Geert Jan Biessels1 and Yael D. Reijmer1,2

Brain Changes UnderlyingCognitive Dysfunction inDiabetes: What Can We LearnFrom MRI?Diabetes 2014;63:2244–2252 | DOI: 10.2337/db14-0348

Diabetes is associated with cognitive dysfunction andan increased risk of dementia. This article addressesfindings with brain MRI that may underlie cognitivedysfunction in diabetes. Studies in adults with type 1diabetes show regional reductions in brain volume. Inthose with a diabetes onset in childhood, these volumereductions are likely to reflect the sum of changes thatoccur during brain development and changes thatoccur later in life due to exposure to diabetes-relatedfactors. Type 2 diabetes is associated with global brainatrophy and an increased burden of small-vessel dis-ease. These brain changes occur in the context of agingand often also in relation to an adverse vascular riskfactor profile. Advanced imaging techniques detectmicrostructural lesions in the cerebral gray and whitematter of patients with diabetes that affect structuraland functional connectivity. Challenges are to furtherunravel the etiology of these cerebral complications byintegrating findings from different imaging modalitiesand detailed clinical phenotyping and by linking struc-tural MRI abnormalities to histology. A better under-standing of the underlying mechanisms is necessaryto establish interventions that will improve long-termcognitive outcomes for patients with type 1 and type 2diabetes.

Interest in the effect of diabetes on the brain is growing.It is now clear that type 1 diabetes is associated withmodest decrements in cognitive functioning, which aremost marked in patients with an early childhood diabetesonset (1). These decrements in adults with type 1 diabetes

are most evident in the domains of general intelligence,psychomotor speed, and mental flexibility (2). On thesedomains, the magnitude of the decrements is ;0.3 to 0.7SD units relative to people without diabetes (2). Thisimplies that, on average, the performance of peoplewith diabetes on these domains is at the 30th to the40th percentile of control values. The progression of cog-nitive decrements in adults with type 1 diabetes, relativeto people without diabetes, is generally slow, except insubgroups of patients with marked microvascular compli-cations, who may show more marked decline (1).

Modest decrements in cognitive functioning, evidenton the domains of verbal and visual memory, informationprocessing speed, and executive functioning, have alsobeen noted in people with type 2 diabetes across all agegroups (3). Similar to type 1 diabetes, effect sizes aresmall to moderate (0.3 to 0.4 SD units) (4) and followa slow progression over time, only modestly exceeding therate of normal aging-related cognitive decline (3). In olderpeople, however, particularly older than the age of 65,type 2 diabetes is also associated with more severe formsof cognitive impairment. Data from large epidemiologicalsurveys link diabetes to an increased dementia risk. Ameta-analysis estimated that people with type 2 diabeteshave a relative risk of vascular dementia of 2.5 (95% CI2.1–3.0) and that of Alzheimer disease is 1.5 (95% CI1.2–1.8) relative to individuals without diabetes (5). Toprevent the progression of subtle cognitive decrements indementia in patients with diabetes, we need to develop anunderstanding of the causative mechanisms at the earli-est stages of cognitive decline.

1Department of Neurology, Brain Center Rudolf Magnus, University Medical Cen-ter, Utrecht, the Netherlands2J. Philip Kistler Stroke Research Center, Massachusetts General Hospital, Har-vard Medical School, Boston, MA

Corresponding author: Geert Jan Biessels, g.j.biessels@umcutrecht.nl.

Received 27 February 2014 and accepted 30 March 2014.

© 2014 by the American Diabetes Association. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

2244 Diabetes Volume 63, July 2014

DIA

BETES

SYMPOSIU

M

This article addresses brain changes that may underliecognitive dysfunction in adults with diabetes, focusing onMRI. We will address the effect of diabetes on brainvolumes, markers of small-vessel disease (SVD), andon structural and functional connectivity. We will alsoconsider methodological aspects that may help to in-terpret the existing data and support the design of futurestudies.

BRAIN VOLUMES

Brain volumes can be assessed with various methods,ranging from visual rating scales to automated segmen-tation methods. Visual rating scales are tolerant todifferences in image quality and scan protocol, but theirdownside is that they are operator dependent and pro-vide only a qualitative and rather insensitive measure ofatrophy (6). Automated segmentation methods, includ-ing voxel-based morphometry (VBM), probabilistic tissueclassification, and surface-based parcellation techniques,offer an operator independent and quantitative assessmentof brain volumes (6). Furthermore, VBM and surface-based parcellation techniques allow for investigationof local differences in brain anatomy by comparing vol-ume measures across brains at every voxel. Because quan-titative techniques are more sensitive to detect subtledifferences in gray or white matter tissue, they are pre-ferred methods to measure brain volume in patients withdiabetes (7).

Automated segmentation techniques, however, also havesome limitations. The accuracy of the automated tissue seg-mentation depends on the image contrast and thereforeon the quality of the MRI scans. Brain images acquiredfrom different scanners or with different scan parameterswill thus lead to variation in volume estimates. A secondlimitation is that VBM comparison methods require spatialnormalization and registration of the brain to a standardtemplate. The degree of spatial normalization or smoothingcan greatly affect the results. If one averages over largerregions, the effect of misregistration is reduced, but toomuch smoothing may mask local volume loss. Finally, whenanalyzing across multiple voxels, statistical correctionsshould be applied to limit the chance of finding false-positive effects (i.e., type I error). Researchers usually tendto choose multiple correction methods that have highpower to detect differences, such as the false discoveryrate method, at the costs of increasing the risk for type Ierrors. These processing steps should be considered wheninterpreting the below-mentioned findings in patientswith diabetes and may also explain some of the discrep-ancies between studies.

