UBC Faculty Research and Publications

The association between cognitive function and white matter lesion location in older adults: a systematic… Bolandzadeh, Niousha; Davis, Jennifer C; Tam, Roger; Handy, Todd C; Liu-Ambrose, Teresa Oct 30, 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12883_2012_Article_687.pdf [ 300.42kB ]
JSON: 52383-1.0223410.json
JSON-LD: 52383-1.0223410-ld.json
RDF/XML (Pretty): 52383-1.0223410-rdf.xml
RDF/JSON: 52383-1.0223410-rdf.json
Turtle: 52383-1.0223410-turtle.txt
N-Triples: 52383-1.0223410-rdf-ntriples.txt
Original Record: 52383-1.0223410-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessThe association between cognitive function andwhite matter lesion location in older adults: asystematic reviewNiousha Bolandzadeh1, Jennifer C Davis2, Roger Tam3, Todd C Handy4 and Teresa Liu-Ambrose1,5,6*AbstractBackground: Maintaining cognitive function is essential for healthy aging and to function autonomously withinsociety. White matter lesions (WMLs) are associated with reduced cognitive function in older adults. However,whether their anatomical location moderates these associations is not well-established. This review systematicallyevaluates peer-reviewed evidence on the role of anatomical location in the association between WMLs andcognitive function.Methods: In accordance with the preferred reporting items for systematic reviews and meta-analysis (PRISMA)statement, databases of EMBASE, PUBMED, MEDLINE, and CINAHL, and reference lists of selected papers weresearched. We limited our search results to adults aged 60 years and older, and studies published in the Englishlanguage from 2000 to 2011. Studies that investigated the association between cognitive function and WMLlocation were included. Two independent reviewers extracted: 1) study characteristics including sample size, samplecharacteristic, and study design; 2) WML outcomes including WML location, WML quantification method (scoring orvolume measurement), strength of the MRI magnet in Tesla, and MRI sequence used for WML detection; and 3)cognitive function outcomes including cognitive tests for two cognitive domains of memory and executivefunction/processing speed.Results: Of the 14 studies included, seven compared the association of subcortical versus periventricular WMLswith cognitive function. Seven other studies investigated the association between WMLs in specific brain regions(e.g., frontal, parietal lobes) and cognitive function. Overall, the results show that a greater number of studies havefound an association between periventricular WMLs and executive function/processing speed, than subcorticalWMLs. However, whether WMLs in different brain regions have a differential effect on cognitive function remainsunclear.Conclusions: Evidence suggests that periventricular WMLs may have a significant negative impact on cognitiveabilities of older adults. This finding may be influenced by study heterogeneity in: 1) MRI sequences, WMLquantification methods, and neuropsychological batteries; 2) modifying effect of cardiovascular risk factors; and 3)quality of studies and lack of sample size calculation.Keywords: White matter lesions, Distribution, Cognition, Aging* Correspondence: tlambrose@exchange.ubc.ca1Department of Physical Therapy, University of British Columbia, 212-2177Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada5Brain Research Centre, University of British Columbia, 212-2177 WesbrookMall, Vancouver, BC V6T 1Z3, CanadaFull list of author information is available at the end of the article© 2012 Bolandzadeh et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of theCreative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use,distribution, and reproduction in any medium, provided the original work is properly cited.Bolandzadeh et al. BMC Neurology 2012, 12:126http://www.biomedcentral.com/1471-2377/12/126BackgroundThe world’s population is aging [1]. Maintaining cogni-tive function is essential for healthy aging and to func-tion autonomously within society.With age, the brain undergoes both structural and func-tional changes [2-5]. Specifically, cerebral white matterlesions (WMLs) are prevalent among adults aged 60 yearsor older [6,7]. These lesions are due to damage to thebrain parenchyma [8], ranging from demyelination tocomplete axonal disruptions [9,10]. Although their patho-genesis is unknown, there is a growning recognition thatWMLs are most likely the result of cerebrovascular disor-ders and cerebral ischemia [8,11-13]. The current goldstandard for diagnosis of WMLs includes various MRIsequences, such as T1, T2, proton density (PD), or fluidattenuated inversion recovery (FLAIR).White matter lesions are associated with both impairedmobility and reduced cognitive performance as measuredby standard neuropsychological testing, which might becaused by impairing the speed or integrity of signal trans-mission [14,15]. Specially, WML load has a negative im-pact on multiple domains of cognitive function such asmemory, processing speed, attention, and executive func-tion [8,16]. Pantoni et al. [16] summarized the results of16 studies focusing on the effect of WMLs on differentcognitive domains. Their results showed that, despite thefact that the probability of finding a positive associationbetween WML load and cognitive decline may be affectedby the cognitive domains assessed, an effect of WML oncognition was present invariably. However, emerging evi-dence suggests that WML distribution, as well as load,may also be a predictor of reduced cognitive performance[17,18]. In a study by Kim et al. [17], it is suggested that aspecific distribution of fiber tract damage is more asso-ciated with cognitive and motor impairment, comparedwith the total WML load. Thus, we conducted a system-atic review to ascertain the role of anatomical location inthe association between WMLs and cognitive function inolder adults.MethodsSearch strategyIn accordance with the preferred reporting items for sys-tematic reviews and meta-analysis (PRISMA) statement[19], we [NB, JCD and TLA] conducted a search ofEMBASE, MEDLINE, PUBMED, and CINAHL supple-mented by manual search of included articles’ referencelists. The search strategy (Figure 1(A)) was developed byApril 19th 2011, and includes studies from 2000 to 2011.We limited our search results to adults aged 60 yearsand older, and studies published in the English language.Study selectionWe excluded case-studies, reviews, and articles lackingWML quantification or measurements of cognitive func-tion, based on their titles and abstracts (Figure 1(B)).Figure 1 (A): Searching strategy retrieved from Ovid, (B): Flowchart of study selection.Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 2 of 10http://www.biomedcentral.com/1471-2377/12/126Also, any study with the primary focus on psychiatricconditions (e.g., depression) or progressive neurodegen-erative diseases (except for Alzheimer’s disease (AD) andcerebrovascular disorders due to the high prevalence ofWMLs) was excluded. Based on full text review, weexcluded studies that: 1) used computed tomography (asit is less sensitive than MRI in detection of WMLs [20]),or used MRI device with a magnet strength of less than1.5T and; 2) assessed only global cognition (measured bymini-mental state examination (MMSE)) as it may notbe sensitive to the differential effects of WML location;and 3) did not detail WML location.Data extraction and quality assessmentWe [NB and TLA] developed a list of extraction itemsincluding: 1) study characteristics; 2) WML outcomes;and 3) cognitive function outcomes. One study [21] didnot report the strength of MRI magnet and NB con-tacted the author.Two authors [NB, TLA] independently evaluated eachstudy based on four quality assessments questions (seeTable 1), and all the discrepancies were reviewed by JCDand RT. Assessing the validity of WML quantificationwas influenced by the difficulty in the differential diag-nosis of WMLs, which requires expert radiologicalknowledge to be done accurately [22]. In addition, theintensity range of lesions typically overlaps with those ofhealthy tissues, so automatic identification methods tendto produce more false positives as compared with man-ual identification by a radiologist [23]. Therefore, our as-sessment favors quantification methods that useradiologist/physician identification of WMLs. We useddichotomized answers (+: yes, -: no) for the quality as-sessment questions.ResultsOverview of studiesThe initial number of articles identified was 490 (Figure 1(B)).After duplicate removal, 156 papers were furtherexcluded using their title and abstract. We conducted afull text review of the remaining 48 articles. In total, 14articles met the inclusion criteria (see Tables 2, 3, 4, 5).These articles were further categorized into two groupsbased on the cognitive status of their study samples: 1)studies that did not compare subjects based on cognitivestatus (i.e., normal, cognitively impaired but not demented,and demented); and 2) studies that classified and com-pared subjects based on cognitive status. Table 6 showsthe most commonly-used cognitive tests in the 14included studies.Studies that did not compare subjects based on cognitivestatusSubcortical vs. periventricular WMLFive studies [24-28] – four cross-sectional studies andone prospective study – compared the association ofsubcortical versus periventricular WMLs with cognitivefunction. In the first cross-sectional study of 1077 olderadults [24], WMLs were defined as T2 and PD hyperin-tensities that were not T1 hypointensities. Four lobes offrontal, parietal, occipital, and temporal were consideredfor subcortical WML scoring. Three regions adjacent tofrontal horns, lateral ventricles wall, and occipital hornswere selected for periventricular WML scoring. TheTable 1 Quality assessment results for included studiesReference Q1. Was the WMLidentification done bya radiologist/physician?Q2. Was the cognitiveperformance measuredusing a standardized method?Q3. Was therea sample sizecalculation?Q4. Were age oreducation consideredas confounders?Groot et al. et al. [24] + + - +Shenkin et al. [25] + + - -Baune et al. [26] - + - +Kim et al. [27] - + - +Silbert et al. [28] - + - +McClleland et al. [21] + + - +Wright et al. [29] - + - +Kaplan et al. [30] - + - +Wakefield et al. [31] - + - +O’Brien et al. [32] + + - +Smith et al. [14] - + - +Burns et al. [33] + + - +Ishii et al. [34] + + - +Tullberg et al. [35] - + - -Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 3 of 10http://www.biomedcentral.com/1471-2377/12/126neuropsychological battery evaluated two domains ofmemory and executive function/processing speed. Theresults showed that when controlled for subcorticalWML severity, increased periventricular WML severityin all the three regions was associated with reduced per-formance in both cognitive domains (p<0.01). However,when controlled for periventricular WMLs, no such as-sociation was found for subcortical WMLs.In the second cross-sectional study of 105 older adults[25], WMLs were identified using T2 and FLAIR scans.Results showed that higher periventricular and subcor-tical WML scores were not significantly associated withreduced memory and executive function/processingspeed.In a sample of 268 older adults [26], WMLs were cate-gorized into three groups of: 1) large subcortical WMLsdefined as PD and T2 hyperintensities that were not T1hypointensities; 2) infarction lesions defined as lesions of≥2 mm that were either T2 hyperintensities, or PD andT1 hypointensities; and 3) periventricular WMLs. Theresults indicated that large subcortical WMLs were sig-nificantly associated with memory, and infarction lesionswere significantly associated with executive function/processing