-
sed functional connectivity was found between the
precuneus/posteriorcingulate cortex (PCC) and the left
lateral parietal cortex (LPC), and the abnormally
decreased coupling was found between the bilateral
anterior portion of dlPFC [26,38]. In the seed-based
analysis, the most common region of interest (ROI) is
the amygdala. The abnormally increased connectivity
was found between the amygdala and the prefrontal
cortex, the precentral gyrus (prCG), poCG, PCC, the
uncus,STG,thesuperioroccipitalgyrus(SOG),andthe
insula [31,34,43]. Moreover, abnormally decreased
connectivitywasfound between theamygdalaandIPL,
the middle frontal gyrus (MFG), the insula, the
cerebellumposteriorlobe(CBPL), the cerebellar tonsil,
IFG,thetemporal pole,andtheventromedial prefrontal
cortex(vmPFC)[31,35,43].Asforstudieswithmultiple
ROIs across brain networks, abnormally increased
connectivity within the reward network and the default
modenetwork(DMN)aswellasbetweenDMNandthe
executive control network was reported [32,43,48].
Meanwhile,abnormallydecreasedwithinDMNconnec-
tivity, within-network superior occipital and superior
temporalconnectivity,andbetween-networkhippocam-
palconnectivitywasreported[37,42,43,59].
Among these findings, though the increased and
decreasedconnectivitywithinorbetweennetworkswere

Table 3 Altered functional connectivity findings reported by R-fMRI studies on MDD
Study Method
Principalfindings
Findingsofalteredincreasing
connectivity
Findingsofaltereddecreasing
connectivity
Hermesdorfet al.2016[28] VMHC STG,insula,andprecuneus
Posneret al.2016[26] ICA Precuneus/PCCandleftLPC BilateralanteriorportionofdlPFC
Panet al.2017[32] Seed-basedanalysis:11ROIsinthe
valuationsystem
LeftVS
Yeet al.2017[31] Seed-basedanalysis:amygdala LeftamygdalawiththePFC,right
amygdalawiththeleftpoCG,leftPCC,
leftuncus,rightSTG,rightprCG,right
SOG,rightinsulaandrightuncus
LeftamygdalawiththeleftIPL,right
MFG,rightIPL,rightinsula,right
CBPLandrightCBT;rightamygdala
withtheleftIFG,leftMFG,left
temporalpoleandbilateralCBPL.
Lopezet al.2018[34] Seed-basedanalysis:amygdala,dlPFC dlFCwithdACC
Mehtaet al.2018[35] Seed-basedanalysis:amygdala RightamygdalaandvmPFC
(increasingplasmaC-reactiveprotein)
Tokudaet al.2018[37] Seed-basedanalysis:78ROIsacross
14brainnetworks
RightAGwithotherareaswithin
DMN
Tuet al.2018[38] ICA,PPI Positivemodulatoryinteractionsinthe
auditorynetwork
Negativemodulatoryinteractionsin
DMN
Shaet al.2018a[59] Modularityanalysis:WMD,PCof
nodesacross7networks
VN DMN,FPN
Wanget al.2019[39] Seed-basedanalysis:hypothalamus Bilateralhypothalamuswiththeright
insula,STG,IFG,andRolandic
operculum
Yaoet al.2019[42] Seed-basedanalysis:90ROIsacross
14brainnetworks
SOG,STG
Yanget al.2020[48] Functionalconnectivitystrength
analysis
LeftAG
Zhuet al.2020[44] Seed-basedanalysis:cuneus RightcuneustorightLTC(LES)
ChinFattet al.2020[43] Seed-basedanalysis:a100-brain-
regionparcellationandhippocampus,
VS,thalamus,andamygdala
parcellationsacross7brainnetworks
WithintheDMN,between-network
connectivityoftheDMNandECN
Between-networkhippocampal
connectivity
aMeta-analysis.Abbreviations:VMHC(voxel-mirroredhomotopicconnectivity),ICA(independentcomponentanalysis),ROI(regionofinterest),
PPI(physiophysiologicalinteraction)STG(superiortemporalgyrus),PCC (posteriorcingulatecortex),LPC(lateralparietalcortex), dlPFC
(dorsolateralprefrontal cortex),VS(ventralstriatum),poCG(postcentralgyrus), prCG( precentralgyrus), SOG(superioroccipitalgyrus),
IPL(inferiorparietallobule),MFG(topfrontalgyrus), CBPL(cerebellumposteriorlobe),CBT(cerebellartonsil), IFG(inferiorfrontalgyrus),
dACC(dorsalanteriorcingulatecortex),vmPFC(ventralmedialprefrontalcortex),AG(angulargyrus),DMN(defaultmodenetwork),VN(visual
network),FPN(frontoparietalnetwork),LTC(lateraltemporal cortex),ECN (executivecontrolnetworks);LSE(low sleepefficacy group);
WMD(within-moduledegree),PC(participantcoefficient).


SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 371

reported,DMN wasthemost involvedbrainnetwork in
MDD,suggestingnotonlythelimitedstaticalpowerbut
the complex neuropathobiology underlying the interac-
tionsofDMNandotherconfoundingfactors.Onerecent
studyinvestigatedtheDMNfunctionalconnectivityina
largesampleof1,300depressedpatientsand1,128HCs
and then found a significantly decreased functional
connectivitywithin DMNinrecurrent MDDvs.HCs as
well as recurrent MDD vs. first-episode drug-naïve
MDD patients. Furthermore, this effect was associated
with medication usage rather than MDD duration [66].
Moreover, the abnormal connectivity between the
amygdalaandtheinsula,the precuneus, PCC, IPL, and
CBPLwasreportedinmorethanonestudy.
Results from task-based fMRI studies
AlteredactivationsinlocalbrainregionsofMDD
Fourteen task-based fMRI studies reported abnormal
activationsinlocalbrainregions in patients with MDD
(Table4).Tasks in these studies can beroughly classi-
fiedintothreecategories:theemotionalprocessingtasks
(including the angry faces processing task and the
emotion regulation task), the reward learning and
valuationtasks(includingthe reward guessing task, the
monetary reward tasks, probabilistic selection task, the
probabilistic reward task, and other reward-related
tasks) and the cognitive tasks (including the working
memorytasksandtheTowerofLondonparadigm).
In the studies performing the emotional processing
tasks[21,25,40,45,50],abnormallyincreasedactivations
were found in the right ventrolateral prefrontal cortex
(vlPFC), OFC, the bilateral MTG, STG, and MFG in
dlPFC,theamygdala,thestriatum,theparahippocampal,
the cerebellar, the fusiform and ACC, and abnormally
decreased activation was found in the dACC. In the
studies performing the reward learning and valuation
tasks [27,33,47,51,58], abnormally increased activation