Type 1 DiabetesCross-sectional case-control studies in adults (20–60years) with type 1 diabetes have used automated seg-mentation methods to demonstrate modest reductionsin brain volume compared with control subjects (8–11).One study reported an overall 7% reduction in gray

matter volume (8), but most studies do not reportoverall differences in brain volumes and only show re-gional differences in gray matter volume or densitywithin the frontal (8–11), posterior (9,12), and temporalcortex (9) and in subcortical gray matter (10). The exactlocation of the alterations in brain volume varies greatlyacross studies, which highlights the difficulty of detect-ing small regional differences with techniques that aresensitive to interindividual differences in head size andshape.

Two of the above-mentioned studies directly relatedalterations in brain volume to cognitive functioning.Results showed an association between decreased brainvolume and slowing of information processing speed,reduced attention, and lower IQ subtest scores (12,13).Furthermore, smaller brain volumes in patients were re-lated with poor metabolic control, reflected by chronichyperglycemia (9,12) and hypoglycemic events (9,10).However, not all studies observed these associations,even with large sample sizes (n . 100) (8).

Importantly, type 1 diabetes commonly has its onset inchildhood or adolescence, when the brain is still devel-oping. Brain volume reductions in adults may thus reflectchanges that already occurred during brain development.Studies in children with type 1 diabetes indeed showthat alterations in brain volume are already detectablein childhood (14). This may be due to the interaction ofdiabetes with processes such as neuronal pruning andneuronal growth, resulting in regions of decreased aswell as increased brain volume (14). That smaller brainvolumes are observed in adults with early onset of diabe-tes relative to those with a later onset supports the viewthat childhood brain changes leave their fingerprintup until adulthood (12,13). Longitudinal imaging studiesare needed to disentangle the effects of abnormal braindevelopment from atrophy as a result of exposure todiabetes-related factors later in life.

Type 2 DiabetesType 2 diabetes has consistently been associated withglobal brain atrophy in cross-sectional studies (Fig. 1)(15). The reduction in mean total brain volume is 0.2–0.6 SD units, comparable with 3–5 years of normal aging(16–18). Longitudinal case-control and population-basedstudies have demonstrated brain volume loss in patientswith diabetes that is similar to or up to three times theatrophy rate of normal aging (16–19). The loss of braintissue is most clearly reflected by accelerated expansion ofthe ventricles (16,17,19). These findings may indicate thattype 2 diabetes–associated atrophy is most pronounced inregions surrounding the ventricles, such as subcortical graymatter or white matter regions. Another possible explana-tion is that the ventricles are less sensitive to segmentationerrors due to the relatively high signal contrast and smoothborder between brain tissue and cerebrospinal fluid. Thiscan also explain why global brain volume loss is oftendetected in these regions.

diabetes.diabetesjournals.org Biessels and Reijmer 2245

Given the association between diabetes and Alzheimerdisease, much attention has been directed to the hippo-campus. Studies that used manual and automated seg-mentation methods to assess hippocampal volume in type2 diabetes found that hippocampal volumes of patients areindeed smaller than those of age-matched control partic-ipants (20). However, a pooled analysis of three cohortstudies showed that the degree of hippocampal volumeloss in type 2 diabetes is comparable to the degree of totalbrain volume loss (Wisse et al., unpublished observations),indicating that the hippocampus is not more severely af-fected than the rest of the brain. Furthermore, studieshave found little evidence that brain atrophy specificallycontributes to memory deficits in diabetes. Most cross-sectional studies report an association between brain vol-ume and executive functioning or processing speed (21).One longitudinal study found an association betweenbaseline brain atrophy and decline in immediate memorybut not in delayed memory (18). Two studies directly ex-amined whether brain volume mediated the relationshipbetween type 2 diabetes and cognitive dysfunction andreported contrasting results: gray matter volume signifi-cantly mediated the relationship with executive function-ing and memory in one study (22), but in another study,between-group differences in cognition were largely inde-pendent of MRI markers of brain atrophy (19). Smallerbrain volumes in diabetes have been associated withgreater insulin resistance and longer diabetes duration(23,24), suggesting that long-term exposure to diabetes-related risk factors is especially harmful to the brain.

In conclusion, type 1 and type 2 diabetes are bothassociated with alterations in brain volume. However, it isimportant to note that there is as yet no definite proofthat (patterns of) atrophy relates to specific etiologies.Pathological processes leading to brain atrophy areheterogeneous and not necessarily indicative of neuronal

loss (6). Loss of glial cells and axons, white matter rare-faction and shrinkage, and arteriolosclerosis and venouscollagenosis may all produce changes in brain volume. Tobetter understand the pathological process of brain atro-phy in diabetes, imaging and histopathological findingsshould be integrated. Techniques are now being developedthat can register blocks of serially stained histologicalsections to postmortem brain MRI scans and will allowus in the near future to directly relate local brain imagingfindings to their pathophysiological substrates.