speed (p<0.05). Contrary to the results of twopreviously mentioned studies, this study found no sig-nificant relationship between periventricular WMLs andcognitive performance.In the last cross-sectional study, Kim et al. [27] definedWMLs as T2 and FLAIR hyperintensities. Over the 84older adults, only periventricular WML was significantlycorrelated with memory and executive function/proces-sing speed, when both the periventricular and subcorticalWMLs were entered simultaneously into the regressionmodel (p<0.05).The one longitudinal study [28] used a sample of 104subjects to investigate the impact of WML volume pro-gression on the rate of cognitive decline. White matterlesions were defined as PD and T2 hyperintensities. In-farction lesions – detected by their clean or sharp edges,and if they were relatively dark on PD scans – wereexcluded from WML analysis. The neuropsychologicalbattery assessed only memory. Higher rate of subcortical(but not periventricular) WML volume change was asso-ciated with increased rate of decline in memory scores(p<0.001).Regional WMLsSix cross-sectional studies [14,21,29-32] examined theassociation between WMLs in specific brain regions(e.g., frontal, parietal, etc.) and cognitive performance.McClelland et al. [21] defined WMLs as PD and T2hyperintensities that were T1 hypointensity. The resultsin 3647 older adults suggested that WMLs located incerebellar and cerebral white matter and basal gangliawere significantly associated with reduced processingspeed performance (p<0.05).Among 656 older adults, Wright et al. [29] differen-tiated subclinical infarction lesions from the rest ofTable 2 Characteristics of studies included in thissystematic reviewReference Sample size Study designPublishing year Sample characteristicsGroot et al. [24] 1077 Cross-Sectional2000 Subsample of Rotterdamand Zeotemeer StudiesShenkin et al. [25] 105 Cross-Sectional2005 Random Sample ofCommunity-DwellingParticipantsBaune et al. [26] 268 Cross-Sectional2009 Subsample of MEMO StudyKim et al. [27] 84 Cross-Sectional2011 Random Sample ofNormals/Recruited fromMemory ClinicSilbert et al. [28] 104 Longitudinal2008 Subsample of Oregon BrainAging StudyMcClleland et al. [21] 3647 Cross-Sectional2000 Subsample of CHS CohortWright et al. [29] 656 Cross-Sectional2008 Subsample of NOMAS CohortstudyKaplan et al. [30] 95 Cross-Sectional2009 Random Sample of ParticipantsWakefield et al. [31] 99 Cross-Sectional2010 Sample Selected for aLongitudinal StudyO’Brien et al. [32] 149 Cross-Sectional2002 Subsample of SCOPE StudySmith et al. [14] 145 Cross-Sectional2011 Subsample of ProspectiveStudyBurns et al. [33] 156 Cross-Sectional2005 88 Normal (CDR=0), 68Early-Stage AD (CDR=0.5,1)Ishii et al. [34] 453 Cross-Sectional2007 340 (CDR=0), 113 (CDR=0.5)Tullberg et al. [35] 78 Cross-Sectional2004 22 Normal (CDR=0), 30 CIND(CDR=0.5), 26 Demented(CDR≥1)Abbreviations: MEMO = Memory and Morbidity in Augsburg Elderly; CDR =Clinical Dementia Rating Scale; CHS = Cardiovascular Health Study; NOMAS =Northern Manhattan Study; SCOPE = Study on Cognition and Prognosis inElderly; CIND = Cognitively Impaired not Demented.Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 4 of 10http://www.biomedcentral.com/1471-2377/12/126WMLs based on the size, location, and imaging charac-teristics obtained from PD, T2, and FLAIR scans. Theywere grouped by location into frontal, deep andoccipital-temporal-parietal networks. The neuropsycho-logical battery assessed only executive function/proces-sing speed. The results demonstrated that individualswith infarction lesions in frontal and deep locations hadsignificantly worse cognitive performance (p<0.05).Kaplan et al. [30] studied a sample of 95 older adults.White matter lesions were defined as FLAIR and T2hyperintensities, and were categorized into frontal andposterior regions. The results showed that frontal WMLswere associated with memory (p<0.05) and executivefunction/processing speed (p<0.001).Furthermore, Wakefield et al. [31] detected WMLsbased on FLAIR and T1 scans in a sample of 99community-dwelling older adults. The following regionsof interest were segmented for WMLs: anterior, superior,and posterior corona radiata; cingulate gyrus, genu, body,and splenium of corpus callusum; anterior and posteriorlimb of internal capsule; and superior longitudinal fascic-ulus. The neuropsychological battery assessed only execu-tive function/processing speed. In regions of posteriorcorona radiata and splenium of corpus callosum, the totalamount of WMLs was significantly associated with execu-tive function/processing speed (p<0.05).O’Brien et al. [32] detected WMLs based on FLAIRand T2 scans, in 149 older adults. The focus of theirTable 3 Outcome Measures: white matter lesion quantificationReference Sequence WML Location MRIMagnetWML Type WML QuantificationGroot et al. [24] PD, T1, T2 S: Four lobes of Frontal, Parietal, Occipital, and Temporal 1.5 TP, S, Regions Scoring P: Adjacent frontal horns, lateral ventricles wall, and occipital hornsShenkin et al. [25] T2, FLAIR - 1.5 TS, P ScoringBaune et al. [26] PD, T1, T2 - 1.5 TS, P ScoringKim et al. [27] T2, FLAIR - 1.5 TS, P VolumSilbert et al. [28] PD, T2 - 1.5 TS, P VolumeMcClleland et al. [21] PD, T1, T2 Cerebral White Matter, Cerebellar White Matter, Basal Ganglia 1.5 TRegions ScoringWright et al. [29] PD, T2, FLAIR Frontal, Deep, and Occipital-Temporal-Parietal 1.5 TS, I, Regions VolumeKaplan et al. [30] T2, FLAIR Frontal and Posterior Regions 3.0 TRegions VolumeWakefield et al. [31] T1, FLAIR Anterior, Superior, Posterior Corona Radiata 3.0 TRegions Volume Cingulate Gyrus, Genu, Body, Splenium of Corpus CallusumAnterior and Posterior Limb of Internal CapsuleSuperior Longitudinal FasciculusO’Brien et al. [32] T2, FLAIR Internal and External Capsule 1.5 TRegions ScoringSmith et al. [14] PD, T1, T2 Whole Brain 1.5 TRegions VolumeBurns et al. [33] T1, T2 S: Frontal, Parietal, Temporal, and Occipital Lobes 1.5 TS, P, Regions Scoring P: Right and Left Frontal Horns, Posterior Horns, and Ventricular BodiesIshii et al. [34] T2 S: Left and Right 1.5 TP, S, Regions Scoring P: Anterior and PosteriorTullberg et al. [35] T1, T2 Orbitofrontal, Prefrontal, Dorsolateral Frontal, Parietal, and Occipitotemporal 1.5 TRegions VolumeAbbreviations: PD = Proton Density; FLAIR = Fluid Attenuated Inversion Recovery.Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 5 of 10http://www.biomedcentral.com/1471-2377/12/126analysis was on the distribution of WMLs in the internaland external capsule. They found that WMLs from bothregions were significantly associated with cognitive per-formance of speed of memory retrieval and executivefunction/processing speed (p<0.05).Smith et al. [14] analyzed WML distribution using PD,T2, and T1 scans in the whole brain of 147 older adults.The total volume of WMLs was associated with the cog-nitive performance of memory (p<0.01) and executivefunction (p=0.05). In the following locations, WMLswere significantly associated with memory: right inferiortemporal-occipital, left temporal-occipital periventricu-lar, and right parietal periventricular; and anterior limbof internal capsule. Also, WMLs in the following regionswere significantly associated with executive function: thebilateral inferior frontal, temporal-occipital periventricu-lar, right parietal periventricular, and prefrontal whitematter; and the anterior limb of the internal capsulebilaterally.Studies that classified and compared subjects based oncognitive statusSubcortical vs. periventricular WMLAmong the studies that classified participants based ontheir cognitive status, two cross-sectional studies [33,34]compared the effects of subcortical and periventricularWMLs. Burns et al. [33] included 88 non-demented par-ticipants (clinical dementia rating (CDR) score=0), 68with early-stage AD (48 with very mild AD (CDR=0.5),and 20 with mild AD (CDR=1)). White matter lesionswere defined as T2 hyperintensities that were T1 hypoin-tensities. Subcortical WMLs were rated in regions offrontal, parietal, temporal, and occipital lobes. Periven-tricular WMLs were rated in right and left frontal horns,posterior horns, and ventricular bodies. For non-demented participants, only associate memory was asso-ciated with periventricular WMLs (p<0.01). For participantswith early-stage AD, memory and executive function/processing speed were associated with both periventricularand subcortical WMLs (p<0.05).Ishii et al. [34] detected WMLs based on T2 hyperin-tensities. Sample of 453 older adults were categorizedinto two groups of CDR=0 and CDR=0.5. Anterior andposterior periventricular WMLs, as well as left and rightsubcortical WMLs were segmented. The results sug-gested that, for CDR=0 group, anterior periventricularWMLs and a test of executive function/processing speedwere significantly correlated (p=0.001).Regional WMLThe last study [35] detected WMLs based on T1 and T2scans. They categorized 78 older adults into threeTable 4 Outcome measures: Cognitive tests used for two cognitive domains of memory and executive function/processing speedReference Executive function / Processing speed MemoryGroot et al. et al. [24] Stroop, Letter-Digit Substitution Task,Verbal FluencyRey’s Auditory, Memory Scanning TaskShenkin et al. [25] Verbal Fluency, Controlled Word Association,Moray House Test, Raven’s Progressive MatricesWechsler Memory ScaleBaune et al. [26] Stroop, Letter-Digit Substitution Task 3-Word RecallKim et al. [27] Boston Naming, Buccofacial Praxis Test,Semantic Controlled Oral Word AssociationTest, Stroop Color, Word TestSeoul Verbal Learning Test, Ray Complex Figure Test,Delayed Recall and Recognition, Digit Span TestsSilbert et al. [28] - Delayed Story RecallMcClleland et al. [21] Digit-Symbol Substitution Task -Wright et al. [29] Color Trail 1 & 2 -Kaplan et al. [30] Stroop, Trail Making, CalCap Repeated Battery for Neuropsychological StatusWakefield et al. [31] Stroop, Trail Making 1 & 2, CalCap -O’Brien et al. [32] Verbal Fluency, Trail Making 1 & 2 Memory Component of CDRSmith et al. [14] Letter Fluency, Trail Making 2 Episodic Memory, Alpha Span TestBurns et al. [33] Trail Making 1 & 2, Short Blessing Test,Boston NamingWechsler Memory Scale, Wechsler AdultIntelligence ScaleIshii et al. [34] Verbal Fluency, Trail Making Test,Benton’s Visual Form TestADAS-Cog, 10 Word Recall, Digit Span ForwardTullberg et al. [35] Verbal Fluency Wechsler Memory Scale, Word List Learning,Digit Span BackwardAbbreviations: CalCap = California Computerized Assessment Package; CDR = Clinical Dementia Rating Scale; ADAS-Cog = Alzheimer’s Disease Assessment Scale-Cognitive.Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 6 of 10http://www.biomedcentral.com/1471-2377/12/126cognitive groups: normal (CDR=0), cognitively impairedbut not demented (CDR=0.5), and demented (CDR≥1),either by AD or vascular dementia. WMLs were ana-lyzed in regions of orbitofrontal, prefrontal, dorsolateralfrontal, parietal, and occipitotemporal. In non-dementedindividuals, increased volumes of frontal (specifically,prefrontal and dorsolateral), parietal, and occipital WMLwere separately associated with lower executive func-tion/processing speed scores (p<0.05). Frontal WMLswere also associated with reduced memory function innon-demented group (p<0.05). No association was foundfor individuals with dementia.Quality assessmentThe quality assessment results for each of the four ques-tions are presented in Table 1: 1) in seven studies, WMLidentification is done by a radiologist/physician, whilethe remaining used automatic methods; 2) all the articlesemployed standard methods for cognitive assessment; 3)none of the studies provided sample size calculation;and 4) the statistical analyses of twelve