Table 4 Altered activations in local brain regions of MDD reported by task-based fMRI studies
Study Task
Principalfindings
Findingsofalteredincreasing
activation
Findingsofaltereddecreasing
activation
Demenescuet al.2011[21] Emotionalfacesprocessingtask dlPFC
vanTolet al.2011[22] TowerofLondonparadigm LeftdlPFC
Casementet al.2016[27] Rewardguessingtask dmPFC
Gollier-Briantet al.2016[25] Angryfacesprocessingtask RightvlPFC,OFC,MFGinthedlPFC
andinthebilateralMTGandSTG
Yükselet al.2017[30] Workingmemoryn-backtask
(0-back,2-backand3-back)
BilateralMOG,bilateralMFG,right
prCG,bilateralcerebellum,leftIPL
Admonet al.2017[33] Monetaryincentivedelaytask,
Probabilisticselectiontask
Striatum
Fitzgeraldet al.2019[40] Block-designreappraisal-based
Emotionregulationtask
dACC
Rupprechter et al.2020[47] Probabilisticrewardlearningtask NAcc
Hillandet al.2020[45] Emotionregulationtask ACCandamygdala(MDDwithout
ABMtraining)
Groenewoldet al.2013a[50] Emotionalprocessingtasks Amygdala,striatum,parahippocampal,
cerebellar,fusiformandACC
(negativestimuli)
Amygdala,striatum,parahippocampal,
cerebellar,fusiformandACC
(positivestimuli)
Grahamet al.2013a[49] Emotional,cognitiveandothertasks BilateralMTG,leftIFC,leftsgACC,
leftprCG,leftthalamus,leftMFG;
RightMFG,rightparahippocampus,
leftIFC,bilateralcaudate,rightSTG,
MTG,rightaACC,rightinsula,right
amygdalaandleftoccipitalregions
Zhanget al.2013a[51] Moneyrewardtasksandemotion
processingtasks
MFGanddACC Caudate
Wang et al.2015a[53] Workingmemorytasks LeftIFCandMFC,leftprCG,left
insula,rightSTGandrightSG
RightprCG,rightprecuneusandright
insula
Keren et al.2018a[58] Reward-relatedtasks Caudate,putamenandglobuspallidus
aMeta-analysis.Abbreviations:MDD(majordepressivedisorder),ABM(attentionalbiasmodification),dlPFC(dorsolateral prefrontalcortex),
dmPFC(dorsalmedialprefrontalcortex),vlPFC(ventrallateralprefrontalcortex),OFC(orbitofrontalcortex),MFG(topfrontalgyrus),MTG(top
temporalgyrus),STG(superiortemporalgyrus),MOG(topoccipitalgyri),prCG(precentralgyrus),IPL(inferiorparietallobule),ACC(anterior
cingulatecortex),dACC(dorsalanteriorcingulatecortex),sgACC(subgenualanteriorcingulate),NAcc(nucleusaccumbens),IFC(inferiorfrontal
cortex),poCG(postcentralgyrus),SG(supramarginalgyrus).


Xue-YingLietal.

372 ©TheAuthors(2022).PublishedbyHigherEducationPress

was found in the dorsal medial prefrontal cortex
(dmPFC) and MFG. Abnormally decreased activations
were found in the striatum and the nucleus accumbens
(NAcc), the caudate, the putamen, and the globus
pallidus. In the studies performing the cognitive tasks
[22,30,53], abnormally increased activation was found
intheleftdlPFC, and abnormally decreased activations
were found in the middle occipital gyri (MOG), MFG,
the right prCG, the cerebellum, and the left IPL.
Moreover, a meta-analysis [53] reviewed the altered
brain responses to working memory loads in MDD
patients vs. HCs and reported the increased activations
intheleftIFGand MFG, the left prCG, the left insula,
the right STG, and the right supramarginal gyrus (SG)
andthedecreasedactivationsintherightprCG,theright
precuneus,andtherightinsula.
In sum, studies on working memory generally repor-
tedthedecreasedactivationintherightprCG.Withboth
cognitive tasks and emotional processing tasks,
researchersgenerally foundalteredactivationsindlPFC
andSTG.
AlteredfunctionalconnectivityduringtasksinMDD
Five task-based fMRI studies reported the altered
functionalconnectivityinpatients withMDD(Table5).
Tasks performed in the five studies can be roughly
classified into four categories, three of which are the
same as the above-mentioned tasks: the emotion
processingtasks(including the emotionregulationtask,
consciousandnon-consciousemotionalfacesprocessing
tasks), the reward learning and valuation tasks (inclu-
ding the monetary incentive delay task, probabilistic
selection task, and the probabilistic reward learning
task), and the cognitive tasks (including the external
attentiontask,theauditoryoddballtask, the continuous
performance task, and the Go-No Go task). Moreover,
the self-appraisal task is classified into the fourth
categoryasself-perceptionandself-understanding.
In studies performing the emotion processing tasks,
decreasedfunctionalconnectivitybetweentheamygdala
andvlPFCinpatientswithMDDwasreported[40].In
studies with the reward learning and valuation tasks,
decreased connectivity between the prefrontal cortex
andtheventralstriatum(VS) as well as between NAcc
and the midcingulate cortex (MCC) was reported
[33,47].InthestudyofKorgaonkaret al.[46],cognitive
andemotion processing abilities were assessed through
5 fMRI tasks, and decreased DMN–FPN connectivity
was observed in MDD non-remitters. And in the study
with self-appraisal task, the negative modulatory effect
of the medial prefrontal cortex (mPFC) on IPL was
reported[29].
DISCUSSION
Due to the lack of consensus and reproducibility in
fMRIstudiesonMDD,agrowingbodyofliteraturehas
highlighted the methodological issues in neuroimaging
research. According to these studies, we screened
previousfMRI studiesonMDDusingcriteriaincluding
sample size, preprocessing pipelines, and multiple
comparison correction strategies. Contrary to our
assumptions, after screening the previous studies with
the above-mentioned criteria, both convergent and
contradictedresults werereported.Here,we focusedon
those convergent findings and discussed some implica-
tionsaccordingly.

Table 5 Altered functional connectivity during tasks in MDD
Study Task
Principalfindings
Findingsofalteredincreasing
connectivity
Findingsofaltereddecreasing
connectivity
Daveyet al.2017[29] Self-appraisaltask,externalattentiontask MPFCnegativelymodulatesIPL
Fitzgeraldet al.2019[40] Block-designreappraisal-basedEmotion
RegulationTask
AmygdalawithvlPFC
Korgaonkaret al.2020[46] iSPOT-Dstudyprotocolwith5fMRItasks:
auditoryoddballtask,continuous
performancetask,Go-NoGotask,conscious
andnon-consciousemotionalfaces
processingtasks
BetweenDMNandFPN
Rupprechteret al.2020[47] Probabilisticrewardlearningtask PFCwithVS
Admonet al.2017a[33] Monetaryincentivedelaytask;Probabilistic
selectiontask
NACcandmidcingulatecortex
aMeta-analysis.Abbreviations:iSPOT-D(InternationalStudytoPredictOptimizedTreatmentforDepression),mPFC(medialprefrontalcortex),
IPL(inferior parietallobule), vlPFC(ventral lateralprefrontal cortex),DMN (defaultmodenetwork),FPN(frontoparietalnetwork),PFC
(prefrontalcortex),VS(ventralstriatum),NAcc(nucleusaccumbens),MCC(midcingulatecortex).


SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 373

Dysfunctions of brain regions and networks in
MDD
Evidence from the R-fMRI studies reviewed here
suggests that the alterations in DMN (especially in
ACC, PCC, and the precuneus) and its couplings with
otherbrainnetworksmay play an important role in the
pathologyofMDD.AsakeynodeofDMN,convergent
dysfunction of the precuneus regarding abnormal
regional activities and functional connectivity [26] in
MDDpatientsversusHCsinresting-statewerereported.
The precuneus dysfunction has been found in several
different mental disorders, even in migraines [67,68].
Thus the dysfunction of the precuneus may be a
generalizedfunctionalmarkeracrossmentaldisorders.It
has been proposed that DMN underlies the self-
referentialprocessandthenegativerumination inMDD
[69–72]. It is also found that the hyper-connectivity
within DMN may predict the better outcomes of
sertralinetreatment[43],which was consistent with the
recent study reporting the decreased functional
connectivity within DMN in recurrent MDD patients
[66]. To sum up, these studies implicated that DMN
may be a potential biomarker of MDD and a possible
targetforfutureMDDtreatment.
We found that task-based fMRI studies get some
convergent results about cognitive and emotional
processing.Specifically,the hyperactivityindlPFCand
alteredfunctional connectivitybetweenDMNandFPN,
especially dlPFC, were reported in studies applying
paradigms with both cognitive or emotional stimuli.
Thesefindingsalignwithpreviousstudiesindicatingthe
recruitment of dlPFC in emotion regulation [73].
Specifically, dlPFC has been repetitively reported for
blunted activity in R-fMRI studies and proved as an
effectivetargetfortheTMStreatmentofMDD[74].Of
note,dlPFCwasakeynodeoftheFPN,whichunderlies
executive and control functions [75]. FPN does not
generate emotions directly but may underlie the reap-
praisal and reactive processes regarding emotions.
Emotion regulation is identified as a process with
conscious or non-conscious strategies to change the
initial emotional reaction, especially negative emotions
[76]. Regulating strategies can impact the generation
andreactivitytoemotionsat different time points [77].
Thus, the causal relationship between cognition and
emotionyieldsa confounding pattern askingforfurther
examinations.
Caveats for future fMRI studies on MDD
We also observed inconsistency among these results.
Forexample, bothdecreasedandincreasedspontaneous
activities in the PHG and ACC were reported
[36,55,57].Andbothincreasedanddecreasedfunctional
connectivity between the amygdala and the insula, as
well as within- and between-network connectivity
regarding DMN was reported [28,31,43,78]. Issues
including the lack of integrative assessments and
interpretations of MDD, differences in experimental
design, data acquisition and preprocessing procedures,
the mixture of the heterogeneous participant popula-
tions, and inappropriate methodologies to depict the
anatomical and functional brain may lead to such
inconsistencyinR-fMRIstudies.
It is not surprising that task-based fMRI studies
yielded such inconsistent results, given the differences
in the psychological processes they investigated. In
previous studies, these processes were commonly
investigated by three sets of paradigms, which corres-
pond to three functional systems according to the
researchdomaincriteria(RDoC)[79,80]framework:the
emotion processing tasks, the reward learning and
valuation tasks and the cognitive tasks. The cognitive
tasksmainlyassess the working memory and attention,
whichcorrespondstothecognitivesystem.Thereward
learning and valuation tasks mainly assess the learning
and valuating ability to rewards, which corresponds to
the positive valence system. The emotion processing
tasksmainlyassessedtheresponsesto threats and loss,
whicharecloselyrelatedtothenegativevalencesystem
[81]. Note that these psychological functions were
investigated separately in most previous MDD studies
lacking integrative assessments and interpretations.
However, as a disorde characterized by a constellation
ofbehavioral,emotional,and cognitive symptoms [82],
MDDispositedtoinvolvedysfunctionsinmanyaspects
of cognitive and emotional processes including inhibi-
toryprocesses,deficitsinworkingmemory, rumination
and reappraisal [83]. This suggests that applying a
carefully selected and comprehensive set of neuro-
behavioral measurements that covers multiple psycho-
logical functions may shed new light on depression
research.
Furthermore, the lack of convergent results across
previousfMRI studies investigating the same cognitive
or emotional function indicates that differences in
experimental design and data acquisition procedures
mayalso contributetoinconsistency[84].Forexample,
in task-based studies included in the current review,
workingmemorywastestedbydifferentversionsofn-
backtasks(e.g.,continuousperformancetask,orn-back
tasks that consist of conditions of 0-back, 2-back, and
3-back),mentalarithmetictasks,TowerofLondontasks
andother paradigms. In these experiments, participants
were instructed to respond to different stimuli, such as

Xue-YingLietal.