SVD

Cerebral SVD refers to pathological processes that affectthe small arteries, arterioles, venules, and capillaries ofthe brain in the context of aging and vascular riskfactors. Brain MRI is a powerful tool to detect SVD,although we note that MRI shows the consequences ofSVD in the brain tissue (i.e., parenchymal lesions) ratherthan abnormalities in the small vessels themselves. Aninternational working group has recently proposedstandardized definitions for the core MRI features ofSVD, including:

� lacunes, which are round or ovoid, subcortical, fluid-filled cavities 3–15 mm in diameter, compatible witha previous acute small deep-brain infarct or hemor-rhage in the territory of a single perforating arteriole;

� white matter hyperintensities (WMH), noted as signalabnormalities of variable size in the white matter thatare hyperintense on fluid-attenuated inversion recoveryand T2/proton density-weighted images without cavi-tation (Fig. 2); and

� cerebral microbleeds, characterized as small, mostly2–5 mm areas of signal void with associated “blooming”on T2* or other MR sequences sensitive to paramag-netic material (Fig. 2) (6).

Figure 1—VBM analysis was used to create a probability map of areas of gray matter atrophy (highlighted voxels) in patients with type 2diabetes when compared with control subjects. A: Inferior region. B: Temporal region. C: Medial region. D: Superior region. Reproducedwith permission from Moran et al. (22), copyright 2013 by the American Diabetes Association (for details, see original paper).

2246 Brain MRI Findings and Cognitive Dysfunction Diabetes Volume 63, July 2014

The presence of lacunes and microbleeds is generally ratedvisually. Presence and severity of WMH can also beassessed visually, with ordinal rating scales, or with(semi)automated volumetric methods (6). The introduc-tion of high-field-strength MRI now allows for a detectionof even smaller SVD lesions, in particular microinfarcts(25), which may also be relevant to diabetes. In addition,whole-body MRI allows for the assessment of central aswell as peripheral vascular MRI abnormalities, which mayprovide a more complete estimate of vascular health andthe burden of vascular disease in diabetes (26).

The etiology of the different SVD lesions on MRI isheterogeneous. Lacunes are most often due to occlusivedisease of a perforating arteriole but can also be due to anembolus. Postmortem studies have shown that WMHreflect tissue abnormalities that range from slight disen-tanglement of the white matter structure to varyingdegrees of myelin and axonal loss (27). The etiologyincludes ischemia, hypoperfusion, blood-brain barrierleakage, inflammation, degeneration, and amyloid angio-pathy (27). Postmortem validation of MRI lesions sugges-tive of microbleeds shows hemosiderin-laden macrophagesconsistent with previous vascular leakage of blood cells(6,27). The underlying vascular pathology most com-monly involves hypertensive vasculopathy or cerebralamyloid angiopathy (27). The heterogeneity in etiologiesof SVD lesions on MRI is important to consider wheninterpreting findings from studies in patients with dia-betes. It implies that underlying mechanisms cannot bereliably inferred from lesion patterns or severity.

Type 1 DiabetesIn light of the increasing number of papers on brain MRIin patients with type 1 diabetes, that there are only a fewreports on SVD is surprising. It could well be that severalstudies assessed SVD but did not report negative findings.Indeed, studies that have reported WMH in patients with

type 1 diabetes observed no difference on visual ratingscales compared with control subjects (28,29). Moreover,although type 1 diabetes is an established risk factor forlacunar stroke (30), there are limited data on the occur-rence of lacunes on brain MRI in people with type 1 di-abetes (29). One recent study on microbleeds reported nodifference in the overall occurrence of these lesions rela-tive to control subjects in a small cohort of patients withtype 1 diabetes but did observe an increased occurrence inthe subgroup of patients with proliferative diabetic reti-nopathy (31).

It is important to note that all but one (29) of theavailable MRI studies on SVD in type 1 diabetes predom-inantly involved patients younger than 50 years of age.Because SVD on MRI is quite uncommon in young adults,definite insights into the relationship between type 1diabetes and SVD will require studies with sufficient sta-tistical power, preferentially also involving older partici-pants. Nevertheless, on the basis the currently availabledata, MRI markers of SVD do not appear to be a keydeterminant of type 1 diabetes–associated cognitive dec-rements in young adults.

Type 2 DiabetesType 2 diabetes is clearly associated with the occurrenceof lacunes on MRI. A meta-analysis reported odds ratiosfor lacunes in patients relative to control subjects thatvaried from 1.3 to 2.1, depending on the source of thestudy cohort (e.g., hospital- or population-based) (15).The relation between type 2 diabetes and WMH is morecontroversial. Several studies that assessed WMH vol-umes with (semi)automated methods reported no differ-ence between patients with type 2 diabetes and controlsubjects in cross-sectional (18,22,32) or longitudinal anal-yses (18). Other studies reported that WMH volumeswere increased in patients with diabetes (7,17,19,33). Im-portantly, in these latter studies, the relative increase inWMH volumes compared with control subjects was;20%, which is only a modest difference in light of themarked interindividual variation in WMH volumes inolder people. Taken together, it can therefore be con-cluded that type 2 diabetes does not have a major effecton WMH volumes. The same appears to be true for micro-bleeds. Two recent studies did not observe a relation be-tween type 2 diabetes and microbleed occurrence (22,33),although the proportion of individuals with more thanone microbleed was higher among people with type 2 di-abetes (33).

MRI markers of SVD, in particular WMH and infarcts,have been related to slowing of processing speed anddecrements in attention and executive functioning inpatients with type 2 diabetes (21), and progression ofWMH has been linked to accelerated cognitive decline(34). Some studies observe that diabetes-associated cog-nitive decline is independent of SVD burden (18), whereasothers report slight (19) or evident mediation of SVD(infarcts, WMH, but not microbleeds) in the associations

Figure 2—Markers of SVD on MRI. The image on the left showsa T2-weighted fast field echo image with a microbleed (large open△) and blood vessels that can mimic microbleeds (small open △).The image on the right shows a fluid-attenuated inversion recoveryimage with periventricular (small closed△) and deep (large closed△) WMH.

diabetes.diabetesjournals.org Biessels and Reijmer 2247

of diabetes with memory, processing speed, and executivefunctioning (33).

