Detection, risk factors, and functional consequences of cerebral microinfarcts

Susanne J. van Veluw PhD1,2, Andy Y. Shih PhD3, Eric E. Smith MD4, Christopher Chen MD5, Prof. Julie

A. Schneider MD6, Prof. Joanna M. Wardlaw MD7, Prof. Steven M. Greenberg MD2, Prof. Geert Jan

Biessels MD1

1 Brain Center Rudolf Magnus, University Medical Center Utrecht, Utrecht, the Netherlands2 Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA3 Department of Neuroscience, Medical University of South Carolina, Charleston, SC, USA4 Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada5 Memory Ageing and Cognition Centre, National University Health System, Singapore6 Rush Alzheimer's Disease Center, Rush University Medical Center, Chicago, IL, USA7 Centre for Clinical Brain Sciences and Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, Edinburgh, United Kingdom

Word count: 4963 including Table 1 (excluding Box 1, 2, and Figure legends)

Figures: 5

Supplemental Figures: 2

Tables: 1

Boxes: 2

References: 78

Corresponding author

Geert Jan Biessels MD PhD, Brain Center Rudolf Magnus, Department of Neurology, University

Medical Center Utrecht, G03.218, P.O. Box 85500 3508 GA Utrecht, the Netherlands,

g.j.biessels@umcutrecht.nl

Abstract

Cerebral microinfarcts (CMIs) are small, presumed ischemic lesions that exist at the crossroads of

cerebrovascular disease and dementia. Although by definition tiny, CMIs number in the hundreds or

thousands in affected individuals, cause measurable disruption to structural brain connections, and

associate with dementia apparently independently of Alzheimer’s disease pathology or larger

infarcts. There is substantial recent progress in the understanding of CMIs, driven in large part by

new in vivo detection methods. Moreover, experimental models have been established that closely

mimic multiple aspects of human CMIs. Insights derived from these recent advances suggest that

CMIs can be manifestations of both small vessel and large vessel disease, that CMIs are

independently associated with cognitive impairment, and that they likely cause damage to brain

structure and function that extends beyond their actual lesion boundaries. Criteria for the

identification of CMIs with in vivo magnetic resonance imaging are provided to support further

studies into the role of CMIs in cerebrovascular disease and dementia.

1. Introduction

Cerebral microinfarcts (CMIs) are small lesions of presumed ischemic origin. Although they are often

classified as ‘not visible to the naked eye’ on autopsy, definitions on specific size cut-offs vary and

depend on different detection modalities. CMIs are a common finding at brain autopsy, particularly in

individuals with dementia and in those with other manifestations of cerebrovascular disease. (1,2) A

systematic review of neuropathological studies reported a CMI prevalence of 62% in patients with a

diagnosis of vascular dementia, 43% in Alzheimer’s disease and 24% in aged individuals who came to

autopsy without a prior diagnosis of dementia.(2) It has been estimated that CMIs likely number in

the hundreds to thousands in affected individuals. (3,4) A Lancet Neurology review in 2012 highlighted

the importance of these lesions(1), which are considered the most widespread form of brain

infarction, with impact extending beyond the tissue injury that can be appreciated at autopsy. It is

now clear that CMIs cause measurable disruption to structural brain connections, and associate with

dementia apparently independently of Alzheimer’s disease pathology. These findings suggest that

CMIs are an important mechanistic link between cerebrovascular disease and dementia. (5)

While in 2012 CMIs were still referred to as ‘the invisible lesions’(1), it has since become possible to

visualize the largest CMIs in vivo with high-resolution structural magnetic resonance imaging (MRI).

In addition, experimental models have been established to study the impact of CMIs on the brain

with unprecedented temporal and spatial resolution. This review addresses these recent

developments and provides an update on the frequency, risk factors, and functional consequences of

CMIs. We propose criteria for identification of CMIs with MRI and provide a framework to translate

candidate CMI mechanisms and outcome markers from experimental models to MRI-guided clinical

trials.

2. Detection of microinfarcts

There are three different modalities for the detection of CMIs: neuropathology, diffusion-weighted

imaging (DWI) and high-resolution structural MRI. An important consideration is that none of these

three modalities captures the entire CMI burden in the brain (Table 1). Neuropathology is currently

considered the reference standard for the diagnosis of a CMI and is able to detect the smallest acute

and chronic CMIs. However, detection by most neuropathologic methods has limited brain coverage.

At a standard brain autopsy only a few samples (usually 20 or less) are taken for processing into ~4-6

μm-thick paraffin sections for histopathological analysis; as such <0·01% of the entire brain is

sampled. It has been estimated that one or several CMIs detected at routine brain autopsy may

indicate the presence of hundreds or thousands of CMIs throughout the brain that go undetected. (3)

DWI can detect only relatively large lesions (spatial resolution 1-2 mm), but offers full brain coverage.

However, temporal resolution is limited, because DWI can only detect lesions in up to the first two

weeks after they occurred (and occasionally longer). (6,7) A recent study used mathematical models to

estimate the likely annual rates of new CMIs based on the presence of incidental DWI lesions. (4) The

calculations indicated that the detection of one or two incidental small DWI lesions - presumed to be

acute CMIs large enough to produce a DWI signal - could reflect an annual incidence of hundreds of

new CMIs of any size.(4)

Finally, it has become clear that CMIs can be detected in acute, subacute and chronic stages on

structural MRI.(8,9) Detection is dependent on spatial resolution and contrast of the scan protocol,

field strength, and degree of scrutiny.(10) Currently, CMI detection with structural MRI has only been

validated histologically for the cortical grey matter.(9,11,12) In the white matter occurrence of CMI

mimics, such as white matter hyperintensities (WMHs), and enlarged perivascular spaces (PVS)

currently limit reliable detection.(11)

It is important to note that the neuropathologic reference standard is not directly translatable to

neuroimaging definitions. Furthermore, it is not clear yet whether the different subtypes of CMIs -

detected by different modalities - have the same etiological and functional implications. Further

cross-modal studies are warranted to determine this.

Table 1. Comparison of microinfarct detection modalities in terms of spatial and temporal resolution and brain coverage.

Modality Spatial

(microinfarct size)

Temporal

(microinfarct stage)

Volumetric

(brain coverage)

Neuropathology high

100 μm - few mm

high

(hyper)acute - chronic

low (<0·01%)

few sections

Diffusion-weighted MRI moderate

>1 mm

low

(hyper)acute (<2 weeks)

high

whole brain

High-resolution structural MRI moderate

≥1 mm

high

acute - chronic (> days)

partial

cortical grey matter

a. Neuropathology

CMIs were initially and most extensively described in neuropathology studies. (1,2) On microscopic

examination of standard histological Hematoxylin & Eosin (H&E)-stained sections, typical old

(chronic) CMIs can be discriminated as focal lesions of tissue cavitation with evidence of gliosis and

often a few remaining macrophages, a histological signature compatible with ischemia. Less

extensive chronic lesions may show only cell loss with or without macrophages or significant gliosis,

with tissue pallor or puckering (Figure 1). Evidence of old hemorrhage in the form of hemosiderin or

hematoidin may also be present, but unlike microhemorrhages it is not the primary signature of the

lesion. Recent (acute) CMIs show well defined foci of cell injury (red neurons in the grey matter)

without cavitation or cell loss (Figure 1). Most neuropathologic studies focus on chronic CMIs since

these can be most readily related to cognitive impairment often assessed months before death.

Acute CMIs may also be related to peri-mortem factors rather than factors relevant during life. Some

studies report immunohistochemical approaches for the enhanced detection of CMIs, such as glial

fibrillary acidic protein (GFAP) for gliosis, CD68 for macrophages, human leukocyte antigen (HLA)

markers for microglia(2), and calcineurin(13). However the sensitivity and specificity of CMI detection

using these approaches is not clear.