studies includedage or education as confounders.DiscussionSubcortical vs. periventricular WMLsBased on their proximity to ventricles, WMLs were clas-sified as subcortical or periventricular in seven studies[24-28,33,34]. The results show that more studies havefound an association between periventricular WMLswith the cognitive domain of executive function, thansubcortical WMLs.Table 5 Association between the structural location of white matter lesion (i.e., subcortical, periventricular, orregional) with two domains of cognitive function (i.e., memory and executive function/processing speed)Reference AssociationGroot et al. et al. [24] Controlled for subcortical, periventricular WMLs were associated with memory and executive function/processing speed.Shenkin et al. [25] Subcortical and periventricular WMLs were not associated with any of the cognitive measurements.Baune et al. [26] Subcortical WMLs were associated with memory.As a subgroup of subcortical WMLs, infarction lesions were associated with executive function/processing speed.Periventricular WMLs were not associated with any of the cognitive functions.Kim et al. [27] Only periventricular WML was significantly correlated with memory and executive function/processing speed, whenboth the periventricular and subcortical WMLs were entered simultaneously into the regression model.Silbert et al. [28] Change in subcortical WMLs (excluding infarction lesions) was associated with memory decline. This association wasnot true for periventricular WMLs.McClleland et al. [21] White matter lesions were associated with executive function/processing speed, in all white matter regions of cerebrum,cerebellum, and basal ganglia.Wright et al. [29] Subcortical WMLs (including infarction lesions) were associated with executive function/processing speed, in regionsof frontal and deep white matter.Kaplan et al. [30] White matter lesions were associated with memory and executive function/processing speed, in frontal regions.Wakefield et al. [31] White matter lesions were associated with executive function/processing speed in white matter regions of posteriorcorona radiata and splenium of corpus callosum.O’Brien et al. [32] White matter lesions were associated with speed of memory retrieval and executive function/processing speed.Smith et al. [14] White matter lesions were associated with memory and executive function/processing speed. White matter lesions inthe following locations were significantly associated with memory: right inferior temporal-occipital, left temporal-occipitalperiventricular, and right parietal periventricular white matter; and anterior limb of internal capsule. Also, WMLs in thefollowing regions were significantly associated with executive function: the bilateral inferior frontal, temporal-occipitalperiventricular, right parietal periventricular, and prefrontal white matter; and the anterior limb of the internal capsulebilaterally.Burns et al. [33] For non-demented participants, only associate memory was associated with periventricular WMLs. For participants withearly-stage Alzheimer’s Disease (AD), memory and executive function/processing speed were associated with bothperiventricular and subcortical WMLs.Ishii et al. [34] For CDR=0 group, anterior periventricular WML and a test of executive function/processing speed were significantlycorrelated.Tullberg et al. [35] In non-demented individuals, increased volumes of frontal (specifically prefrontal and dorsolateral), parietal, and occipitalWML were separately associated with lower executive function/processing speed scores.Frontal WMLs were also associated with reduced memory function in non-demented group. No association was foundfor individuals with dementia.Abbreviations: WML = White Matter Lesion; CDR = Clinical Dementia Rating Scale.Table 6 The most commonly-used neuropsychologicaltests in the included studiesExecutive Function Trail-Making Test, Stroop Test, Verbal Fluency TestMemory Wechseler Memory Scale, Word Recall TestBolandzadeh et al. BMC Neurology 2012, 12:126 Page 7 of 10http://www.biomedcentral.com/1471-2377/12/126Subcortical WMLs are believed to primarily disruptshort connections, and thus impairing cognitive per-formance supported by the specific brain region [24].For example, dexterous hand and arm movements aregenerally thought to be primarily supported by themotor cortex. Therefore, subcortical WMLs in this spe-cific region can result in reduced performance in handand arm dextrous movements [36]. In contrast, periven-tricular WMLs disrupt longer connections to spatiallydistant cortical areas, and thus can cause cognitive per-formance decline in multiple domains [24,27]. For ex-ample, executive function tasks typically used inresearch experiments depend on multiple brain regions(i.e., frontal and non-frontal) which are not necessarilylocated spatially close to each other [37]. Therefore, anydisruption in long white matter tracts traversing fromperiventricular areas may initially reduce the axonaltransmission speed [38], and later cause impaired execu-tive function. In summary, cognitive function dependson intact connections within subcortical areas and be-tween cortical and subcortical structures, and any dis-ruption in these connections may impair cognitivefunction.We categorized all included studies into two majorcognitive domains which are sensitive to aging: 1) mem-ory; or 2) executive function/processing speed. The lattercategory was a combination of two cognitive domainsbased on the idea that they are not mutually exclusive,and one needs to control for their mutual relationshipbefore examining their unique effects [39].For memory, out of seven studies, three studies[24,27,33] found a significant association between peri-ventricular WMLs and memory performance, two stud-ies [26,28] found a significant association betweensubcortical WMLs and