In summary, type 2 diabetes is associated with anincreased burden of SVD, and MRI markers of SVD areassociated with cognitive dysfunction in people with type 2diabetes. The relation between SVD and cognitive dysfunc-tion is, however, not specific to diabetes, and the extentto which SVD mediates the link between diabetes andcognitive dysfunction remains to be determined.

STRUCTURAL AND FUNCTIONAL CONNECTIVITY

Structural ConnectivityDiffusion tensor imaging (DTI) is an MRI technique thatcan quantify alterations in the white matter tissue ata microscopic scale by characterizing the diffusion ofwater molecules within the brain (35). Damage to tissuestructure caused by, for example, demyelination or axonalatrophy will lead to alterations in the diffusion of watermolecules reflected by a change in mean diffusivity (MD)and fractional anisotropy (FA). As such, DTI can detectmicrostructural white matter abnormalities that are notvisible on conventional MRI scans. A limitation of DTI isthat tissue properties unrelated to pathology, but due to,for example, the organization of axonal fibers, will alsoaffect diffusion parameters.

The directionality of the diffusion obtained by DTIallows the creation of maps of white matter tract anatomyand study of the connectivity between brain regions (Fig.3). More recently, complex network theory has been usedto assess the organization of whole-brain white matterconnections (36). Network analysis is a promising tech-nique to study the structural basis of cognitive functionsthat rely on the interaction between widely distributed

brain regions, such as executive functioning and informa-tion processing speed, cognitive functions that are pref-erentially affected in diabetes

Type 1 DiabetesTwo studies examined microstructural white matteralterations with DTI in middle-aged adults with type 1diabetes (37,38). Results showed decreased FA in poste-rior brain regions compared with control subjects (37).More widespread reductions in FA throughout the brainwere found in patients with diabetes and microangiop-athy (38). In these studies, lower FA in patients was as-sociated with longer diabetes duration (37) and decreasedinformation processing speed and executive functioning(37,38).

Type 2 DiabetesAn increasing number of studies have examined DTIparameters in type 2 diabetes. Most studies observeda global decrease in FA or increase in MD that could notbe explained by differences in vascular lesion load or totalbrain volume (32,39–41). Disruption of white matter con-nections in the temporal lobe was specifically associatedwith decreased memory performance (39,41), whereasincreased MD in frontal and temporal and posterior fibertracts was related to reduced information processingspeed (39). These findings are in line with current theo-ries about the localization of higher cognitive functions.Interestingly, diffusion alterations have also been observedin patients with diabetes who performed cognitively sim-ilar to control subjects (39,40) and in individuals withthe metabolic syndrome (42), suggesting that diabetes-associated white matter abnormalities occur early in thedisease process.

So far, only one study has used network analysis toexamine differences in whole-brain connectivity patterns.Results showed decreased measures of local and globalconnectivity in patients compared with age-matchedcontrol subjects. Network alterations were related toslower information processing in patients, independentof age, sex, education, WMH, and lacunar infarcts (36).

It is not clear what pathological condition underliesthe alterations in white matter diffusion parametersin diabetes. Possible mechanisms include inflammation,microvascular lesions, enlarged perivascular spaces, andblood-brain barrier disruption (43). In summary, DTI haslow specificity but high sensitivity to white matter pathol-ogy. In patients with diabetes, diffusion measures showrelatively strong and robust correlations with cognition.For these reasons, DTI measures have high potential asan MRI outcome marker to track disease progression inclinical trials.

Functional ConnectivityFunctional brain imaging allows researchers to bridge thegap between variations in brain structure and cognitiveperformance by assessing measures related to neuronalactivity. Imaging modalities that are used to measure

Figure 3—Reconstruction of four major fiber bundles is shown withdiffusion-based fiber tractography (red, uncinate fasciculus; blue,corticospinal tract; green, inferior longitudinal fasciculus; yellow,fornix). Fiber tractography can be used to study the structural con-nectivity between brain regions. Breakdown of connectivity withincertain networks may explain some of the cognitive deficits ob-served in patients with diabetes (36,39).

2248 Brain MRI Findings and Cognitive Dysfunction Diabetes Volume 63, July 2014

brain function include functional MRI (fMRI), electroen-cephalography (EEG), and magnetoencephalography (MEG).Most studies in diabetes have used fMRI, a techniquethat estimates regional increases in the blood-oxygen-level–dependent (BOLD) contrast in response to neuronalactivity. By examining the correlation in BOLD activitybetween brain areas, studies have identified large-scalefunctional networks that are spontaneously active duringrest. These resting-state networks strongly overlap withtask-related networks and are compromised in various cog-nitive disorders, including mild cognitive impairment andAlzheimer disease (44). The interest in resting-state fMRIin the field of dementia is rapidly growing, and studiescontinue to demonstrate the relevance of these techniquesfor detecting functional brain changes in Alzheimer diseaseas well as in vascular cognitive impairment.