The differences between pathologic subtypes of CMI (cavitated, slit-like, hemorrhagic, or recent

(Figure 1)) have not yet been linked to specific pathogenic mechanisms, but do affect their MRI

appearance and detectability (see Box 1). Conventionally, CMIs are only considered microinfarcts (as

opposed to larger infarcts), when they are not visible to the naked eye on gross pathology. Their

reported sizes vary between 100 μm and a few millimetres (reviewed by (2)). Of note, standard

neuropathological evaluation may underestimate the actual sizes of CMIs, as often only a part of the

lesion is captured in a single thin paraffin section.

b. Diffusion-weighted imaging

DWI detects recent brain infarction with high sensitivity (Figure 2).(6) Because of the large amount of

signal generated by DWI, even very small infarcts can be detected, including 1-2 mm diameter

infarcts that are towards the upper limit of what pathologically would be defined as CMIs. However,

a disadvantage of DWI is that the high signal fades over two weeks (7), making it only useful for

identifying recent CMIs.

There is an emerging literature on incidental small DWI lesions, possibly representing CMIs caught

‘red-handed’, in patients with symptomatic ischemic stroke, intracerebral hemorrhage (ICH), or

cognitive impairment. Incidental small DWI lesions have been reported in 23-41% of patients with a

recent ICH (<3 months) where they are associated with MRI manifestations of SVD (14–17), 15% of

patients with cerebral amyloid angiopathy (CAA) and recent ICH (18), 6% of ischemic stroke patients(19),

14% of patients after carotid endarterectomy(20), and 1-4% of patients with cognitive impairment or

dementia(21,22). In contrast, the background prevalence of incidental small DWI lesions in the general

population appears much lower. In a population-based study of 793 participants aged 40-75, no

incidental DWI lesions were seen yielding a 95% confidence upper limit of 0·5% (23), while in a

community-based study of 623 older persons (mean age 71 years) with more vascular risk factors a

prevalence of 1% was seen.(22)

Most studies of incidental small DWI lesions have been cross-sectional. A notable exception is a small

study of five patients with extensive WMHs who had weekly MRI scans over 16 weeks, during which

three patients had nine incidental subcortical DWI lesions during follow-up. (24) Another longitudinal

study found that incidental DWI lesions appeared in 30 (27%) of 113 patients with ICH scanned at 30

days.(16) Based on cross-sectional prevalence data, it has been estimated that the average patient

with CAA may have as many as eight asymptomatic DWI lesions per year.(18)

A key question is the fate of incidental small DWI lesions. Preliminary data suggest that they can

evolve into a small cavity, a T2 hyperintensity without cavitation (Figure 2), or become radiologically

inapparent.(25) In some cases, acute small DWI lesions resolved into <3 mm cavities that in their

chronic stage would not be classifiable as infarcts according to STRIVE criteria. (26)

Most papers on incidental small DWI lesions do not report their actual size, but just state that they

were below a certain cut-off (e.g. 5 mm in diameter). However, review of representative figures from

prior publications(e.g. 14,15,18,25) suggests that most small DWI lesions are only a few mm in size and

potentially consistent with CMIs. The question is whether upper size limits for DWI lesions

compatible with acute CMIs can distinguish them from acute lacunar infarcts or larger acute cortical

infarcts. The STRIVE criteria(26) for recent small subcortical infarcts include an upper size limit of 20

mm, a criterion designed to identify lacunar infarcts in their acute stage, without providing a lower

size limit and without criteria for classification of cortical small infarcts. Another issue is that size of

DWI lesions depends on field strength and the parameters of the DWI sequence.

Based on the currently available insights, we postulate that DWI lesions with an axial diameter ≥5

mm are unlikely to evolve into what would be considered a CMI on structural brain imaging (see

subsequent section) or pathology, and should not be described as CMI. Hence, we propose a size

criterion of <5 mm for acute CMIs (see Box 1). Of note, this cut-off was operationally defined and

may need to be refined in the future as more data on the fate of incidental small DWI lesions and

radiological-pathological correlation become available.

c. High-resolution structural MRI

Detection of CMIs on structural MRI provides a major step forward in determining the causes and

consequences of these minute ischemic lesions. Insights from recent radiologic - histopathologic

correlation studies at high-field strength 7T MRI suggested that such durable ( i.e. not just visible in

the (sub)acute stage) CMIs of at least 1-2 mm in size in the cortical grey matter can be discerned on

structural MRI scans(9,11,27) (Figure 3, Box 1). These post-mortem verification studies indicated high

specificity (i.e. taken together 26 out of 27 lesions on post-mortem MRI that could be retrieved in the

corresponding histological sections were indeed CMIs on histology). (9,11,27) These initial studies chose a

pragmatic approach to specifically focus on the detection of CMIs in cortical areas of the brain

because in the white matter it would be difficult to distinguish CMIs from other lesions such as

WMHs, lacunar infarcts, and enlarged PVS. Indeed, in juxtacortical areas ( i.e. directly adjacent to

rather than within cortical grey matter) lesions that had an appearance quite similar to cortical CMIs

proved to be enlarged PVS on histology.(11) Hence, we consider punctate subcortical lesions on

structural MRI - without prior information of DWI positivity - currently ‘unclassifiable’ as CMIs.

Initial studies on the in vivo detection of cortical CMIs used 7T MRI.(9,28–30) One 7T MRI study in 23

patients showed that a subset (i.e. 27%) of the lesions also proved to be visible on 3T MRI scans

acquired on the same day.(31) Subsequently, cortical CMIs were examined in several studies on 3T

MRI.(19,31–33) CMIs have also occasionally been observed at 1·5T (Supplemental Figure 1), so this

possibility should not be overlooked. Along with field strength, the sensitivity of detection depends

on the scan protocol and level of scrutiny during examination.

Proposed rating criteria for the detection of durable CMIs on in vivo MRI and representative

examples are provided in Box 1, Figure 3, and Supplemental Figure 1. These rating criteria are based

on previous publications and were harmonized and further refined through consensus by the

authors. The shape of durable cortical CMIs on in vivo structural MRI generally matches the perfusion

territory of small single penetrating cortical arterioles (i.e. a cylindrical shape perpendicular to the

cortical surface extending towards the border with the white matter). Importantly, such CMIs can be

detected using existing datasets from large cohort studies, which provides a major opportunity to

advance this field. For a reliable detection of cortical CMIs on structural MRI, datasets should ideally

include high-resolution (≤1x1x1 mm voxel size) 3D T1-weighted and 3D T2-weighted ( i.e. FLAIR

and/or T2) MR images, as well as blood-sensitive scans to rule-out certain CMI mimics (e.g. cerebral

microbleeds (CMBs)). As most existing datasets at 1·5-3T MRI include only one 3D scan - which is

often the T1 - it is recommended to detect cortical CMIs in those datasets on T1, verifying possible

lesion locations on 2D T2-weighted images. Two recent 3T MRI studies measured the actual size of

such lesions smaller than 5 mm, in 75 and 209 patients respectively . (31,34) One reported a median

lesion size of 3 mm, with only 8% of lesions being larger than 3 mm. (31) The other similarly reported

an average size of 2·8 ± 0·8 mm (95% confidence interval [1·1; 4·5]).(34) We therefore propose a size

criterion for durable cortical CMIs of <4 mm (see Box 1).