memory performance, and twostudies [25,34] did not find any association.For executive function/processing speed, out of six stud-ies, three studies [24,27,34] found a significant associationbetween periventricular WMLs and executive function/processing speed, while only one study [26] found a sig-nificant association between subcortical WMLs and ex-ecutive function/processing speed. Two studies [25,33]did not find any association.Thus, our overall results show that greater numberof studies found an association between cognitive im-pairment (in both domains of memory and executivefunction/processing speed) and periventricular WMLs,compared with subcortical WMLs. Moreover, greaternumber of studies showed an association between im-pairment in the domain of executive function/proces-sing speed with periventricular WMLs, compared tosubcortical WMLs.As highlighted earlier, periventricular WMLs may im-pact multiple domains of cognition because they disruptdistant connections. Hence, our findings concur with thegeneral knowledge that the domain of executive func-tion/processing speed may depend on multiple brainregions and spatially distant connections [37,40].Regional WMLSeven studies [14,21,29-32,35] investigated regionalWMLs. No common pattern was evident secondary tothe heterogeneity of regions studied.The following regions demonstrated significant associa-tions between WMLs and cognitive function: cerebralwhite matter, cerebellar white matter, and basal ganglia[21]; frontal (dorsolateral frontal and prefrontal) [29-31],parietal, occipital, and temporal lobes [29,31,35]; internaland external capsule [32]; posterior corona radiata, andsplenium corpus callosum [31]. This systematic reviewprovides researchers with a summary set of brain regionsin which an association have been found between WMLsand cognitive performance. To better understand the roleof anatomical location in the association between WMLand cognitive function, future studies should examine thespatial distribution of WMLs on the whole brain, or spe-cific set of brain regions identified in this review as beinghighly associated with cognitive dysfunction.LimitationsThe discrepancies between the results may be due to theheterogeneous study methodologies and the quality ofincluded studies.Different MRI sequences, WML quantification methods, andneuropsychological batteriesThe included studies were heterogeneous in MRIsequences for WML detection (i.e., PD, T1, T2, or FLAIR),WML quantification method (i.e., scoring or volume mea-surements), and components of neuropsychological bat-teries. This likely contributed to variability in our results.Moreover, two different methods were used for WMLquantification: 1) scoring [24,25,33,34]; and 2) volumemeasurement [26-28]. Scoring measures are usually donemanually, and show a higher accuracy for selection of sub-tle WMLs, compared to automatic volumetric methods.However, these methods vary significantly in terms of le-sion classification and severity scoring. Moreover, eachscoring method has its own specific limitations.For WML volume measurement, there are two steps.The first step is identifying lesions, which can be done ei-ther manually by an expert radiologist or automatically.After the WMLs are identified manually or automatically,one can proceed to the second step, which is measuringWML volumes automatically. It has been shown that bothscoring and volumetric quantification methods are reliablefor measuring WML load [41,42]. However, periventricu-lar and subcortical WMLs quantified by these twoBolandzadeh et al. BMC Neurology 2012, 12:126 Page 8 of 10http://www.biomedcentral.com/1471-2377/12/126quantification methods are differently associated with cog-nitive function [42]. Out of three studies which used vol-ume measurement, two studies [26,28] showed asignificant association between the subcortical WMLs andcognitive performance. Out of four studies which usedscoring, three [24,33,34] showed a significant associationbetween periventricular WMLs and cognitive perform-ance. These results suggest that scoring might have biasedthe results toward periventricular WMLs. Conversely, vol-ume measurement might be problematic for periventricu-lar WMLs due to their similar appearance to CSF onsome MRI sequences (e.g., T2 or T1) [43].Modifying effect of cardiovascular risk factorsThere is a growing recognition that WMLs are associatedwith age and cardiovascular risk factors [8,44]. However,all but one included study [21] considered the modifyingeffect of cardiovascular risk factors in the statistical ana-lysis. We recommend that future studies consider includ-ing cardiovascular risk factors in their analysis.Quality of studies and lack of sample size calculationOne study [25] did not demonstrate a significant associ-ation between any type of WMLs and any of the cogni-tive tasks. Based on our quality assessment, this study isthe only study categorizing WMLs locations as subcor-tical and periventricular that did not consider age oreducation as potential confounders. Therefore, we con-cluded that this study did not provide strong evidencefor the lack of correlation between WMLs and cognitivefunction.Moreover, the lack of sample size calculations in all ofthe included studies might have resulted in possible typeII errors. However, we do recognize that the lack of sam-ple size calculations may be due to the dearth of data inthis research area [45].ConclusionsThis study provides the first in depth analysis of brainregions where an association between WML location andcognitive decline has been found in older adults. Specific-ally, studies that considered periventricular versus subcor-tical WMLs suggest that, compared with subcorticalWMLs, periventricular WMLs may have a greater nega-tive impact on cognitive performance. Moreover, periven-tricular WMLs appear to be more associated to thedomain of