It should be noted, however, that the link between theBOLD response and neural activity (i.e., neurovascularcoupling) may be altered in patients with cerebrovasculardisease. An attenuated BOLD response in patients withdiabetes may thus be caused by impaired vascular re-activity and does not directly imply impaired neuronalactivity. For this reason, the assessment of fMRI inpatients with cerebrovascular disease is often combinedwith the assessment of cerebral vasoreactivity by usinga breath-holding test or CO2 inhalation test. Alternatively,neuronal activity can be measured more directly with EEGor MEG. The studies below have not combined thesetechniques, which should be considered when interpretingtheir results.

Type 1 DiabetesOnly a few studies have used fMRI to examine resting-state functional connectivity in patients with type 1diabetes. One study found reduced functional connectiv-ity in the ventral attention network in patients comparedwith control subjects, but only in patients with micro-angiopathy (45). Patients with diabetes without microan-giopathy were not significantly different from controlsubjects. Evidence for altered functional connectivity intype 1 diabetes also comes from neurophysiological stud-ies using EEG and MEG (46,47). EEG and MEG measuresdo not rely on neurovascular coupling; thus, these reportssupport the hypothesis that neuronal communication indiabetes is altered. However, the results of these twostudies vary considerably with respect to affected brainregions and frequency bands.

Type 2 DiabetesIn patients with type 2 diabetes, reduced functionalconnectivity has been observed between regions of theso-called default mode network, including the medialfrontal gyrus, precuneus, and medial temporal gyrus (48).The default mode network regions are among the mosthighly connected regions in the brain, suggesting an im-portant role in global cognitive processing. Indeed, im-paired default mode network connectivity was related toreduced memory, executive functioning, and processing

speed in older individuals free from Alzheimer pathology(44). Abnormalities in resting-state brain activity havebeen observed in patients with type 2 diabetes, with de-creased low frequency fluctuations in the postcentral gy-rus and occipital cortex (49). Decreased brain activity inthese areas was observed in the absence of structuralbrain changes and was related to worse memory perfor-mance and executive functioning.

Together, these findings clearly indicate that diabetesis associated with alterations in structural as well asfunctional brain connectivity. The next step is to integratefindings from different imaging modalities to determine1) whether the reduced correlations in BOLD responseare indeed due to altered neuronal activity or also reflectaltered vascular reactivity, 2) whether alterations in func-tional connectivity are a direct consequence of disruptionof white matter tracts, and 3) which processes underliedisruption of these tracts. To achieve this, standardiza-tion of processing methods across studies are needed aswell as more adequate sample sizes to prevent type I andtype II errors.

IMPLICATIONS AND FUTURE PERSPECTIVES

Type 1 and type 2 diabetes are both linked to abnormal-ities on brain MRI that are likely to underlie diabetes-associated cognitive decrements. Brain MRI abnormalitiesthat are associated with type 1 diabetes occur ratherdiffusely throughout the brain and are reflected in subtleregional reductions in brain volume and changes inconnectivity (Fig. 4). Focal vascular lesions are not a keyMRI feature of type 1 diabetes in young adults. Uncer-tainty still exists regarding the etiology of brain MRI ab-normalities in type 1 diabetes and their significance forlong-term cognitive decline. In this respect, an importantlimitation of most of the available studies is their cross-sectional design. Part of the MRI abnormalities in adultswith type 1 diabetes may originate from the effects ofdiabetes on brain development in childhood. To iden-tify potentially modifiable risk factors for progressivebrain volume loss (i.e., atrophy) in adulthood, longitu-dinal studies are clearly needed. Such studies should usequantitative volumetric techniques. Inclusion of a non-diabetic reference group, to be able to identify diabetes-attributable changes on top of those of normal aging, isstrongly recommended.

Brain MRI abnormalities that are associated with type 2diabetes include global cortical and subcortical atrophy,with some variation in severity across brain regions, andan increased burden of SVD, in particular lacunar infarcts(Fig. 5). SVD is likely to contribute to abnormalities infunctional and structural connectivity, which is anotherMRI feature of type 2 diabetes that is clearly linked tocognitive dysfunction. Brain abnormalities and associatedcognitive dysfunction in type 2 diabetes occur in the con-text of brain aging. The challenge is to disentangle theeffects of diabetes—and diabetes-associated vascular riskfactors—from those of “normal aging.” The different brain

diabetes.diabetesjournals.org Biessels and Reijmer 2249

MRI features of type 2 diabetes are unlikely to share thesame etiology. On the basis of current insights, chronicexposure to hyperglycemia, insulin resistance, and vascularrisk factors—most importantly hypertension—are potentiallymodifiable risk factors, but other factors, including inflam-mation, oxidative stress, hypoperfusion, and Alzheimer-type degenerative processes, may also play a role (3).

An open question is how diabetes-related brain abnor-malities can be prevented. In this context it is importantto mention the Action to Control Cardiovascular Risk inDiabetes-Memory In Diabetes (ACCORD-MIND) study,which showed that after 40 months, the rate of brainatrophy on MRI was reduced in patients with type 2diabetes receiving intensive glycemic control comparedwith those receiving standard therapy (50). Unfortu-nately, this positive effect on brain structure was notaccompanied by improved cognitive functioning. On onehand, these observations raise the question of whetherthere could be a cerebral “legacy effect” where preserva-tion of cerebral structure leads to improved long-termcognitive outcomes, in particular delaying or preventingdementia. On the other hand, the latest report from theACCORD-MIND study shows that in many aspects, long-term preservation of brain health in diabetes is still anenigma: intensive blood pressure control in patients withdiabetes was accompanied by an accelerated rate of brainatrophy compared with standard therapy (51). It seemsthat what is good for the heart is not necessarily goodfor the brain.