Additional methodological details and CMI examples are available in a recent publication, which

includes a video instruction to guide detection of cortical CMIs on both 7T and 3T in vivo structural

MRI.(10) Currently, visual rating of cortical CMIs is time intensive (approximately 30 minutes per

subject), and inter-rater agreement is modest (two papers reported an intra-class correlation

coefficient of 0·39 and 0·76 respectively, one paper reported a kappa of 0·83). (29,33,35) Development of

semi-automated detection techniques is underway(29,36) and would greatly improve observer

reliability as well as the efficiency of evaluation of existing 3T datasets on large cohort studies that

are already available.

Box 1. Detection and visual rating criteria for CMIs on in vivo MRI

Acute CMIs on DWI should meet all of the following criteria ( i.e. small incidental DWI lesions) (Figure 2A)

- Hyperintense on DWI- Isointense or hypointense on ADC- Isointense or hyperintense on T2*-weighted MRI / blood-sensitive scans (e.g. GRE, SWI)- Any brain parenchymal location- Operationally defined as <5 mm in greatest dimension

Notes: be aware of the possibility of ‘T2-shine-through’ in the case of significant high signal on T2 and double-check the ADC, which should be iso- or hypointense at the same location.

Durable cortical CMIs on structural MRI should meet all of the following criteria (Figure 3, Supplemental Figure 1)

- Hyperintense on T2-weighted MRI (i.e. FLAIR, T2), with or without cavitation on FLAIR- Hypointense on T1-weighted MRI- Isointense on T2*-weighted MRI / blood-sensitive scans (e.g. GRE, SWI)- Operationally defined as strictly intracortical and <4 mm in greatest dimension- Distinct from enlarged PVS- Visible in at least two planes (e.g. sagittal, transversal, coronal)

Notes: important mimics for hemorrhagic cortical CMIs are cortical CMBs (hence always verify signal on blood-sensitive scans), enlarged PVS in the immediate underlying juxtacortical areas that may or may not extend into the cortical ribbon (Figure 3), leptomeningeal vessels (especially in the temporal lobes) (Figure 3), and anatomical variations (e.g. gyral curvatures). We suggest to discard CMIs in close proximity (i.e. <1 cm in the same gyrus) to larger strokes.

Abbreviations in Box 1

CMI: cerebral microinfarct; MRI: magnetic resonance imaging; DWI: diffusion-weighted imaging; ADC: apparent diffusion coefficient; GRE: gradient echo; SWI: susceptibility-weighted imaging; FLAIR: fluid-attenuated inversion recovery; PVS: perivascular spaces; CMBs: cerebral microbleeds

3. Etiology and risk factors of microinfarcts

CMIs have multiple underlying etiologies, which can co-exist in a single patient. Three main etiologies

are 1) cerebral SVD (e.g. CAA, arteriolosclerosis), 2) microemboli, and 3) hypoperfusion. As already

extensively discussed in the 2012 Lancet Neurology review, neuropathological studies show that

CMIs are common in the presence of severe CAA at autopsy (1) (see also papers thereafter(37–39)). More

recently, CMIs were assessed in relation to manifestations of CAA in living subjects. Durable cortical

CMIs were associated with lobar CMBs at 7T MRI(28,29) as well as 3T MRI(8,31,40). Interestingly, a recent

study that assessed cortical CMIs on in vivo 7T MRI in patients with hereditary CAA, showed that

CMIs were one of the earliest markers of the disease, compared to other clinical and MRI

manifestations.(41)

Recent neuropathological studies have further explored the relation of CMIs to CAA, atherosclerosis

and arteriolosclerosis. In one study, examining the brains of ~1,000 community-dwelling individuals

who came to autopsy, cortical CMIs were associated with CAA whereas subcortical CMIs were

associated with atherosclerosis and arteriolosclerosis.(42) In a smaller study of 80 autopsy cases, all

with multiple CMIs on routine pathological examination, it was found that CMIs in the occipital

cortex were associated with local CAA pathology, whereas CMIs in the frontal cortex and

hippocampus were not.(43)

Apart from CAA, CMIs are also associated with other manifestations of SVD on MRI. Both in memory

clinic patients(31) and in population based cohorts(32,33), CMIs co-occur with lacunar infarcts(31,33) and

presence of CMIs is related to higher WMH volume(31–33) and cerebral atrophy(31). CMI presence has

been linked to a history of stroke in population based cohorts(32,33), but also among patients admitted

for stroke(19,34,44) and in patients attending a memory clinic (diagnoses ranging from subjective

cognitive complaints to Alzheimer’s disease and vascular dementia) (31). Interestingly, this association

with stroke may also relate to large vessel disease, as large cortical infarcts on MRI and CMI have

been found to co-occur.(31,33) Of note, CMIs have been related to presence of intracranial stenosis in

studies of patients from a memory clinic(31) and in patients with ischemic stroke(34,44) and appear to be

more common ipsilateral to the stenosis(31,44). A recent study found an incidence of 14% for acute

CMIs on DWI in patients that underwent carotid endarterectomy.(20) Atrial fibrillation has also been

found to be a significant risk factor for cortical CMIs. (19,31,45) Moreover, biomarkers of cardiac disease,

in particular ‘NTpro-BNP’ and ‘hs-cTnT’ were associated with presence of cortical CMIs on 3T MRI in

memory-clinic patients.(45) Whether these links between CMIs and both large vessel atherosclerosis

and cardiac disease reflect microemboli as a causative mechanism or even hypoperfusion, which has

been put forward as an alternative mechanism of CMI formation, is still unclear.(46,47)

There also is an emerging literature on the association between vascular risk factors and CMIs on

MRI. A 7T MRI study assessing cortical CMIs in 48 patients with type 2 diabetes mellitus and 49

control subjects did not find a difference in presence or number of cortical CMIs between patients

with diabetes and controls.(35) This observation is in line with two recent neuropathological studies ,

assessing 1,228 and 2,365 autopsy cases respectively, that both showed an association between

diabetes and large infarcts, but not CMIs.(48,49) In a 3T MRI study in 238 memory-clinic patients CMIs

were not related to diabetes or hypertension, but did show a significant association with

hyperlipidemia.(31) In contrast, among 231 patients with ischemic stroke or transient ischemic attack

none of these risk factors was associated with presence of CMIs (19), whereas in a cohort of 861

participants from the general population CMIs were related to hypertension. (33) Obviously, risk factor

patterns for CMIs clearly depend on the setting in which they are studied and this is an area that

warrants further investigation.

Taken together, these studies clearly show that the etiology of CMIs is heterogeneous, being linked

to different manifestations of small vessel, large vessel, and cardiac disease. Hence, in most settings

presence of CMIs cannot be regarded as a marker of specific etiologies. Yet, in specific patient

subgroups with a high burden of one particular pathology, for example patients with diagnosed

probable CAA, presence of CMIs may be regarded as a likely marker of that pathology. The question

is if particular CMI features, such as location, size, or shape may be indicative of specific etiologies.

The aforementioned autopsy studies do indeed suggest that there may be a link between CMI

location and different subtypes of SVDs, but such studies are hampered by limited brain coverage.

Emerging MRI data from both memory clinic and stroke cohorts suggest a predilection for CMIs in

parietal and frontal areas (Supplemental Figure 2).(29,31,34) Yet, MRI studies are restricted by the fact

that durable subcortical lesions and lesions smaller than 1-2 mm cannot yet be detected. It is

expected that with the development of semi-automated CMI detection techniques a wealth of MRI

data from existing, large population and clinic-based cohorts with detailed information on risk factors

and comorbidities will become available. In addition, there are ongoing efforts for cross-modal

studies combining post-mortem MRI with histopathology. This should provide more unbiased data on

risk factors and etiologies of CMIs.