374 ©TheAuthors(2022).PublishedbyHigherEducationPress

letters or numbers [30,46,53]. Even in R-fMRI studies,
theMRIdatawere acquired on different scanners,with
different parameters and prompted for subjects (eyes
open or closed), and through different scan durations,
followed by various pre-processing workflows
[28,32,36,44].Theabsence of a “goldstandard”forthe
data acquisition in R-fMRI studies may lead to the
recentreplicabilityandreproducibilitycrisesas well as
difficulties in interpreting this inconsistency [85,86].
Some researchers have raised the discussion of
replicabilityissues andcalledfordisciplines toadvance
research transparency and open science [87,88]. In
practice,wehaveinitiatedtheREST-meta-MDDProject
withastandardizedMRIdatasharingandpreprocessing
protocol based on data processing assistant for resting-
state fMRI (DPARSF) [89] and achieved preliminary
successinopendatasharingandcollaborative research
[66].
Oneotherpossiblecontributiontotheinconsistencyof
previousfMRIstudies on MDD is the heterogeneity of
theinvestigatedpopulations[84]regardingmedications,
treatment outcomes, onset ages, and subtypes (e.g.,
melancholic vs. atypical MDD) [90]. Therefore, future
studies on MDD should carefully divide the MDD
samplesintosubgroupsaccordingtothese confounding
factors so that a clearer understanding of the relation-
shipsbetweenrepresentationsoffMRIbrainalternations
andMDDcanbeobtained.
Finally, the functional systems of the human brain
havefeaturesofanintricatenetworkwithmultipletem-
poral and spatial levels, which are largely distributed/
embeddedontheintrinsic two-dimensional structure of
thecorticalsurface[91,92].Thenetworkneuroscientific
approach provided efficient new ways to map, analyze
and model the elements and interactions of neurobio-
logicalsystemsasagraph[93].There have been fMRI
studies [94,95] using the network and graph theory
analysisthatshows somealteredfunctionallocations in
linewiththevoxel-wisemetrics,includingReho,ALFF,
and fALFF and functional connectivity strength find-
ings. For example, Long and his colleagues find local
changes in the default-mode, sensorimotor and subcor-
tical areas using a novel dynamic network-based
approach [94]. What’s more, in order to properly
understandthe brain functional systems, it isnecessary
to obtain an accurate and explicit representation of the
corticalsurfaceconsidering its topology of a 2-D sheet
andahighly folded geometry. The surface-based fMRI
approachisaprincipledwaytoachievethisgoal,which
was reported to be nearly three times better than the
traditional volume-based approaches in special
localization of cortical areas [96]. However, most
previousstudiesarestillbasedontraditional voxel-wise
metrics (e.g., Reho, ALFF, and fALFF) and functional
connectivity. Future studies could leverage a surface-
based network neuroscientific approach to advance
understanding the brain dysfunctions in depressions
fromamoreintegrativeperspective.
Limitations
Despite the strict criteria we applied, controversy still
existed in the remaining studies. Due to the limited
numberofpapersincludedinthepresentreview,wedid
not further group studies according to factors such as
age, race, severity of disease, or medication usage.
Moreover,wenote thatgraphtheorymetrics candepict
thebrainasacomplex networked system and be worth
considering. However, findings from graph theory
studies may need to be interpreted in a more
sophisticatedframeworkandthusareoutofthescopeof
thecurrentreview.
CONCLUSION
In this systematic review, we found that MDD was
consistently characterized by abnormalities within the
DMN and the FPN, as well as altered connectivity
between them and other brain networks. However,
highly inconsistent results remained, probably due to
issuesincludingthe lack of integrativeassessmentsand
interpretations of MDD, differences in experimental
design and data acquisition procedures, the mixture of
the heterogeneous participant populations, and the
relatively inappropriate methodologies to depict the
anatomicalandfunctionalbrain.Apartfromasufficient
sample size, adequate head motion artifact correction,
and multiple comparison correction strategies, future
studies are recommended to perform a comprehensive
set of neuro-behavioral measurements, consider the
heterogeneity of MDD patients and other potentially
confoundingfactors,applysurface-basedneuroscientific
network fMRI approaches and advance research
transparencyandopen sciencebymovementsincluding
developing state-of-the-art pipeline with open data
sharing.
MATERIALS AND METHODS
Literature search strategy
Studiesthatare electronically published untilJune10th,
2020, were searched in PubMed (Fig.1). The search
terms are “((English [Language]) AND ((MRI
[Title/Abstract]) OR (fMRI [Title/Abstract]))) AND
((depression [Title/Abstract]) OR (depressive disorder
[Title/Abstract]))”. The filter of PubMed was used to
constrainthearticletypesofthesearchingresults,which

SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 375

could exclude the nonscientific papers, such as news,
books,and documents. Moreover, the reference listsof
theincludedarticleswerealsoscreened.
Inclusion and exclusion criteria
We reviewed the titles and abstracts to exclude studies
thatarenot fMRIstudiesonMDD, systematic reviews,
and commentary articles. Then the following exclusion
criteriawereapplied:(i)publicationsthathavenotbeen
peer-reviewed; (ii) studies regarding other mental
diseases, such as post-traumatic stress disorder (PTSD)
or bipolar disorders. (iii) studies based solely on the
structural MRI or diffusion MRI. (iv) studies reporting
no fMRI findings; (v) studies focusing on topics other
thanthehumanbrain,suchasgeneticsorgut-brainaxis.
To ensure the quality of the included studies, we
furtherappliedadditional methodological criteria: (i) at
least 40 participants per group; (ii) performed proper
head motion artifact correction; (iii) with proper
correctionformultiplecomparisons,e.g.,FWE-basedor
FDR-based.Exceptforthepermutationtest withTFCE,
theacceptedthresholds forotherFWE-basedcorrection
arevoxel-wiseP<0.001withcluster-wiseP<0.05,and
forFDR-basedcorrectionq<0.05.
ACKNOWLEDGMENTS
This work was supported by the National Key R&D Program of China
(2017YFC1309902 to CY), the National Natural Science Foundation of
China (81671774, 81630031 to CY), the 13th Five-year Informatization
Plan of Chinese Academy of Sciences (XXH13505 to CY), the Key
Research Program of the Chinese Academy of Sciences (ZDBS-SSW-
JSC006 to CY), Beijing Nova Program of Science and Technology
(Z191100001119104 to CY), Scientific Foundation of Institute of
Psychology,Chinese Academyof Sciences(Y9CX422005 toXC), China
PostdoctoralScienceFoundation (2019M660847 toXC), China National
PostdoctoralProgramforInnovativeTalents(BX20200360toXC).
COMPLIANCE WITH ETHICS GUIDELINES
TheauthorsXue-YingLi,XiaoChenandChao-GanYandeclarethatthey
havenoconflictofinterest.
Thisarticledoesnotcontainanystudieswithhumanoranimalmaterials
performedbyanyoftheauthors.
OPEN ACCESS
This article is licensed by the CC By under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons licence, and indicate if
changesweremade.Theimagesorotherthirdpartymaterialinthisarticle
areincludedinthe article’s Creative Commons licence, unless indicated
otherwiseinacreditlinetothematerial. Ifmaterialis notincludedinthe
article’sCreativeCommonslicenceandyourintendeduseisnotpermitted
by statutory regulation or exceeds the permitted use, you will need to
obtainpermissiondirectly from the copyrightholder.To view a copyof
thislicence,visithttp://creativecommons.org/licenses/by/4.0/.
REFERENCES
Friedrich, M. J. (2017) Depression is the leading cause of
disabilityaroundtheworld.JAMA,317,1517
1.
Vos, T., Lim, S. S., Abbafati, C., Abbas, K. M., Abbasi, M.,
Abbasifard, M., Abbasi-Kangevari, M., Abbastabar, H., Abd-
Allah, F., Abdelalim, A., et al. (2020) Global burden of 369
diseasesandinjuriesin204 countriesandterritories,1990‒2019:
a systematic analysis for the Global Burden of Disease Study
2019.Lancet,396,1204−1222
2.


Figure 1. Flowchart of literature screening.


Xue-YingLietal.