Owing to the subtle nature of diabetes-associated MRIabnormalities, they are currently only detectable at thegroup level. In other words, there is no typical MRIsignature of type 1 or type 2 diabetes that can beidentified in an individual patient. As such, there is noplace for brain MRI in the periodic evaluation of patientswith diabetes in current clinical practice. However, asshown in this review, brain MRI has clearly established itsvalue as a research tool to further unravel the trajectoriesof brain changes, identify the primary underlying etiolo-gies, and develop treatment. To fully benefit from thepotential that brain MRI has to offer in this context,rigorous methodology is required in study design and alsoin image acquisition and analysis. Multimodal imagingapproaches are recommended, where different structuraland functional imaging techniques are combined, wherepossible also complimented by histological studies, toverify the true nature and etiology of imaging abnor-malities. Apart from this, rigorous clinical phenotypingof patients is essential, with careful documentation oflong-term functional outcomes and detailed and precisedocumentation of diabetes-related variables and otherrisk factors.

Regarding trajectories of brain changes and cognitiveoutcomes, it would be a major advance if we couldpinpoint those imaging abnormalities that are predictiveof long-term poor cognitive outcome in individualpatients, because this would help to select patients whoare likely to benefit most from novel therapies. Hope-fully, brain imaging can also help to identify etiologicalprocesses at an individual level in the near future thatmay guide the selection of the appropriate therapy. Anexample of this in the field of Alzheimer disease is

Figure 4—Trajectories of brain MRI abnormalities in patients withtype 1 diabetes (T1DM). T1DM is associated with brain volume loss(blue line, evolution of brain volumes with age in general population;red line, estimated trajectories for T1DM) (A) and loss of connectiv-ity (B). In patients with childhood-onset diabetes, brain abnormali-ties in adulthood reflect the sum of changes that have occurredduring brain development in childhood (reproduced with permissionfrom Marzelli et al. [14]; copyright 2014 by the American DiabetesAssociation) (1), and changes that occur later in life under the in-fluence of diabetes-related factors, particularly in vulnerable sub-groups of patients such as those with advanced microvascularcomplications (retinal image courtesy of M.K. Ikram, National Uni-versity of Singapore) (2).

Figure 5—Trajectories of brain MRI abnormalities in patients withtype 2 diabetes. Type 2 diabetes is associated with brain atrophy(blue line, evolution of brain volumes with age in general population;red line, estimated trajectories for type 2 diabetes) (A), loss of con-nectivity (B), and an increased burden of SVD (shaded blue area,SVD in general population with age; red area, SVD in patients withtype 2 diabetes) (C). Brain MRI abnormalities are likely to start inprediabetic stages (dotted red line), probably under the influence ofan adverse vascular risk factor profile.

2250 Brain MRI Findings and Cognitive Dysfunction Diabetes Volume 63, July 2014

amyloid positron emission tomography imaging, whichhelps to identify people at high risk of cognitive declinedue to Alzheimer-type pathologies at an early stage of thedisease when intervention may be most effective.

Funding. The research of G.J.B. is supported by Vidi grant 91711384 fromZonMw, The Netherlands Organisation for Health Research and Development,and grant 2010T073 from the Netherlands Heart Foundation.Duality of Interest. G.J.B. consults for and receives research support fromBoehringer Ingelheim. No other potential conflicts of interest relevant to thisarticle were reported.Author Contributions. G.J.B. and Y.D.R. determined the structure of thereview and each wrote separate sections of the manuscript. G.J.B. compiled thesections. G.J.B. and Y.D.R. critically reviewed and edited the full manuscript. G.J.B.is the guarantor of this work and, as such, had full access to all the data in thestudy and takes responsibility for the integrity of the data and the accuracy of thedata analysis.Prior Presentation. Parts of this study were presented at the 74th Sci-entific Sessions of the American Diabetes Association, San Francisco, CA, 13–17June 2014.

Appendix

Literature Search Strategy and Selection

We searched PubMed from 2000 to February 2014 with the terms (andsynonyms) “diabetes,” “brain,” “white matter,” “cortex,” “hippocampus,” “MRI,”“magnetic resonance,” “atrophy,” “diffusion tensor imaging,” “fMRI,” “smallvessel disease,” “lacunar infarct,” and “microbleed.” We also extracted relevantpapers from our records. We only considered studies on human subjects—particularly involving adults—published in English.