4. Functional impact of microinfarcts

Presence of CMIs at brain autopsy has been associated with ante-mortem cognitive dysfunction.(1)

Recent neuropathological studies have confirmed the link between CMIs on autopsy with ante-

mortem dementia and cognitive impairment, although the trajectories of cognitive decline later in

life in relation to CMIs and other co-occurring pathologies are complex and may need further

investigation.(50–55) Studies suggest a strong relationship with older age(52,53), race(54), and additive

effects with comorbid pathologies(54).

The recent advances in CMI detection during life greatly facilitate studies on their functional impact.

In one of the first studies that looked at CMIs on in vivo 7T MRI, higher numbers of CMIs were found

in 14 patients with Alzheimer’s disease compared to 18 healthy controls, although the proportion of

subjects with CMIs did not differ between the groups (86% and 72% respectively). (28) Another 7T MRI

study similarly found no difference in CMI occurrence between 29 patients with early Alzheimer’s

disease compared to 22 healthy controls (55% and 45% respectively), but did not confirm differences

in CMI numbers between groups.(29) Both of these initial studies were limited by small numbers of

participants, inherent to most in vivo 7T MRI studies. Moreover, one of the studies also considered

juxtacortical lesions28, which according to current insights may include CMI mimics, in particular

enlarged perivascular spaces.11 It has since become clear that CMIs can also be detected on 3T MRI

scans, which allows assessing their relation with cognition in much larger cohorts, likely with greater

generalizability of the findings. In a memory-clinic cohort including 238 patients, CMIs were

associated with dementia, most strongly with vascular dementia (CMIs occurred in 12 (55%) of 22

patients with vascular dementia).(31) In particular, CMIs were related to worse global cognition and

poorer performance on tasks related to visuoconstructive abilities and language. In a sample of

patients with ischemic stroke or transient ischemic attack both cortical CMIs and acute CMIs at

baseline predicted worse cognitive performance at two-year follow-up than in patients without CMIs,

especially in the domain of visuospatial functioning. (19) In an Asian study of a population at risk of

cognitive impairment, including 861 subjects, the presence of cortical CMIs on in vivo 3T MRI was

independently associated with dementia.(33) Furthermore, cortical CMIs were significantly associated

with worse global cognition and poorer performance on tasks in the domains of executive function,

visual memory, and verbal memory.(33) Of note, these studies took possible confounding effects of

other markers of vascular damage and neurodegeneration (i.e. atrophy) in consideration. One

neuropathology study, assessing 850 autopsy cases, found that CMIs were associated with impaired

motor function proximate to death.(56) The impact of CMIs on motor function has not yet been

addressed on in vivo MRI.

Taken together, these initial in vivo MRI studies have so far confirmed neuropathological

observations that CMIs are associated with worse cognition, also independently of other age-related

pathologies. An unresolved issue, however, is whether such associations are causal. It could be

argued that due to their high numbers and widespread distribution throughout the brain, CMIs cause

sufficient disruption resulting in functional impairment. Animal studies, reviewed in the next section,

support this. An alternative explanation would be that CMIs are a marker of other underlying

vascular pathologies that are even more widespread than the lesions themselves and thus impact the

brain also without visible focal injury.

5. Animal models of microinfarcts

a. Mechanisms

Basic scientists have successfully modelled CMIs in the brains of rodents by occluding penetrating

arterioles that form bottlenecks in microvascular perfusion. Penetrating arterioles have emerged as a

key locus for occlusion, because unlike the interconnected pial and capillary systems, blood flow

through a penetrating arteriole cannot be efficiently re-routed around a localized clot.(57) CMIs

induced in rodents show remarkable similarity to human CMIs, with respect to their range of shapes,

size, and location, as well as their temporal evolution (Figure 4). For example, the core of an

experimental CMI becomes packed with microglia and is surrounded by reactive astrocytes, just as

with human CMIs.(58)

Two distinct but complementary strategies have been used to model CMIs. One method involves the

intra-arterial injection of occlusive microbeads(59), cholesterol crystals(60,61) or microthrombi(62). This

abrupt shower of microemboli produces CMIs that are broadly distributed, with shapes ranging from

wedge / column-like lesions in the cortex to smaller, circumscribed CMIs in both cortical and

subcortical tissues. A second, more recently developed model of cortical microinfarction provides a

means to reproducibly target the location and size of CMIs in rodent cortex. Typically coupled with in

vivo multiphoton imaging, single penetrating arterioles are selectively occluded by inducing clots with

precise laser irradiation, with and without intravenous photosensitizing agents.(63–67)

Spontaneously occurring CMIs have also been described in mouse models. When subjected to

forebrain hypoperfusion with bilateral carotid stenosis, the Tg-SwDI mouse model of CAA develops

cortical and subcortical CMIs by 18 to 32 weeks of age. (46) Application of carotid stenosis to non-

transgenic C57BL/6 mice also led to spontaneous CMI formation (as well as microhemorrhages), but

at lower levels compared to Tg-SwDI mice.(68) Mice deficient in endothelial nitric oxide synthase

exhibit cerebral hypoperfusion and also develop both cortical and subcortical CMIs. (69) These models

can provide important clues on the etiology of CMIs in humans, and the vascular changes responsible

for CMI formation should be thoroughly investigated.

b. Peri-lesional and remote effects of microinfarcts

Animal studies have provided evidence that the physiologic impact of CMIs extends well beyond the

non-viable lesion core observed by MRI or depicted in neuropathological studies (Figure 5). By

strategically targeting CMIs to the mouse sensory or motor cortices, in vivo imaging studies have

reported deficits in both neural and hemodynamic function in tissue surrounding the CMI that can

last for many weeks.(70) This involved persistent alterations in neuronal circuitry, observed by

displacement and disorganization of sensorimotor maps.(71) A recent study estimated that a single

CMI, with a lesion core averaging 0·2 mm3 in volume (~0·5 mm in diameter), could impair neuronal

function over cortical regions at least 12-times larger than this core. (58) Even this, however, was likely

to be an underestimate as CMIs in both grey and white matter can also disrupt neighboring fibers of

passage, as evidenced by demyelination and damage of axonal structure.(59–61) Consistent with having

a large impact on brain function, rodent studies using unilateral (60,61) or repeated bilateral cholesterol

crystal injections(72) have reported deficits in cognitive tasks, despite the actual CMI load being small

compared to overall brain size. Two of these studies also showed brain-wide impairment of

glymphatic pathways, a perivascular network for clearance of amyloid β and other cytotoxic

metabolites from the brain.(61,73)

Examination of the peri-lesional areas of experimentally-induced CMIs has revealed numerous

pathological changes that are likely substrates for observed functional and behavioral deficits. There

is evidence for delayed, selective neuronal loss (5-10% decrease) (60), thinning of dendritic processes,

and reduced dendritic spine density(58,61). Non-neuronal changes include widespread astrogliosis,

reduced expression of aquaporin 4, and myelin loss.(58,60,61,63,74) Blood-brain barrier disruption is also

evident based on extravasation of plasma proteins and nitrosative stress at the capillary walls. (63,72)

One striking manifestation of this pathology is the coalescence of isolated CMIs into larger

contiguous infarcts when their peri-lesional zones begin to overlap. (63) This emphasizes the

vulnerability of tissues in peri-lesional areas, and suggests that injury can be shifted toward further

exacerbation, or rescued by therapeutic agents.