376 ©TheAuthors(2022).PublishedbyHigherEducationPress

Pérez-Stable, E. J., Miranda, J., Muñoz, R. F. and Ying, Y. W.
(1990)Depressioninmedical outpatients. Underrecognition and
misdiagnosis.Arch.Intern.Med.,150,1083−1088
3.
Bandettini, P. A. (2012) Twenty years of functional MRI: the
scienceandthestories.Neuroimage,62,575−588
4.
Bijsterbosch, J., Harrison, S. J., Jbabdi, S., Woolrich, M.,
Beckmann,C.,Smith, S. and Duff,E.P. (2020) Challengesand
future directions for representations of functional brain
organization.Nat.Neurosci.,23,1484−1495
5.
Ioannidis,J.P.(2005)Why mostpublished researchfindingsare
false.PLoSMed.,2,e124
6.
Button,K. S., Ioannidis,J.P., Mokrysz,C.,Nosek, B.A.,Flint,
J.,Robinson,E.S.andMunafò,M.R.(2013)Powerfailure:why
smallsamplesizeunderminesthereliabilityofneuroscience.Nat.
Rev.Neurosci.,14,365−376
7.
Zuo, X. N., Xu, T. and Milham, M. P. (2019) Harnessing
reliability for neuroscience research. Nat. Hum. Behav., 3,
768−771
8.
Ciric,R.,Wolf, D. H.,Power,J. D., Roalf, D.R.,Baum, G. L.,
Ruparel, K., Shinohara, R. T., Elliott, M. A., Eickhoff, S. B.,
Davatzikos, C., et al. (2017) Benchmarking of participant-level
confound regression strategies for the control of motion artifact
instudiesoffunctionalconnectivity.Neuroimage,154,174−187
9.
Yan, C.-G., Cheung, B., Kelly, C., Colcombe, S., Craddock, R.
C., Di Martino, A., Li, Q., Zuo, X. N., Castellanos, F. X. and
Milham, M. P. (2013) A comprehensive assessment of regional
variation in the impact of head micromovements on functional
connectomics.Neuroimage,76,183−201
10.
Ciric,R., Rosen, A.F. G., Erus,G., Cieslak, M.,Adebimpe,A.,
Cook, P. A., Bassett, D. S., Davatzikos, C., Wolf, D. H. and
Satterthwaite, T. D. (2018) Mitigating head motion artifact in
functionalconnectivityMRI.Nat.Protoc.,13,2801−2826
11.
Eklund, A., Nichols, T. E. and Knutsson, H. (2016) Cluster
failure: Why fMRI inferences for spatial extent have inflated
false-positiverates.Proc.Natl.Acad.Sci.USA,113,7900−7905
12.
Algermissen,J. and Mehler,D. M.A.(2018) Maythepower be
withyou:are there highly powered studiesinneuroscience, and
howcanwegetmoreofthem?JNeurophysiol.,119,2114−2117
13.
Chen, X., Lu, B. and Yan, C. G. (2018) Reproducibility of R-
fMRI metrics on the impact of different strategies for multiple
comparisoncorrection and samplesizes. Hum. BrainMapp.,39,
300−318
14.
VanDijk, K.R., Sabuncu,M. R.and Buckner,R. L.(2012) The
influence of head motion on intrinsic functional connectivity
MRI.Neuroimage,59,431−438
15.
Power,J. D.,Barnes,K. A.,Snyder, A.Z.,Schlaggar, B.L. and
Petersen, S. E. (2012) Spurious but systematic correlations in
functionalconnectivityMRInetworksarise fromsubjectmotion.
Neuroimage,59,2142−2154
16.
Eklund, A., Nichols, T. E. and Knutsson, H. (2016) Cluster
failure: Why fMRI inferences for spatial extent have inflated
false-positiverates.Proc.Natl.Acad.Sci.USA,113,7900−7905
17.
Nichols,T.andHayasaka, S. (2003) Controlling the familywise
errorrateinfunctionalneuroimaging:acomparativereview.Stat.
18.
MethodsMed.Res.,12,419−446
Genovese, C. R., Lazar, N. A. and Nichols, T. (2002)
Thresholdingofstatisticalmapsinfunctionalneuroimagingusing
thefalsediscoveryrate.Neuroimage,15,870−878
19.
Grimes,D.A.andSchulz,K.F.(1996) Determiningsamplesize
and power in clinical trials: the forgotten essential. Semin.
Reprod.Endocrinol.,14,125−131
20.
Demenescu, L. R., Renken, R., Kortekaas, R., van Tol, M. J.,
Marsman, J. B., van Buchem, M. A., van der Wee, N. J.,
Veltman, D. J., den Boer, J. A. and Aleman, A. (2011) Neural
correlates of perception of emotional facial expressions in out-
patients with mild-to-moderate depression and anxiety. A
multicenterfMRIstudy.Psychol.Med.,41,2253−2264
21.
vanTol,M.J., vanderWee,N.J.,Demenescu,L.R.,Nielen,M.
M.,Aleman, A.,Renken,R., vanBuchem,M. A.,Zitman, F. G.
and Veltman, D. J. (2011) Functional MRI correlates of
visuospatialplanning inout-patient depressionandanxiety.Acta
Psychiatr.Scand.,124,273−284
22.
Bermingham,R.,Carballedo,A.,Lisiecka,D.,Fagan,A.,Morris,
D.,Fahey,C.,Meaney, J.,Gill,M.andFrodl,T.(2012)Effectof
genetic variant in BICC1 on functional and structural brain
changesindepression.Neuropsychopharmacology,37,285−2862
23.
Yang,X.,Ma,X.,Li,M.,Liu,Y.,Zhang,J.,Huang,B.,Zhao,L.,
Deng, W., Li, T. and Ma, X. (2015) Anatomical and functional
brain abnormalities in unmedicated major depressive disorder.
Neuropsychiatr.Dis.Treat.,11,2415−2423
24.
Gollier-Briant, F., Paillère-Martinot, M. L., Lemaitre, H.,
Miranda, R., Vulser, H., Goodman, R., Penttilä, J., Struve, M.,
Fadai, T., Kappel, V., et al. (2016) Neural correlates of three
types of negative life events during angry face processing in
adolescents.Soc.Cogn.Affect.Neurosci.,11,1961−1969
25.
Posner,J., Cha,J., Wang,Z.,Talati, A.,Warner, V.,Gerber,A.,
Peterson, B. and Weissman, M. (2016) Increased default mode
network connectivity in individuals at high familial risk for
depression.Neuropsychopharmacology,41,1759−1767
26.
Casement, M. D., Keenan, K. E., Hipwell, A. E., Guyer, A. E.
andForbes,E. E.(2016)Neural reward processingmediatesthe
relationship between insomnia symptoms and depression in
adolescence.Sleep,39,439−447
27.
Hermesdorf, M., Sundermann, B., Feder, S., Schwindt, W.,
Minnerup,J.,Arolt,V.,Berger,K.,Pfleiderer,B.andWersching,
H. (2016) Major depressive disorder: Findings of reduced
homotopicconnectivityandinvestigationofunderlyingstructural
mechanisms.Hum.BrainMapp.,37,1209−1217
28.
Davey,C.G.,Breakspear,M.,Pujol,J.andHarrison,B.J.(2017)
A brain model of disturbed self-appraisal in depression. Am. J.
Psychiatry,174,895−903
29.
Yüksel,D.,Dietsche, B., Forstner, A. J.,Witt,S.H., Maier, R.,
Rietschel, M., Konrad, C., Nöthen, M. M., Dannlowski, U.,
Baune,B. T.,et al. (2017)Polygenic riskfordepression andthe
neural correlates of working memory in healthy subjects. Prog.
Neuropsychopharmacol.Biol.Psychiatry,79,67−76
30.
Ye,J.,Shen,Z.,Xu,X.,Yang,S.,Chen,W.,Liu,X.,Lu,Y.,Liu,
F.,Lu, J.,Li,N., et al.(2017) Abnormalfunctional connectivity
31.

SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 377

of the amygdala in first-episode and untreated adult major
depressive disorder patients with different ages of onset.
Neuroreport,28,214−221
Pan,P.M.,Sato,J. R.,Salum,G.A.,Rohde,L.A.,Gadelha, A.,
Zugman,A., Mari,J., Jackowski,A.,Picon,F.,Miguel,E.C.,et
al.(2017) Ventralstriatum functionalconnectivityas apredictor
of adolescent depressive disorder in a longitudinal community-
basedsample.Am.J.Psychiatry,174,1112−1119
32.
Admon, R., Kaiser, R. H., Dillon, D. G., Beltzer, M., Goer, F.,
Olson, D. P., Vitaliano, G. and Pizzagalli, D. A. (2017)
Dopaminergic enhancement of striatal response to reward in
majordepression.Am.J.Psychiatry,174,378−386
33.
Lopez,K.C.,Luby,J.L.,Belden,A.C.andBarch,D.M.(2018)
Emotion dysregulation and functional connectivity in children
with and without a history of major depressive disorder. Cogn.
Affect.Behav.Neurosci.,18,232−248
34.
Mehta, N. D., Haroon, E., Xu, X., Woolwine, B. J., Li, Z. and
Felger, J. C. (2018) Inflammation negatively correlates with
amygdala-ventromedial prefrontal functional connectivity in
associationwith anxiety inpatients withdepression:Preliminary
results.BrainBehav.Immun.,73,725−730
35.
Qi, S., Yang, X., Zhao, L., Calhoun, V. D., Perrone-Bizzozero,
N., Liu, S., Jiang, R., Jiang, T., Sui, J. and Ma, X. (2018)
MicroRNA132 associated multimodal neuroimaging patterns in
unmedicatedmajordepressivedisorder.Brain,141,916−926
36.
Tokuda, T., Yoshimoto, J., Shimizu, Y., Okada, G., Takamura,
M., Okamoto, Y., Yamawaki, S. and Doya, K. (2018)
Identification of depression subtypes and relevant brain regions
usingadata-drivenapproach.Sci.Rep.,8,14082
37.
Tu,Z.,Jia,Y.Y.,Wang,T.,Qu,H.,Pan,J.X.,Jie,J.,Xu,X.Y.,
Wang, H. Y. and Xie, P. (2018) Modulatory interactions of
resting-state brain functional connectivity in major depressive
disorder.Neuropsychiatr.Dis.Treat.,14,2461−2472
38.
Wang,D.,Xue, S. W.,Tan,Z., Wang, Y.,Lian,Z. and Sun,Y.
(2019) Altered hypothalamic functional connectivity patterns in
majordepressivedisorder.Neuroreport,30,1115−1120
39.
Fitzgerald,J.M.,Klumpp, H., Langenecker, S. and Phan, K.L.
(2019) Transdiagnostic neural correlates of volitional emotion
regulation in anxiety and depression. Depress. Anxiety, 36,
453−464
40.
Xia, M., Si, T., Sun, X., Ma, Q., Liu, B., Wang, L., Meng, J.,
Chang,M.,Huang,X.,Chen, Z.,et al.(2019)Reproducibilityof
functional brain alterations in major depressive disorder:
Evidencefromamultisiteresting-statefunctionalMRIstudywith
1,434individuals.Neuroimage,189,700−714
41.
Yao,Z.,Zou,Y., Zheng, W., Zhang, Z.,Li,Y.,Yu, Y., Zhang,
Z.,Fu, Y.,Shi, J.,Zhang,W., et al. (2019)Structuralalterations
of the brain preceded functional alterations in major depressive
disorder patients: Evidence from multimodal connectivity. J.
Affect.Disord.,253,107−117
42.
ChinFatt,C.R.,Jha,M.K.,Cooper,C.M.,Fonzo,G.,South,C.,
Grannemann, B., Carmody, T., Greer, T. L., Kurian, B., Fava,
M., et al. (2020) Effect of intrinsic patterns of functional brain
connectivity in moderating antidepressant treatment response in
43.
majordepression.Am.J.Psychiatry,177,143−154
Zhu,D.M.,Zhang,C., Yang, Y., Zhang, Y., Zhao, W., Zhang,
B., Zhu, J. and Yu, Y. (2020) The relationship between sleep
efficiencyandclinicalsymptomsismediatedbybrainfunctionin
majordepressivedisorder.J.Affect.Disord.,266,327−337
44.
Hilland, E., Landrø, N. I., Harmer, C. J., Browning, M.,
Maglanoc, L. A. and Jonassen, R. (2020) Attentional bias
modification is associated with fMRI response toward negative
stimuli in individuals with residual depression: a randomized
controlledtrial.J.PsychiatryNeurosci.,45,23−33
45.
Korgaonkar, M. S., Goldstein-Piekarski, A. N., Fornito, A. and
Williams, L. M. (2020) Intrinsic connectomes are a predictive
biomarker of remission in major depressive disorder. Mol.
Psychiatry,25,1537−1549
46.
Rupprechter, S., Romaniuk, L., Series, P., Hirose, Y., Hawkins,
E.,Sandu, A. L.,Waiter, G. D.,McNeil, C.J.,Shen, X.,Harris,
M. A., et al. (2020) Blunted medial prefrontal cortico-limbic
reward-relatedeffective connectivityanddepression. Brain,143,
1946−1956
47.
Yang,Y.,Zhu,D. M., Zhang, C., Zhang, Y.,Wang,C.,Zhang,
B., Zhao, W., Zhu, J. and Yu, Y. (2020) Brain structural and
functional alterations specific to low sleep efficiency in major
depressivedisorder.Front.Neurosci.,14,50
48.
Graham, J., Salimi-Khorshidi, G., Hagan, C., Walsh, N.,
Goodyer, I., Lennox, B. and Suckling, J. (2013) Meta-analytic
evidenceforneuroimagingmodelsofdepression:stateortrait?J.
Affect.Disord.,151,423–431
49.
Groenewold,N.A.,Opmeer,E.M.,deJonge,P.,Aleman,A.and
Costafreda, S. G. (2013) Emotional valence modulates brain
functional abnormalities in depression: evidence from a meta-
analysisoffMRIstudies.Neurosci.Biobehav.Rev.,37,152−163
50.
Zhang,W.N.,Chang,S.H.,Guo,L.Y.,Zhang,K.L.andWang,
J. (2013) The neural correlates of reward-related processing in
majordepressivedisorder:ameta-analysisoffunctionalmagnetic
resonanceimagingstudies.J.Affect.Disord.,151,531−539
51.
Iwabuchi,S.J.,Krishnadas, R.,Li,C.,Auer,D.P.,Radua,J.and
Palaniyappan, L. (2015) Localized connectivity in depression: a
meta-analysis of resting state functional imaging studies.
Neurosci.Biobehav.Rev.,51,77−86
52.
Wang,X.L.,Du,M.Y.,Chen,T.L.,Chen,Z.Q.,Huang,X.Q.,
Luo,Y.,Zhao, Y.J.,Kumar, P. andGong,Q. Y. (2015)Neural
correlates during working memory processing in major
depressive disorder. Prog. Neuropsychopharmacol. Biol.
Psychiatry,56,101−108
53.
Zhong, X., Pu, W. and Yao, S. (2016) Functional alterations of
fronto-limbic circuit and default mode network systems in first-
episode, drug-naïve patients with major depressive disorder: A
meta-analysisof resting-state fMRIdata. J.Affect.Disord., 206,
280−286
54.
Wang, W., Zhao, Y., Hu, X., Huang, X., Kuang, W., Lui, S.,
Kemp, G. J. and Gong, Q. (2017) Conjoint and dissociated
structuralandfunctionalabnormalitiesinfirst-episodedrug-naive
patients with major depressive disorder: a multimodal meta-
analysis.Sci.Rep.,7,10401
55.