References1. Biessels GJ, Deary IJ, Ryan CM. Cognition and diabetes: a lifespan per-spective. Lancet Neurol 2008;7:184–1902. Brands AM, Biessels GJ, de Haan EH, Kappelle LJ, Kessels RP. The effectsof type 1 diabetes on cognitive performance: a meta-analysis. Diabetes Care2005;28:726–7353. Biessels GJ, Strachan MW, Visseren FL, Kappelle LJ, Whitmer RA. Dementiaand cognitive decline in type 2 diabetes and prediabetic stages: towards targetedinterventions. Lancet Diabetes Endocrinol 2014;2:246–2554. Palta P, Schneider AL, Biessels GJ, Touradji P, Hill-Briggs F. Magnitude ofcognitive dysfunction in adults with type 2 diabetes: a meta-analysis of sixcognitive domains and the most frequently reported neuropsychological testswithin domains. J Int Neuropsychol Soc 2014;20:278–2915. Cheng G, Huang C, Deng H, Wang H. Diabetes as a risk factor for dementiaand mild cognitive impairment: a meta-analysis of longitudinal studies. InternMed J 2012;42:484–4916. Wardlaw JM, Smith EE, Biessels GJ, et al.; STandards for ReportIng Vascularchanges on nEuroimaging (STRIVE v1). Neuroimaging standards for research intosmall vessel disease and its contribution to ageing and neurodegeneration. LancetNeurol 2013;12:822–8387. Jongen C, van der Grond J, Kappelle LJ, Biessels GJ, Viergever MA, PluimJP; Utrecht Diabetic Encephalopathy Study Group. Automated measurement ofbrain and white matter lesion volume in type 2 diabetes mellitus. Diabetologia2007;50:1509–15168. Hughes TM, Ryan CM, Aizenstein HJ, et al. Frontal gray matter atrophy inmiddle aged adults with type 1 diabetes is independent of cardiovascular riskfactors and diabetes complications. J Diabetes Complications 2013;27:558–5649. Musen G, Lyoo IK, Sparks CR, et al. Effects of type 1 diabetes on graymatter density as measured by voxel-based morphometry. Diabetes 2006;55:326–333

10. Northam EA, Rankins D, Lin A, et al. Central nervous system function inyouth with type 1 diabetes 12 years after disease onset. Diabetes Care 2009;32:445–45011. Frøkjær JB, Brock C, Søfteland E, et al. Macrostructural brain changes inpatients with longstanding type 1 diabetes mellitus - a cortical thickness analysisstudy. Exp Clin Endocrinol Diabetes 2013;121:354–36012. Franc DT, Kodl CT, Mueller BA, Muetzel RL, Lim KO, Seaquist ER. Highconnectivity between reduced cortical thickness and disrupted white mattertracts in long-standing type 1 diabetes. Diabetes 2011;60:315–31913. Ferguson SC, Blane A, Wardlaw J, et al. Influence of an early-onset age oftype 1 diabetes on cerebral structure and cognitive function. Diabetes Care 2005;28:1431–143714. Marzelli MJ, Mazaika PK, Barnea-Goraly N, et al.; Diabetes Research inChildren Network (DirecNet). Neuroanatomical correlates of dysglycemia in youngchildren with type 1 diabetes. Diabetes 2014;63:343–35315. van Harten B, de Leeuw FE, Weinstein HC, Scheltens P, Biessels GJ. Brainimaging in patients with diabetes: a systematic review. Diabetes Care 2006;29:2539–254816. de Bresser J, Tiehuis AM, van den Berg E, et al.; Utrecht Diabetic Enceph-alopathy Study Group. Progression of cerebral atrophy and white matter hyper-intensities in patients with type 2 diabetes. Diabetes Care 2010;33:1309–131417. Kooistra M, Geerlings MI, Mali WP, Vincken KL, van der Graaf Y, BiesselsGJ; SMART-MR Study Group. Diabetes mellitus and progression of vascular brainlesions and brain atrophy in patients with symptomatic atherosclerotic disease.The SMART-MR study. J Neurol Sci 2013;332:69–7418. van Elderen SG, de Roos A, de Craen AJ, et al. Progression of brain atrophy

and cognitive decline in diabetes mellitus: a 3-year follow-up. Neurology 2010;75:997–100219. Espeland MA, Bryan RN, Goveas JS, et al.; WHIMS-MRI Study Group. In-fluence of type 2 diabetes on brain volumes and changes in brain volumes:results from the Women’s Health Initiative Magnetic Resonance Imaging studies.Diabetes Care 2013;36:90–9720. Gold SM, Dziobek I, Sweat V, et al. Hippocampal damage and memoryimpairments as possible early brain complications of type 2 diabetes. Dia-betologia 2007;50:711–71921. Manschot SM, Brands AM, van der Grond J, et al.; Utrecht Diabetic En-cephalopathy Study Group. Brain magnetic resonance imaging correlates of im-paired cognition in patients with type 2 diabetes. Diabetes 2006;55:1106–111322. Moran C, Phan TG, Chen J, et al. Brain atrophy in type 2 diabetes: regionaldistribution and influence on cognition. Diabetes Care 2013;36:4036–404223. Saczynski JS, Siggurdsson S, Jonsson PV, et al. Glycemic status and braininjury in older individuals: the Age Gene/Environment Susceptibility-ReykjavikStudy. Diabetes Care 2009;32:1608–161324. Tan ZS, Beiser AS, Fox CS, et al. Association of metabolic dysregulation withvolumetric brain magnetic resonance imaging and cognitive markers of sub-clinical brain aging in middle-aged adults: the Framingham Offspring Study.Diabetes Care 2011;34:1766–177025. van Veluw SJ, Zwanenburg JJ, Engelen-Lee J, et al. In vivo detection ofcerebral cortical microinfarcts with high-resolution 7T MRI. J Cereb Blood FlowMetab 2013;33:322–32926. Bamberg F, Parhofer KG, Lochner E, et al. Diabetes mellitus: long-termprognostic value of whole-body MR imaging for the occurrence of cardiac andcerebrovascular events. Radiology 2013;269:730–73727. Gouw AA, Seewann A, van der Flier WM, et al. Heterogeneity of small vesseldisease: a systematic review of MRI and histopathology correlations. J NeurolNeurosurg Psychiatry 2011;82:126–13528. Weinger K, Jacobson AM, Musen G, et al. The effects of type 1 diabetes oncerebral white matter. Diabetologia 2008;51:417–42529. Brands AM, Kessels RP, Hoogma RP, et al. Cognitive performance, psy-chological well-being, and brain magnetic resonance imaging in older patientswith type 1 diabetes. Diabetes 2006;55:1800–1806