The peri-lesional and remote effects of CMIs likely involve the same pathophysiological processes

that occur with larger strokes, but on a smaller scale. For example, CMIs are able to evoke waves of

ischemic spreading depression that emanate several millimetres away from the CMI core, which may

contribute to neuronal damage and hypoperfusion.(63) Neuronal dysfunction may also involve

diaschisis, where death of neurons within the CMI core disables the cortical and subcortical circuits

to which they were previously integrated. Maladaptive increases of inhibitory tone in peri-lesional

tissues may also depress neuronal excitation.(75) Additionally, CMIs located within or in close

proximity to white matter tracts likely disrupt communication between distant brain regions by

damaging axonal structure. Finally, impairment of glymphatic drainage may elevate toxic metabolites

in a brain-wide fashion.(61,73) These potential mechanisms of CMI-induced pathology will need to be

studied in greater detail using experimental systems, as the relative importance of each mechanism

remains poorly understood.

6. Conclusions and future directions

The course of CMI research has largely depended on the relatively complex problem of how to assess

their total burden in the brain. Accurate quantification of CMIs is central for identifying their specific

impact on brain function, and the pathogenic processes involved in their development. It is thus

exciting to note substantial progress over recent years in CMI detection, particularly the publication

of pathologically validated structural MRI methods and proposed rating criteria for cortical CMIs (Box

1). Further development of these methods are likely, including the possibility of semi-automated

detection methods for cortical CMIs that could be used to analyze large, population-derived 3T MRI

datasets. At the same time, it is chastening to note that current methods remain able to detect only a

small fraction of CMIs, hence are biased towards lesions with a particular set of spatial, temporal, or

anatomic characteristics (Table 1). In practice, these limitations imply that all studies of the effects of

CMIs have considerable built-in error due to miscounting and biased counting. Moreover, the

reliability of visual and emerging computational assessments remains largely to be assessed (Box 2).

Another area of both notable progress and substantial remaining uncertainty is the extent to which

CMIs disrupt the immediately surrounding brain and the larger scale brain network structure and

function. Such potential widespread effects of CMIs potentially allow these numerous, spatially

distributed lesions to exert greater impact on brain function than predicted by their small aggregate

volume (which is likely to be true for other brain lesions in the context of SVD as well). As the

strongest evidence for more widespread effects of CMIs currently comes from animal model-based

studies (section 5b; Figure 5) it will be important for future studies to translate imaging markers from

animals to human research participants (Box 2).

The results of future improvements in CMI detection and measurement of their widespread effects

will determine how central these lesions are, not just as markers of microvascular injury and

cognitive impairment, but as direct contributors to cognitive decline. Identification of a major

causative role for CMIs would clearly place these lesions high on the list of candidate outcome

markers for trials aimed at preventing vascular cognitive impairment. If the accrual of CMIs does play

a major causal role in progressive cognitive impairment, it would also bring renewed focus on

ischemic neuroprotection - a biology that has been extensively studied in acute stroke without

clinical efficacy.(76) The potential impact of improvements in the CMI field thus extends beyond simply

providing another imaging biomarker of cerebral SVD to the larger goal of guiding future therapeutic

approaches.

Box 2. Directions for future research

Identification and definition of microinfarcts

Re-fine operational rating criteria for the detection of acute CMIs on diffusion-weighted MRI (e.g. size cut-off) and verify the ischemic nature of these small incidental DWI lesions by MRI-histopathologic correlation.

Establish sensitivity, specificity, and rater reliability for proposed criteria for detection of CMIs on MRI accounting for differences in field strength and other imaging parameters.

Establish MRI detection methods for durable subcortical CMIs. Investigate topographical distribution of CMIs, e.g. with regards to lobar predilections, grey

vs. white matter, and underlying SVDs. Perform cross-modal studies on different pathologic subtypes of CMIs (e.g. cavitated,

hemorrhagic, slit-like) as well as lesion sizes, and their MRI appearance.

Etiology, risk factors, and functional impact

Perform longitudinal MRI studies in different populations, both general population, and specific patient cohorts, with rich phenotyping of risk factors, etiologies, and functional outcomes.

Identify pathophysiologic mechanisms underlying CMI formation in different patient populations focusing on the roles of vascular risk factors, small vessel diseases, (micro)emboli, and hypoperfusion.

Establish if features such as size, distribution, and appearance of CMIs reflect different pathogenic mechanisms.

Study the remote effects of CMIs on brain structure and function. Translate insights from animal models to humans.

Use animal models that spontaneously develop CMIs to better understand vascular factors leading to CMIs.

Treatment

Establish if CMIs can be used as reliable and informative biomarkers in multicenter clinical trials.

Study the role of aggressive blood pressure reduction and other vascular interventions in the formation of CMIs.

Search strategy and selection criteria

References for this review were identified by searches on PubMed focusing on papers that appeared

between 2012 and April 2017. The search terms (and synonyms) ‘microinfarct*’, in combination with

‘brain or cerebral’, and ‘autopsy’, ‘neuropathology’, ‘MRI’, ‘imaging’, or ‘diffusion’ were used. For the

animal section, the search terms ‘microinfarct*’, ‘microemboli’, ‘ministroke’, or ‘infarct’ and ‘brain or

cerebral’, and ‘rodent or mouse’, and ‘penetrating vessel’ were used. From the papers identified

through these searches and additional papers from our records a selection was made by SJvV and

GJB, with feedback from the co-authors. For the selection we focused on original papers, published in

English, and - where appropriate - high impact reviews. The final reference list was generated on the

basis of relevance and originality to the topics covered in this review.

Author contributions

SJvV and GJB coordinated the overall process of the manuscript. SJvV, GJB, and SMG were involved in

study design. SJvV, AYS, EES, JAS, and JMW performed literature searches, and SJvV and GJB made a

selection of identified papers. SJvV, AYS, EES, SMG, and GJB wrote the manuscript. SJvV and AYS

prepared figures. All authors were involved in critically reading and editing the manuscript. All

authors gave final approval for submission. SJvV and GJB prepared the manuscript for submission.

Acknowledgments

SJvV received funding support from the Netherlands Organization for Scientific Research

(019.153LW.014). AYS received support from the National Institute of Neurological Disorders and

Stroke (NS085402, NS096997), the Dana Foundation, The National Science Foundation (1539034),

South Carolina Clinical and Translational Institute (UL1TR000062), Alzheimer’s Association (New

Investigator Research Grant), and the National Institute of General Medical Sciences (P20GM12345),

and the Charleston Conference on Alzheimer’s Disease New Vision Award. EES received support from

the University of Calgary, Alberta Innovates Health Solutions, the Canadian Institutes of Health

Research, Brain Canada, Heart and Stroke Foundation of Canada, Canadian Partnership Against

Cancer, and the European Union Joint Programme-Neurodegenerative Disease Research. CC received

support from the National Medical Research Council Singapore (NMRC/CIRG/1446/2016,

NMRC/CG/013/2013). JAS received support from the National Institute of Health (UH2 NS100599-

01). JMW received support from Fondation Leducq (CVD-05), European Union Horizon 2020 (PHC-03-

15 SVDs@Target grant agreement 666881), Medical Research Council (MR/K026992/1), Age UK,

Alzheimer Society (AS-PG-14-033), Welcome Trust (088134/Z/09), Innovate UK, British Heart

Foundation, and Chief Scientist Office of the Scottish Executive. SMG has received support from the

National Institute of Health (R01 AG26484, R01 NS096730). GJB has received support from the Dutch

Heart Association (2010T073), ZonMw (91711384, 91816616), the Netherlands Organization for

Health Research and Development, European Union Horizon 2020 (SVDs@Target grant agreement

No. 666881).

Declaration of interests

[JAS] Consultant Navidea Biopharmaceuticals, Eli Lilly Inc., Genentech, Michael J. Fox Foundation

References

1. Smith EE, Schneider JA, Wardlaw JM, Greenberg SM. Cerebral microinfarcts: the invisible lesions. Lancet Neurol 2012; 11: 272–82.