executive function/processing speed, than tothe domain of memory. To further clarify the associationof cognitive function with WML locations, we suggest thatfuture studies consider spatial distribution of WMLs onthe whole brain.We did not proceed to a meta-analysis of the results, pri-marily because of the small number of studies systematic-ally found on this topic. Moreover, the neuropsychologicalbatteries used for assessing cognitive status, the WMLquantification method, and MRI sequences used for WMLdetection varied vastly between studies. Thus, it was notfeasible to conduct a meta-analysis.Source of fundingThis work was supported by the Canadian Stroke Net-work, the Heart and Stroke Foundation of Canada, andthe Canadian Institutes of Health Research (MOB-93373)to TLA. TLA is a Canada Research Chair (Tier II) inPhysical Activity, Mobility, and Cognitive Neuroscience, aMichael Smith Foundation for Health Research Scholar, aCanadian Institutes of Health Research New Investigator,and a Heart and Stroke Foundation of Canada's Henry JMBarnett's Scholarship recipient. JCD is a Michael SmithFoundation for Health Research Post-Doctoral Fellow anda Canadian Institutes of Health Research Post-DoctoralFellow. NB is a Heart and Stroke Foundation of CanadaDoctoral Trainee.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionAll authors participated, read, and approved the final manuscript.Author details1Department of Physical Therapy, University of British Columbia, 212-2177Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada. 2Centre for ClinicalEpidemiology and Evaluation, University of British Columbia, 212-2177Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada. 3Department of Radiology,University of British Columbia, 212-2177 Wesbrook Mall, Vancouver, BC V6T1Z3, Canada. 4Department of Psychology, University of British Columbia,212-2177 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada. 5Brain ResearchCentre, University of British Columbia, 212-2177 Wesbrook Mall, Vancouver,BC V6T 1Z3, Canada. 6Centre for Hip Health and Mobility, Vancouver CoastalHealth Research Institute, 212-2177 Wesbrook Mall, Vancouver, BC V6T 1Z3,Canada.Received: 28 March 2012 Accepted: 12 October 2012Published: 30 October 2012References1. Kinsella K: Demographic dimensions of global aging. J Fam Issues 2000,21(5):541–558.2. Park DC, Reuter-Lorenz P: The adaptive brain: aging and neurocognitivescaffolding. Annu Rev Psychol 2009, 60:173–196.3. Haug H, Eggers R: Morphometry of the human cortex cerebri and corpusstriatum during aging. Neurobiol Aging 1991, 12(4):336–338. discussion352–335.4. Adolfsson R, Gottfries CG, Roos BE, Winblad B: Post-mortem distribution ofdopamine and homovanillic acid in human brain, variations related toage, and a review of the literature. J Neural Transm 1979, 45(2):81–105.5. de Keyser J, De Backer JP, Vauquelin G, Ebinger G: The effect of aging onthe D1 dopamine receptors in human frontal cortex. Brain Res 1990,528(2):308–310.6. Breteler MM, van Swieten JC, Bots ML, Grobbee DE, Claus JJ, van den HoutJH, van Harskamp F, Tanghe HL, de Jong PT, van Gijn J, et al: Cerebralwhite matter lesions, vascular risk factors, and cognitive function in apopulation-based study: the rotterdam study. Neurology 1994,44(7):1246–1252.7. Lindgren A, Roijer A, Rudling O, Norrving B, Larsson EM, Eskilsson J, Wallin L,Olsson B, Johansson BB: Cerebral lesions on magnetic resonance imaging,heart disease, and vascular risk factors in subjects without stroke. Apopulation-based study. Stroke 1994, 25(5):929–934.Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 9 of 10http://www.biomedcentral.com/1471-2377/12/1268. Kuo HK, Lipsitz LA: Cerebral white matter changes and geriatricsyndromes: is there a link? J Gerontol A Biol Sci Med Sci 2004,59(8):818–826.9. Galluzzi S, Lanni C, Pantoni L, Filippi M, Frisoni GB: White matter lesions inthe elderly: pathophysiological hypothesis on the effect on brainplasticity and reserve. J Neurol Sci 2008, 273(1–2):3–9.10. Frisoni GB, Galluzzi S, Pantoni L, Filippi M: The effect of white matterlesions on cognition in the elderly—small but detectable. Nat Clin PractNeurol 2007, 3(11):620–627.11. Pantoni L, Garcia JH: Pathogenesis of leukoaraiosis: a review. Stroke 1997,28(3):652–659.12. Jokinen H, Kalska H, Ylikoski R, Madureira S, Verdelho A, Gouw A, ScheltensP, Barkhof F, Visser MC, Fazekas F: MRI-defined subcortical ischemicvascular disease: baseline clinical and neuropsychological findings.Cerebrovasc Dis 2009, 27(4):336–344.13. Pantoni L: Pathophysiology of age-related cerebral white matterchanges. Cerebrovasc Dis 2002, 13(2):7–10.14. Smith EE, Salat DH, Jeng J, McCreary CR, Fischl B, Schmahmann JD,Dickerson BC, Viswanathan A, Albert MS, Blacker D, et al: Correlationsbetween MRI white matter lesion location and executive function andepisodic memory. Neurology 2011, 76(17):1492–1499.15. Pantoni L, Poggesi A, Inzitari D: Cognitive decline and dementia related tocerebrovascular diseases: some evidence and concepts. Cerebrovasc Dis2009, 27(1):191–196.16. Pantoni L, Poggesi A, Inzitari D: The relation between white-matter lesionsand cognition. Curr Opin Neurol 2007, 20(4):390–397.17. Kim SH, Park JS, Ahn HJ, Seo SW, Lee JM, Kim ST, Han SH, Na DL: Voxel-based analysis of diffusion tensor imaging in patients with subcorticalvascular cognitive impairment: correlates with cognitive and motordeficits. J Neuroimaging 2011, 21(4):317–324.18. Desmond DW: Cognition and white matter lesions. Cerebrovasc Dis 2002,13(2):53–57.19. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, ClarkeM, Devereaux PJ, Kleijnen J, Moher D: The PRISMA statement for reportingsystematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. J Clin Epidemiol 2009,62(10):e1–e34.20. Gunning-Dixon FM, Raz N: The cognitive correlates of white matterabnormalities in normal aging: a quantitative review. Neuropsychology2000, 14(2):224–232.21. McClelland RL, Kronmal RA, Bryan RN, Manolio TA, Herskovits E, Kuller LH,O’Leary DH: Neurologic correlates of infarction-like lesion location onmagnetic resonance imaging in the cardiovascular health study. J StrokeCerebrovasc Dis 2000, 9(5):218–228.22. Filippi M, Falini A, Arnold DL, Fazekas F, Gonen O, Simon JH, Dousset V,Savoiardo M, Wolinsky JS: Magnetic resonance techniques for the in vivoassessment of multiple sclerosis pathology: consensus report of thewhite matter study group. J Magn Reson Imaging 2005, 21(6):669–675.23. McAusland J, Tam R, Wong E, Riddehough A, Li D: Optimizing the use ofradiologist seed points for improved multiple sclerosis lesionsegmentation. IEEE Trans Biomed Eng 2010, .24. de Groot JC, de Leeuw FE, Oudkerk M, van Gijn J, Hofman A, Jolles J,Breteler MM: Cerebral white matter lesions and cognitive function: therotterdam scan study. Ann Neurol 2000, 47(2):145–151.25. Shenkin SD, Bastin ME, Macgillivray TJ, Deary IJ, Starr JM, Rivers CS, WardlawJM: Cognitive correlates of cerebral white matter lesions and waterdiffusion tensor parameters in community-dwelling older people.Cerebrovasc Dis 2005, 20(5):310–318.26. Baune BT, Roesler A, Knecht S, Berger K: Single and combined effects ofcerebral white matter lesions and lacunar infarctions on cognitivefunction in an elderly population. J Gerontol A Biol Sci Med Sci 2009,64(1):118–124.27. Kim JH, Hwang KJ, Lee YH, Rhee HY, Park KC: Regional white matterhyperintensities in normal aging, single domain amnestic mild cognitiveimpairment, and mild Alzheimer's disease. J Clin Neurosci 2011,18(8):1101–1106.28. Silbert LC, Nelson C, Howieson DB, Moore MM, Kaye JA: Impact of whitematter hyperintensity volume progression on rate of cognitive andmotor decline. Neurology 2008, 71(2):108–113.29. Wright CB, Festa JR, Paik MC, Schmiedigen A, Brown TR, Yoshita M, DeCarliC, Sacco R, Stern Y: White matter hyperintensities and subclinicalinfarction: associations with psychomotor speed and cognitive flexibility.Stroke 2008, 39(3):800–805.30. Kaplan RF, Cohen RA, Moscufo N, Guttmann C, Chasman J, Buttaro M, HallCH, Wolfson L: Demographic and biological influences on cognitivereserve. J Clin Exp Neuropsychol 2009, 31(7):868–876.31. Wakefield DB, Moscufo N, Guttmann CR, Kuchel GA, Kaplan RF, Pearlson G,Wolfson L: White matter hyperintensities predict functional decline invoiding, mobility, and cognition in older adults. J Am Geriatr Soc 2010,58(2):275–281.32. O'Brien JT, Wiseman R, Burton EJ, Barber B, Wesnes K, Saxby B, Ford GA:Cognitive associations of subcortical white matter lesions in olderpeople. Ann N Y Acad Sci 2002, 977:436–444.33. Burns JM, Church JA, Johnson DK, Xiong C, Marcus D, Fotenos AF, SnyderAZ, Morris JC, Buckner RL: White matter lesions are prevalent butdifferentially related with cognition in aging and early Alzheimerdisease. Arch Neurol 2005, 62(12):1870–1876.34. Ishii H, Meguro K, Yamaguchi S, Ishikawa H, Yamadori A: Prevalence andcognitive performances of vascular cognitive impairment no dementiain Japan: the Osaki-Tajiri Project. Eur J Neurol 2007, 14(6):609–616.35. Tullberg M, Fletcher E, DeCarli C, Mungas D, Reed BR, Harvey DJ, WeinerMW, Chui HC, Jagust WJ: White matter lesions impair frontal lobefunction regardless of their location. Neurology 2004, 63(2):246–253.36. Darling WG, Pizzimenti MA, Rotella DL, Peterson CR, Hynes SM, Ge J, SolonK, McNeal DW, Stilwell-Morecraft KS, Morecraft RJ: Volumetric effects ofmotor cortex injury on recovery of dexterous movements. Exp Neurol2009, 220(1):90–108.37. Alvarez JA, Emory E: Executive function and the frontal lobes: a meta-analytic review. Neuropsychol Rev 2006, 16(1):17–42.38. van den Heuvel DM, ten Dam VH, de Craen AJ, Admiraal-Behloul F, OlofsenH, Bollen EL, Jolles J, Murray HM, Blauw GJ, Westendorp RG, et al: Increasein periventricular white matter hyperintensities parallels decline inmental processing speed in a non-demented elderly population. J NeurolNeurosurg Psychiatry 2006, 77(2):149–153.39. Albinet CT, Boucard G, Bouquet CA, Audiffren M: Processing speed andexecutive functions in cognitive aging: How to disentangle their mutualrelationship? Brain Cogn 2012, 79(1):1–11.40. Carpenter PA, Just MA, Reichle ED: Working memory and executivefunction: evidence from neuroimaging. Curr Opin Neurobiol 2000,10(2):195–199.41. Gouw A, Van Der Flier W, Van Straaten E, Pantoni L, Bastos-Leite A, InzitariD, Erkinjuntti T, Wahlund L, Ryberg C, Schmidt R: Reliability and sensitivityof visual scales versus volumetry for evaluating white matterhyperintensity progression. Cerebrovasc Dis 2008, 25(3):247–253.42. Tiehuis A, Vincken K, Mali WPTM, Kappelle L, Anbeek P, Algra A, Biessels G:Automated and visual scoring methods of cerebral white matterhyperintensities: relation with age and cognitive function. Cerebrovasc Dis2008, 25(1–2):59–66.43. Sajja BR, Datta S, He R, Mehta M, Gupta RK, Wolinsky JS, Narayana PA:Unified approach for multiple sclerosis lesion segmentation on brainMRI. Ann Biomed Eng 2006, 34(1):142–151.44. Debette S, Markus HS: The clinical importance of white matterhyperintensities on brain magnetic resonance imaging: systematicreview and meta-analysis. BMJ 2010, 341:c3666.45. Abdul Latif L, Daud Amadera JE, Pimentel D, Pimentel T, Fregni F: Samplesize calculation in physical medicine and rehabilitation: a systematicreview of reporting, characteristics, and results in randomized controlledtrials. Arch Phys Med Rehabil 2011, 92(2):306–315.doi:10.1186/1471-2377-12-126Cite this article as: Bolandzadeh et al.: The association betweencognitive function and white matter lesion location in older adults: asystematic review. BMC Neurology 2012 12:126.Bolandzadeh et al. BMC Neurology 2012, 12:126 Page 10 of 10http://www.biomedcentral.com/1471-2377/12/126


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items