diabetes.diabetesjournals.org Biessels and Reijmer 2251

30. Luitse MJ, Biessels GJ, Rutten GE, Kappelle LJ. Diabetes, hyperglycaemia,and acute ischaemic stroke. Lancet Neurol 2012;11:261–27131. Woerdeman J, van Duinkerken J,, Wattjes MP, et al. Proliferative retinop-athy in type 1 diabetes is associated with cerebral microbleeds, which is part ofgeneralized microangiopathy. Diabetes Care 2014;37:1165–116832. Falvey CM, Rosano C, Simonsick EM, et al.; Health ABC Study. Macro- andmicrostructural magnetic resonance imaging indices associated with diabetesamong community-dwelling older adults. Diabetes Care 2013;36:677–68233. Qiu C, Sigurdsson S, Zhang Q, et al. Diabetes, markers of brain pathologyand cognitive function: the Age, Gene/Environment Susceptibility-ReykjavikStudy. Ann Neurol 2014;75:138–14634. Reijmer YD, van den Berg E, de Bresser J, et al.; Utrecht Diabetic En-cephalopathy Study Group. Accelerated cognitive decline in patients with type 2diabetes: MRI correlates and risk factors. Diabetes Metab Res Rev 2011;27:195–20235. Jones DK, Leemans A. Diffusion tensor imaging. Methods Mol Biol 2011;711:127–14436. Reijmer YD, Leemans A, Brundel M, Kappelle LJ, Biessels GJ; UtrechtVascular Cognitive Impairment Study Group. Disruption of the cerebral whitematter network is related to slowing of information processing speed in patientswith type 2 diabetes. Diabetes 2013;62:2112–211537. Kodl CT, Franc DT, Rao JP, et al. Diffusion tensor imaging identifies deficitsin white matter microstructure in subjects with type 1 diabetes that correlate withreduced neurocognitive function. Diabetes 2008;57:3083–308938. van Duinkerken E, Schoonheim MM, Ijzerman RG, et al. Diffusion tensorimaging in type 1 diabetes: decreased white matter integrity relates to cognitivefunctions. Diabetologia 2012;55:1218–122039. Reijmer YD, Brundel M, de Bresser J, Kappelle LJ, Leemans A, Biessels GJ;Utrecht Vascular Cognitive Impairment Study Group. Microstructural white matterabnormalities and cognitive functioning in type 2 diabetes: a diffusion tensorimaging study. Diabetes Care 2013;36:137–14440. Hsu JL, Chen YL, Leu JG, et al. Microstructural white matter abnormalitiesin type 2 diabetes mellitus: a diffusion tensor imaging study. Neuroimage 2012;59:1098–1105

41. Yau PL, Javier D, Tsui W, et al. Emotional and neutral declarative memoryimpairments and associated white matter microstructural abnormalities in adultswith type 2 diabetes. Psychiatry Res 2009;174:223–23042. Segura B, Jurado MA, Freixenet N, Falcón C, Junqué C, Arboix A. Micro-structural white matter changes in metabolic syndrome: a diffusion tensor im-aging study. Neurology 2009;73:438–44443. Nelson PT, Smith CD, Abner EA, et al. Human cerebral neuropathology oftype 2 diabetes mellitus. Biochim Biophys Acta 2009;1792:454–46944. Chhatwal JP, Sperling RA. Functional MRI of mnemonic networks across thespectrum of normal aging, mild cognitive impairment, and Alzheimer’s disease.J Alzheimers Dis 2012;31(Suppl. 3):S155–S16745. van Duinkerken E, Schoonheim MM, Sanz-Arigita EJ, et al. Resting-statebrain networks in type 1 diabetic patients with and without microangiopathy andtheir relation to cognitive functions and disease variables. Diabetes 2012;61:1814–182146. Cooray GK, Hyllienmark L, Brismar T. Decreased cortical connectivity andinformation flow in type 1 diabetes. Clin Neurophysiol 2011;122:1943–195047. van Duinkerken E, Klein M, Schoonenboom NS, et al. Functional brainconnectivity and neurocognitive functioning in patients with long-standing type 1diabetes with and without microvascular complications: a magnetoencephalographystudy. Diabetes 2009;58:2335–234348. Musen G, Jacobson AM, Bolo NR, et al. Resting-state brain functionalconnectivity is altered in type 2 diabetes. Diabetes 2012;61:2375–237949. Cui Y, Jiao Y, Chen YC, et al. Altered spontaneous brain activity in type 2diabetes: a resting-state functional MRI study. Diabetes 2014;63:749–76050. Launer LJ, Miller ME, Williamson JD, et al.; ACCORD MIND investigators.Effects of intensive glucose lowering on brain structure and function in peoplewith type 2 diabetes (ACCORD MIND): a randomised open-label substudy. LancetNeurol 2011;10:969–97751. Williamson JD, Launer LJ, Bryan RN, et al.; Action to Control CardiovascularRisk in Diabetes (ACCORD) Memory in Diabetes (MIND) Investigators. Cognitivefunction and brain structure in persons with type 2 diabetes mellitus after in-tensive lowering of blood pressure and lipid levels: a randomized clinical trial.JAMA Intern Med 2014;174:324–333

2252 Brain MRI Findings and Cognitive Dysfunction Diabetes Volume 63, July 2014