2. Brundel M, De Bresser J, Van Dillen JJ, Kappelle LJ, Biessels GJ. Cerebral microinfarcts: a systematic review of neuropathological studies. J Cereb Blood Flow Metab 2012; 32: 425–36.

3. Westover MB, Bianchi MT, Yang C, Schneider JA, Greenberg SM. Estimating cerebral microinfarct burden from autopsy samples. Neurology 2013; 13: 1–5.

4. Auriel E, Westover MB, Bianchi MT, Reijmer Y, Martinez-Ramirez S, Ni J, et al. Estimating Total Cerebral Microinfarct Burden From Diffusion-Weighted Imaging. Stroke 2015; 46: 2129–35.

5. Sonnen JA, Larson EB, Crane PK, Haneuse S, Li G, Schellenberg GD, et al. Pathological correlates of dementia in a longitudinal, population-based sample of aging. Ann Neurol 2007; 62: 406–13.

6. Keir SL, Wardlaw JM. Systematic review of diffusion and perfusion imaging in acute ischemic stroke. Stroke 2000; 31: 2723–31.

7. Lansberg MG, Thijs VN, Brien MW, Ali JO, De Crespigny AJ, Tong DC, et al. Evolution of Apparent Diffusion Coefficient, Diffusion-weighted, and T2-weighted Signal Intensity of Acute Stroke. Stroke 2001; 22: 637–44.

8. Ii Y, Maeda M, Kida H, Matsuo K, Shindo A, Taniguchi A, et al. In vivo Detection of Cortical Microinfarcts on Ultrahigh-Field MRI. J Neuroimaging 2012; 23: 28–32.

9. Van Veluw SJ, Zwanenburg JJ, Engelen-Lee J, Spliet WG, Hendrikse J, Luijten PR, et al. In vivo detection of cerebral cortical microinfarcts with high-resolution 7T MRI. J Cereb Blood Flow Metab 2013; 33: 322–9.

10. Van Veluw SJ, Biessels GJ, Luijten PR, Zwanenburg JJ. Assessing Cortical Cerebral Microinfarcts on High Resolution MR Images. J Vis Exp 2015; 105: 1–8.

11. Van Veluw SJ, Zwanenburg JJ, Rozemuller AJ, Luijten PR, Spliet WG, Biessels GJ. The spectrum of MR detectable cortical microinfarcts: a classification study with 7-tesla postmortem MRI and histopathology. J Cereb Blood Flow Metab 2015; 35: 676–83.

12. De Reuck J, Deramecourt V, Auger F, Durieux N, Cordonnier C, Devos D, et al. Post-mortem 7.0-tesla magnetic resonance study of cortical microinfarcts in neurodegenerative diseases and vascular dementia with neuropathological correlates. J Neurol Sci 2014; 346: 85–9.

13. Pleiss MM, Sompol P, Kraner SD, Abdul HM, Furman JL, Guttmann RP, et al. Calcineurin proteolysis in astrocytes: Implications for impaired synaptic function. Biochim Biophys Acta - Mol Basis Dis 2016; 1862: 1521–32.

14. Prabhakaran S, Gupta R, Ouyang B, John S, Temes RE, Mohammad Y, et al. Acute brain infarcts after spontaneous intracerebral hemorrhage: a diffusion-weighted imaging study. Stroke 2010; 41: 89–94.

15. Gregoire SM, Charidimou A, Gadapa N, Dolan E, Antoun N, Peeters A, et al. Acute ischaemic brain lesions in intracerebral haemorrhage: Multicentre cross-sectional magnetic resonance imaging study. Brain 2011; 134: 2376–86.

16. Menon RS, Burgess RE, Wing JJ, Gibbons MC, Shara NM, Fernandez S, et al. Predictors of highly prevalent brain ischemia in intracerebral hemorrhage. Ann Neurol 2012; 71: 199–205.

17. Garg RK, Liebling SM, Maas MB, Nemeth AJ, Russell EJ, Naidech AM. Blood pressure reduction, decreased diffusion on MRI, and outcomes after intracerebral hemorrhage. Stroke 2012; 43: 67–71.

18. Kimberly WT, Gilson A, Rost NS, Rosand J, Viswanathan A, Smith EE, et al. Silent ischemic infarcts are associated with hemorrhage burden in cerebral amyloid angiopathy. Neurology 2009; 72: 1230–5.

19. Wang Z, Van Veluw SJ, Wong A, Liu W, Shi L, Yang J, et al. Risk Factors and Cognitive Relevance of Cortical Cerebral Microinfarcts in Patients with Ischemic Stroke or Transient Ischemic Attack. Stroke 2016; 47: 2450–5.

20. Gwon JG, Kwon TW, Cho YP, Kang DW, Han Y, Noh M. Analysis of Risk Factors for Cerebral Microinfarcts after Carotid Endarterectomy and the Relevance of Delayed Cerebral Infarction. J Clin Neurol 2017; 13: 32–7.

21. Saini M, Ikram K, Hilal S, Qiu A, Venketasubramanian N, Chen C. Silent stroke: Not listened to rather than silent. Stroke 2012; 43: 3102–4.

22. Saini M, Suministrado MS, Hilal S, Dong YH, Venketasubramanian N, Ikram MK, et al. Prevalence and risk factors of acute incidental infarcts. Stroke 2015; 46: 2722–7.

23. Batool S, O’Donnell M, Sharma M, Islam S, Dagenais GR, Poirier P, et al. Incidental magnetic resonance diffusion-weighted imaging-positive lesions are rare in neurologically asymptomatic community-dwelling adults. Stroke 2014; 45: 2115–7.

24. Conklin J, Silver FL, Mikulis DJ, Mell DM. Are acute infarcts the cause of leukoaraiosis? Brain mapping for 16 consecutive weeks. Ann Neurol 2014; 76: 899–904.

25. Auriel E, Edlow BL, Reijmer YD, Fotiadis P, Ramirez-Martinez S, Ni J, et al. Microinfarct disruption of white matter structure: A longitudinal diffusion tensor analysis. Neurology 2014; 83: 182–8.

26. Wardlaw JM, Smith EE, Biessels GJ, Cordonnier C, Fazekas F, Frayne R, et al. Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol 2013; 12: 822–38.

27. Van Veluw SJ, Charidimou A, Van der Kouwe AJ, Lauer A, Reijmer YD, Costantino I, et al. Microbleed and microinfarct detection in amyloid angiopathy: a high-resolution MRI-histopathology study. Brain 2016; 139: 3151–62.

28. Van Rooden S, Goos JD, Van Opstal AM, Versluis MJ, Webb AG, Blauw GJ, et al. Increased Number of Microinfarcts in Alzheimer Disease at 7-T MR Imaging. Radiology 2014; 270: 205–11.

29. Van Veluw SJ, Heringa SM, Kuijf HJ, Koek HL, Luijten PR, Biessels GJ. Cerebral cortical microinfarcts at 7Tesla MRI in patients with early Alzheimer’s disease. J Alzheimer’s Dis 2014; 39: 163–7.

30. Van Veluw SJ, Jolink WM, Hendrikse J, Geerlings MI, Luijten PR, Biessels GJ, et al. Cortical microinfarcts on 7T MRI in patients with spontaneous intracerebral hemorrhage. J Cereb Blood Flow Metab 2014; 34: 1104–6.

31. Van Veluw SJ, Hilal S, Kuijf HJ, Ikram MK, Xin X, Yeow TB, et al. Cortical microinfarcts on 3T MRI: Clinical correlates in memory-clinic patients. Alzheimer’s Dement 2015; 11: 1500–9.