Xue-YingLietal.

378 ©TheAuthors(2022).PublishedbyHigherEducationPress

Kambeitz,J., Cabral,C., Sacchet,M.D.,Gotlib,I. H.,Zahn, R.,
Serpa,M. H.,Walter, M.,Falkai, P.andKoutsouleris,N.(2017)
Detecting neuroimaging biomarkers for depression: A meta-
analysis of multivariate pattern recognition studies. Biol.
Psychiatry,82,330−338
56.
Zhou, M., Hu, X., Lu, L., Zhang, L., Chen, L., Gong, Q. and
Huang, X. (2017) Intrinsic cerebral activity at resting state in
adults with major depressive disorder: A meta-analysis. Prog.
Neuropsychopharmacol.Biol.Psychiatry,75,157−164
57.
Keren, H., O’Callaghan, G., Vidal-Ribas, P., Buzzell, G. A.,
Brotman,M.A., Leibenluft, E., Pan, P. M.,Meffert,L.,Kaiser,
A.,Wolke, S., et al.(2018) Rewardprocessingin depression: A
conceptual and meta-analytic review across fMRI and EEG
Studies.Am.J.Psychiatry,175,1111−1120
58.
Sha,Z.,Xia, M., Lin, Q., Cao, M.,Tang,Y., Xu, K., Song, H.,
Wang, Z., Wang, F., Fox, P.T., et al. (2018) Meta-connectomic
analysis reveals commonly disrupted functional architectures in
network modules and connectors across brain disorders. Cereb.
Cortex,28,4179−4194
59.
Wu, J. C., Buchsbaum, M. S., Johnson, J. C., Hershey, T. G.,
Wagner, E. A., Tung, C. and Lottenberg, S. (1993) Magnetic
resonance and positron emission tomography imaging of the
corpus callosum: size, shape and metabolic rate in unipolar
depression.J.Affect.Disord.,28,15−25
60.
Biswal,B.,Yetkin,F.Z.,Haughton,V.M.andHyde,J.S.(1995)
Functional connectivity in the motor cortex of resting human
brainusingecho-planarMRI.Magn.Reson.Med.,34,537−541
61.
Friston,K.J.,Ashburner,J.,Frith,C.D.,Poline,J.-B.,Heather,J.
D. and Frackowiak, R. S. J. (1995) Spatial registration and
normalizationofimages.Hum.BrainMapp.,3,165−189
62.
Yan, C. G., Wang, X. D., Zuo, X. N. and Zang, Y. F. (2016)
DPABI: data processing & analysis for (resting-state) brain
imaging.Neuroinformatics,14,339−351
63.
Jenkinson, M., Bannister, P., Brady, M. and Smith, S. (2002)
Improved optimization for the robust and accurate linear
registrationand motion correctionofbrain images. Neuroimage,
17,825−841
64.
Cox,R.W.(1996)AFNI: softwarefor analysisandvisualization
offunctionalmagneticresonanceneuroimages.Comput.Biomed.
Res.,29,162−173
65.
Yan,C.G.,Chen,X.,Li,L.,Castellanos,F.X.,Bai,T.J.,Bo,Q.
J., Cao, J., Chen, G. M., Chen, N. X., Chen, W., et al. (2019)
Reduced default mode network functional connectivity in
patients with recurrent major depressive disorder. Proc. Natl.
Acad.Sci.USA,116,9078−9083
66.
Castellanos, F. X., Margulies, D. S., Kelly, C., Uddin, L. Q.,
Ghaffari,M., Kirsch,A., Shaw,D., Shehzad,Z.,DiMartino,A.,
Biswal,B., et al.(2008)Cingulate-precuneusinteractions:anew
locus of dysfunction in adult attention-deficit/hyperactivity
disorder.Biol.Psychiatry,63,332−337
67.
Burstein, R., Noseda, R. and Borsook, D. (2015) Migraine:
multiple processes, complex pathophysiology. J. Neurosci., 35,
6619−6629
68.
Zhou,H.X.,Chen,X.,Shen,Y.Q.,Li, L.,Chen, N.X., Zhu,Z.69.
C.,Castellanos,F.X.and Yan,C.G.(2020)Rumination andthe
default mode network: Meta-analysis of brain imaging studies
andimplicationsfordepression.Neuroimage,206,116287
Chen, X., Chen, N.-X., Shen, Y.-Q., Li, H.-X., Li, L., Lu, B.,
Zhu, Z. C., Fan, Z. and Yan, C. G. (2020) The subsystem
mechanism of default mode network underlying rumination: A
reproducibleneuroimagingstudy.Neuroimage,221,117185
70.
Greicius, M. D., Flores, B. H., Menon, V., Glover, G. H.,
Solvason, H. B., Kenna, H., Reiss, A. L. and Schatzberg, A. F.
(2007)Resting-state functionalconnectivity inmajor depression:
abnormally increased contributions from subgenual cingulate
cortexandthalamus.Biol.Psychiatry,62,429−437
71.
Hamilton, J. P., Farmer, M., Fogelman, P. and Gotlib, I. H.
(2015)Depressiverumination,thedefault-modenetwork,andthe
dark matter of clinical neuroscience. Biol. Psychiatry, 78,
224−230
72.
Silvers,J.A.,Weber,J.,Wager,T.D.andOchsner,K.N.(2015)
Bad and worse: neural systems underlying reappraisal of high-
and low-intensity negative emotions. Soc. Cogn. Affect.
Neurosci.,10,172−179
73.
George, M. S., Taylor, J. J. and Short, E. B. (2013) The
expandingevidencebaseforrTMStreatmentofdepression.Curr.
Opin.Psychiatry,26,13−18
74.
Ochsner,K. N.,Silvers, J.A. andBuhle, J.T.(2012)Functional
imaging studies of emotion regulation: a synthetic review and
evolvingmodelof the cognitive control ofemotion.Ann. N. Y.
Acad.Sci.,1251,E1−E24
75.
Eysenck, M. and Keane, M. (2020) Cognitive Psychology: A
Student’sHandbook.London:Taylor&FrancisGroup
76.
Gross, J. J. and Thompson, R. (2007) Emotion regulation:
conceptual foundations. In: Handbook of Emotion Regulation,
Gross,J.J.(Ed.),pp.3–24.NewYork:TheGuilfordPress
77.
Wang,L., Zhao, Y.,Edmiston,E. K., Womer,F.Y., Zhang,R.,
Zhao, P., Jiang, X., Wu, F., Kong, L., Zhou, Y., et al. (2020)
Structuralandfunctionalabnormitiesofamygdalaandprefrontal
cortexin major depressivedisorder withsuicideattempts. Front.
Psychiatry,10,923
78.
Peterson, B. S. (2015) Editorial: Research Domain Criteria
(RDoC): a new psychiatric nosology whose time has not yet
come.J.ChildPsychol.Psychiatry,56,719−722
79.
Sanislow,C.A., Ferrante, M., Pacheco,J.,Rudorfer, M. V. and
Morris, S. E. (2019) Advancing translational research using
NIMH research domain criteria and computational methods.
Neuron,101,779−782
80.
Tozzi,L., Staveland,B.,Holt-Gosselin, B.,Chesnut, M., Chang,
S.E.,Choi,D.,Shiner,M.,Wu,H.,Lerma-Usabiaga,G.,Sporns,
O., et al. (2020) The human connectome project for disordered
emotional states: Protocol and rationale for a research domain
criteria study of brain connectivity in young adult anxiety and
depression.Neuroimage,214,116715
81.
Association, A. P. (2013) Diagnostic and Statistical Manual of
Mental Disorders (DSM-5®). American Psychiatric Publishing,
Inc
82.
Gotlib,I.H. and Joormann,J.(2010) Cognition and depression:83.