32. Van Dalen JW, Scuric EE, Van Veluw SJ, Caan MW, Nederveen AJ, Biessels GJ, et al. Cortical microinfarcts detected in vivo on 3 tesla MRI: Clinical and radiological correlates. Stroke 2015;

46: 255–7.

33. Hilal S, Sikking E, Shaik MA, Chan QL, Van Veluw SJ, Vrooman H, et al. Cortical cerebral microinfarcts on 3T MRI: A novel marker of cerebrovascular disease. Neurology 2016; 87: 1583-1590.

34. Fu R, Wang Y, Wang Y, Liu L, Zhao X, Wang DZ, et al. The Development of Cortical Microinfarcts Is Associated with Intracranial Atherosclerosis: Data from the Chinese Intracranial Atherosclerosis Study. J Stroke Cerebrovasc Dis 2015; 24: 2447–54.

35. Brundel M, Reijmer YD, Van Veluw SJ, Kuijf HJ, Luijten PR, Kappelle LJ, et al. Cerebral microvascular lesions on high-resolution 7-tesla MRI in patients with type 2 diabetes. Diabetes 2014; 63: 3523–9.

36. Kuijf HJ, Zijlstra F, Van Veluw SJ, Viergever MA, Vincken KL, Biessels GJ. Detecting cortical cerebral microinfarcts in 7.0 T MR images. IEEE International Symposium on Biomedical Imaging: From Nano to Macro. 2013; 2: 970–3.

37. Lauer A, Van Veluw SJ, William CM, Charidimou A, Vashkevich A, Ayres A, et al. Microbleeds on MRI are associated with microinfarcts on autopsy in cerebral amyloid angiopathy. Neurology 2016; 87: 1488–92.

38. Ringman JM, Sachs MC, Zhou Y, Monsell SE, Saver JL, Vinters HV. Clinical predictors of severe cerebral amyloid angiopathy and influence of APOE genotype in persons with pathologically verified Alzheimer disease. JAMA Neurol 2014; 71: 878–83.

39. Kövari E, Herrmann FR, Hof PR, Bouras C. The relationship between cerebral amyloid angiopathy and cortical microinfarcts in brain ageing and Alzheimer’s disease. Neuropathol Appl Neurobiol 2013; 39: 498–509.

40. Ueda Y, Satoh M, Tabei KI, Kida H, Ii Y, Asahi M, et al. Neuropsychological Features of Microbleeds and Cortical Microinfarct Detected by High Resolution Magnetic Resonance Imaging. J Alzheimer’s Dis 2016; 53: 315–25.

41. Van Rooden S, Van Opstal AM, Labadie G, Terwindt GM, Wermer MJ, Webb AG, et al. Early Magnetic Resonance Imaging and Cognitive Markers of Hereditary Cerebral Amyloid Angiopathy. Stroke 2016; 47: 3041–44.

42. Arvanitakis Z, Capuano AW, Leurgans SE, Buchman AS, Bennett DA, Schneider JA. The Relationship of Cerebral Vessel Pathology to Brain Microinfarcts. Brain Pathol 2016; 27: 77–85.

43. Kövari E, Herrmann FR, Gold G, Hof PR, Charidimou A. Association of cortical microinfarcts and cerebral small vessel pathology in the ageing brain. Neuropathol Appl Neurobiol 2016; [Epub ahead of print].

44. Dieleman N, Van der Kolk AG, Zwanenburg JJ, Brundel M, Harteveld AA, Biessels GJ, et al. Relations between location and type of intracranial atherosclerosis and parenchymal damage. J Cereb Blood Flow Metab 2016; 36: 1271–80.

45. Hilal S, Chai YL, Van Veluw SJ, Shaik MA, Ikram MK, Venketasubramanian N, et al. Association between subclinical cardiac biomarkers and clinically manifest cardiac diseases with cortical cerebral microinfarcts. JAMA Neurol 2017; 74: 403–10.

46. Okamoto Y, Yamamoto T, Kalaria RN, Senzaki H, Maki T, Hase Y, et al. Cerebral hypoperfusion accelerates cerebral amyloid angiopathy and promotes cortical microinfarcts. Acta Neuropathol 2012; 123: 381–94.

47. Suter OC, Sunthorn T, Kraftsik R, Straubel J, Darekar P, Khalili K, et al. Cerebral Hypoperfusion Generates Cortical Watershed Microinfarcts in Alzheimer Disease. Stroke 2002; 33: 1986–92.

48. Pruzin JJ, Schneider JA, Capuano AW, Leurgans SE, Barnes LL, Ahima RS, et al. Diabetes, Hemoglobin A1C, and Regional Alzheimer Disease and Infarct Pathology. Alzheimer Dis Assoc Disord 2017; 31: 41–47.

49. Abner EL, Nelson PT, Kryscio RJ, Schmitt FA, Fardo DW, Woltjer RL, et al. Diabetes is associated with cerebrovascular but not Alzheimer neuropathology. Alzheimer’s Dement 2016; 12: 1–8.

50. Ince PG, Minett T, Forster G, Brayne C, Wharton SB. Microinfarcts in an older population-representative brain donor cohort (MRC CFAS): Prevalence, relation to dementia and mobility, and implications for the evaluation of cerebral Small Vessel Disease. Neuropathol Appl Neurobiol 2016; [Epub ahead of print].

51. Boyle PA, Yu L, Wilson RS, Schneider JA, Bennett DA. Relation of neuropathology with cognitive decline among older persons without dementia. Front Aging Neurosci 2013; 5: 1–8.

52. Corrada MM, Sonnen JA, Kim RC, Kawas CH. Microinfarcts are common and strongly related to dementia in the oldest-old: The 90+ study. Alzheimer’s Dement 2016; 12: 900–8.

53. Kawas CH, Kim RC, Sonnen JA, Bullain SS, Trieu T, Corrada MM. Multiple pathologies are common and related to dementia in the oldest-old: The 90+ Study. Neurology 2015; 85: 535–42.

54. White LR, Edland SD, Hemmy LS, Montine KS, Zarow C, Sonnen JA, et al. Neuropathologic comorbidity and cognitive impairment in the Nun and Honolulu-Asia Aging Studies. Neurology 2016; 86: 1000–8.

55. Boyle PA, Yang J, Yu L, Leurgans SE, Capuano AW, Schneider JA, et al. Varied effects of age-related neuropathologies on the trajectory of late life cognitive decline. Brain 2017; 140: 804–12.

56. Buchman AS, Yu L, Boyle PA, Levine SR, Nag S, Schneider JA, et al. Microvascular brain pathology and late-life motor impairment. Neurology 2013; 80: 712–8.

57. Blinder P, Shih AY, Rafie C, Kleinfeld D. Topological basis for the robust distribution of blood to rodent neocortex. Proc Natl Acad Sci 2010; 107: 12670–5.

58. Summers PM, Hartmann DA, Hui ES, Nie X, Deardorff RL, McKinnon ET, et al. Functional deficits induced by cortical microinfarcts. J Cereb Blood Flow Metab 2017; [Epub ahead of print].

59. Silasi G, She J, Boyd JD, Xue S, Murphy TH. A mouse model of small-vessel disease that produces brain-wide-identified microocclusions and regionally selective neuronal injury. J Cereb Blood Flow Metab 2015; 35: 734–8.

60. Wang M, Iliff JJ, Liao Y, Chen MJ, Shinseki MS, Venkataraman A, et al. Cognitive Deficits and Delayed Neuronal Loss in a Mouse Model of Multiple Microinfarcts. J Neurosci 2012; 32: 17948–60.

61. Venkat P, Chopp M, Zacharek A, Cui C, Zhang L, Li Q, et al. White matter damage and glymphatic dysfunction in a model of vascular dementia in rats with no prior vascular pathologies. Neurobiol Aging 2017; 50: 96–106.

62. Rapp JH, Pan XM, Yu B, Swanson RA, Higashida RT, Simpson P, et al. Cerebral ischemia and infarction from atheroemboli <100 micrometer in size. Stroke 2003; 34: 1976–80.

63. Shih AY, Blinder P, Tsai PS, Friedman B, Stanley G, Lyden PD, et al. The smallest stroke: occlusion of one penetrating vessel leads to infarction and a cognitive deficit. Nat Neurosci 2012; 16: 55–63.

64. Nishimura N, Schaffer CB, Friedman B, Lyden PD, Kleinfeld D. Penetrating arterioles are a bottleneck in the perfusion of neocortex. Proc Natl Acad Sci 2007; 104: 365–70.

65. Taylor ZJ, Hui ES, Watson AN, Nie X, Deardorff RL, Jensen JH, et al. Microvascular basis for growth of small infarcts following occlusion of single penetrating arterioles in mouse cortex. J Cereb Blood Flow Metab 2016; 36: 1357–73.

66. Nishimura N, Rosidi NL, Iadecola C, Schaffer CB. Limitations of collateral flow after occlusion of a single cortical penetrating arteriole. J Cereb Blood Flow Metab 2010; 30: 1914–27.

67. Zhang Q, Lan Y, He XF, Luo CM, Wang QM, Liang FY, et al. Allopurinol protects against ischemic insults in a mouse model of cortical microinfarction. Brain Res 2015; 1622: 361–7.

68. Holland PR, Searcy JL, Salvadores N, Scullion G, Chen G, Lawson G, et al. Gliovascular disruption and cognitive deficits in a mouse model with features of small vessel disease. J Cereb Blood Flow Metab 2015; 35: 1005–14.

69. Tan XL, Xue YQ, Ma T, Wang X, Li JJ, Lan L, et al. Partial eNOS deficiency causes spontaneous thrombotic cerebral infarction, amyloid angiopathy and cognitive impairment. Mol Neurodegener 2015; 10: 24.

70. Anenberg E, Arstikaitis P, Niitsu Y, Harrison TC, Boyd JD, Hilton BJ, et al. Ministrokes in channelrhodopsin-2 transgenic mice reveal widespread deficits in motor output despite maintenance of cortical neuronal excitability. J Neurosci 2014; 34: 1094–104.

71. Harrison TC, Silasi G, Boyd JD, Murphy TH. Displacement of sensory maps and disorganization of motor cortex after targeted stroke in mice. Stroke 2013; 44: 2300–6.

72. Rapp JH, Pan XM, Neumann M, Hong M, Hollenbeck K, Liu J. Microemboli composed of cholesterol crystals disrupt the blood-brain barrier and reduce cognition. Stroke 2008; 39: 2354–61.

73. Wang M, Ding F, Deng S, Guo X, Wang W, Iliff JJ, et al. Focal Solute Trapping and Global Glymphatic Pathway Impairment in a Murine Model of Multiple Microinfarcts. J Neurosci 2017; 37: 2870–77.

74. Zheng W, Watts LT, Holstein DM, Prajapati SI, Keller C, Grass EH, et al. Purinergic receptor stimulation reduces cytotoxic edema and brain infarcts in mouse induced by photothrombosis by energizing glial mitochondria. PLoS One 2010; 5: e14401.

75. Clarkson AN, Huang BS, Macisaac SE, Mody I, Carmichael ST. Reducing excessive GABA-mediated tonic inhibition promotes functional recovery after stroke. Nature 2010; 468: 305–9.

76. Sutherland BA, Minnerup J, Balami JS, Arba F, Buchan AM, Kleinschnitz C. Neuroprotection for ischaemic stroke: Translation from the bench to the bedside. Int J Stroke 2012; 7: 407–18.

77. Blinder P, Tsai PS, Kaufhold JP, Knutsen PM, Suhl H, Kleinfeld D. The cortical angiome: an interconnected vascular network with noncolumnar patterns of blood flow. Nat Neurosci 2013; 16: 889–97.

78. Duvernoy HM, Delon S, Vannson JL. Cortical blood vessels of the human brain. Brain Res Bull 1981; 7: 519–79.

Figure legends

Figure 1. Cerebral microinfarcts on neuropathology.

Hematoxylin & Eosin (H&E)-stained sections of (A) an acute microinfarct in the midfrontal cortex;

inset shows a red (i.e. hypoxic) neuron (arrow), (B) a chronic slit-like microinfarct in the inferior

parietal cortex (i.e. puckering), (C) a chronic microinfarct with cavitation in the inferior parietal

cortex, (D) a chronic microinfarct with hemosiderin deposits in the inferior parietal cortex; inset

shows hemosiderin deposits (arrows).

Figure 2. Incidental small DWI lesion and its evolution over time.

(A) Diffusion-weighted imaging (DWI) 1·5T MRI shows a hyperintense subcortical lesion (arrow).

Follow-up imaging at 3 months reveals that the lesion (arrow) is still visible, appearing (B)

hypointense on T1 and (C) hyperintense on FLAIR.

Figure 3. Durable cortical cerebral microinfarcts and mimics on structural MRI.

Top row: Schematic representation of a cortical cerebral microinfarct (CMI), which is hypointense on

T1, hyperintense on T2 and FLAIR, and isointense on T2*-weighted images. The schematic lesion on

the right (indicated with asterisk) represents a cortical CMI with cavitation, which appears

hypointense on FLAIR with a hyperintense rim. Middle two rows: examples of a cortical CMI (arrows)

without cavitation in the same individual captured on the same day at 7T and 3T MRI. Bottom row:

Examples of common CMI mimics, such as microbleeds (CMB; arrows), which are distinct from CMIs

because they appear hypointense on T2*-weighted MRI; enlarged perivascular spaces (PVS; arrow),

which are distinct from cortical CMIs because they are located in juxtacortical areas; and vessels,

which are distinct from cortical CMIs because they appear hypointense on T2*-weighted MRI and can

be traced over several slices (as indicated by the two arrows in the inset).

Figure 4. Modeling microinfarcts in rodent cortex by targeted photothrombotic occlusion of

individual penetrating arterioles.

(A) Under the guidance of multiphoton imaging, a single penetrating arteriole is visualized and

occluded with a second fixed green laser following intravenous injection of rose bengal, a

photosensitizing agent. (B) Example of in vivo multiphoton images of a single penetrating arteriole

before and after occlusion. (C) A cortical microinfarct visualized with in vivo T2-weighted MRI.

Despite a large difference in brain size, rodent microinfarcts are similar in absolute size to human

microinfarcts due to a relatively conserved scale of microvascular topology between species. (77,78)

Figure 5. Functional deficits in tissues beyond the non-viable microinfarct core.

Peri-lesional deficits are caused by secondary effects of ischemic injury, such as spreading

depression, blood-brain barrier (BBB) disruption and neuroinflammation. Remote deficits arise when

microinfarcts damage white matter fibers, or occur in areas that are connected to and critical for the

function of other brain regions.

Supplemental Figure 1. Additional examples of cortical cerebral microinfarcts, captured at 3T and

1·5T MRI.

Supplemental Figure 2. A 3D representation of total cortical cerebral microinfarct distribution as

detected on 3T MRI in a memory-clinic population. Microinfarcts are represented by red dots in a (A)

transversal, (B) sagittal, and (C) coronal view of the brain. Re-printed with permission from (31).