SystematicreviewoffMRIstudiesonMDD

©TheAuthors(2022).PublishedbyHigherEducationPress 379

currentstatusandfuturedirections.Annu.Rev.Clin.Psychol.,6,
285−312
Müller,V. I.,Cieslik, E.C.,Serbanescu,I.,Laird,A.R.,Fox,P.
T. and Eickhoff, S. B. (2017) Altered brain activity in unipolar
depression revisited: Meta-analyses of neuroimaging studies.
JAMAPsychiatry,74,47−55
84.
Cesario, J. (2014) Priming, replication, and the hardest science.
Perspect.Psychol.Sci.,9,40−48
85.
Maxwell, S. E., Lau, M. Y. and Howard, G. S. (2015) Is
psychology suffering from a replication crisis? What does
“failuretoreplicate”reallymean?AmPsychol.,70,487−498
86.
Tackett, J. L., Lilienfeld, S. O., Patrick, C. J., Johnson, S. L.,
Krueger, R. F., Miller, J. D., Oltmanns, T. F. and Shrout, P. E.
(2017) It’s time to broaden the replicability conversation:
Thoughts for and from clinical psychological Science. Perspect.
Psychol.Sci.,12,742−756
87.
Tackett, J. L., Brandes, C. M., King, K. M. and Markon, K. E.
(2019) Psychology’s Replication Crisis and Clinical
PsychologicalScience.Annu.Rev.Clin.Psychol.,15,579−604
88.
Chao-Gan, Y. and Yu-Feng, Z. (2010) DPARSF: A MATLAB
toolboxfor “pipeline”dataanalysis ofresting-state fMRI.Front.
Syst.Neurosci.,4,13
89.
Ahmed, A. T., Frye, M. A., Rush, A. J., Biernacka, J. M.,
Craighead,W. E., McDonald,W.M., Bobo, W.V.,Riva-Posse,
P.,Tye, S. J.,Mayberg, H. S.,et al.(2018)Mapping depression
90.
ratingscale phenotypesontoresearch domaincriteria (RDoC)to
informbiologicalresearch in mooddisorders. J. Affect.Disord.,
238,1−7
Bullmore, E. and Sporns, O. (2009) Complex brain networks:
graph theoretical analysis of structural and functional systems.
Nat.Rev.Neurosci.,10,186−198
91.
Dale,A.M.,Fischl,B.andSereno,M.I.(1999)Corticalsurface-
based analysis. I. Segmentation and surface reconstruction.
Neuroimage,9,179−194
92.
Bassett,D.S.andSporns, O.(2017) Networkneuroscience.Nat.
Neurosci.,20,353−364
93.
Long,Y., Cao,H.,Yan, C.,Chen,X., Li,L.,Castellanos, F.X.,
Bai,T.,Bo,Q.,Chen,G.,Chen,N.,et al.(2020)Alteredresting-
state dynamic functional brain networks in major depressive
disorder: Findings from the REST-meta-MDD consortium.
NeuroimageClin.,26,102163
94.
Ye, M., Yang, T., Qing, P., Lei, X., Qiu, J. and Liu, G. (2015)
Changes of functional brain networks in major depressive
disorder: A graph theoretical analysis of resting-state fMRI.
PLoSOne,10,e0133775
95.
Coalson,T. S., VanEssen, D.C.and Glasser,M.F. (2018)The
impact of traditional neuroimaging methods on the spatial
localization of cortical areas. Proc. Natl. Acad. Sci. USA, 115,
E6356−E6365
96.

Xue-YingLietal.

380 ©TheAuthors(2022).PublishedbyHigherEducationPress