You are here

Article Text

Article menu
Cerebrovascular disease
Research paper
Association of silent lacunar infarct with brain atrophy and cognitive impairment
  1. Jamie Yu Jin Thong1,
  2. Saima Hilal2,
  3. Yanbo Wang1,
  4. Hock Wei Soon1,
  5. Yanhong Dong2,
  6. Simon Lowes Collinson3,
  7. Tuan Ta Anh1,
  8. Mohammad Kamran Ikram4,5,
  9. Tien Yin Wong4,5,
  10. Narayanaswamy Venketasubramanian6,
  11. Christopher Chen2,
  12. Anqi Qiu1,7,8
  1. 1Department of Bioengineering, National University of Singapore, Singapore, Singapore
  2. 2Department of Pharmacology, National University of Singapore, Singapore, Singapore
  3. 3Department of Psychology, National University of Singapore, Singapore, Singapore
  4. 4Singapore Eye Research Institute, National University of Singapore, Singapore, Singapore
  5. 5Department of Ophthalmology, National University of Singapore, Singapore, Singapore
  6. 6Division of Neurology, University Medicine Cluster, National University Health System, Singapore
  7. 7Singapore Institute for Clinical Sciences, the Agency for Science, Technology and Research, Singapore, Singapore
  8. 8Clinical Imaging Research Centre, National University of Singapore, Singapore, Singapore
  1. *Correspondence to Dr Anqi Qiu, Department of Bioengineering, National University of Singapore, 9 Engineering Drive 1, Block EA 03-12, Singapore 117576, Singapore; bieqa{at}nus.edu.sg

Abstract

Objective Silent lacunar infarct (SLI) is associated with cognitive decline and linked to an increased risk of stroke and dementia. We examined the association of SLI with MRI measures of cortical thickness, subcortical and lateral ventricular shapes and cognition in 285 ethnic Chinese elderly.

Methods SLI, cortical thickness, shapes of subcortical and ventricular structures were quantified using MRI. The cognitive performance was assessed using comprehensive neuropsychological tests. Linear regression was used to examine associations among SLI, brain measures and cognition.

Results SLI was associated with atrophy in multiple subcortical structures, ventricular enlargement and widespread cortical thinning. Both SLI and atrophy were independently related to poorer performance in attention, memory and language domains. Only SLI was associated with visuomotor speed and executive function, while atrophy mediated the association between SLI and visuoconstruction.

Conclusions Our findings support a vascular contribution to neurodegeneration and cognitive impairment.

  • Stroke
  • Neuropsychology
  • Neuroanatomy
  • MRI

https://doi.org/10.1136/jnnp-2013-305310

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Silent lacunar infarcts (SLI) are often detected with MRI or CT, but are not accompanied by any overt symptoms. SLI is now known to be linked to an increased risk of stroke and dementia.1 Additionally, SLI is associated with reductions in global cognitive ability and specific cognitive domains such as memory and psychomotor speed.2 ,3 Blum et al4 recently reported that the hippocampus is smaller in elderly people with SLI than in controls and that SLI contributes to memory deficits independent of hippocampal atrophy. Blum et al4 focused only on hippocampal atrophy; however, infarcts in subcortical regions can also lead to hypoperfusion and hypometabolism in both subcortical and cortical regions5–7 and may lead to global neuronal loss.8 Nonetheless, it remains largely unknown to what extent SLI is associated with cortical and subcortical atrophy in the elderly. It is also unclear if SLI as well as cortical and subcortical atrophy independently contribute to impairment in multiple cognitive domains (eg, executive function, attention, language, memory, visuomotor speed and visuoconstruction) or to which extent atrophy mediates the relationship between SLI and impairment in specific cognitive domains.

    Here, we employed an advanced brain mapping technique, large deformation diffeomorphic metric mapping (LDDMM),9 to examine cortical thickness as well as subcortical and lateral ventricular shapes in association with SLI and cognition in Chinese elderly. We hypothesised that SLI would be associated with widespread cortical atrophy and that subcortical shape abnormalities in SLI would be observed not only in the hippocampus but also in the basal ganglia and thalamus. Moreover, in the view of the work in Blum et al,4 we anticipated independent contributions of SLI to memory loss as well as to cognitive deficits in other domains, including executive function, attention, language, visuomotor speed and visuoconstruction, beyond the contributions of cortical and subcortical atrophy.

    Methods

    Subjects

    The present study involved 330 ethnic Chinese subjects drawn from the Epidemiology of Dementia in Singapore (EDIS) study, which is a population-based study among Chinese, Malays and Indians. In the present study, we restricted analysis to the Chinese component of EDIS, as the recruitment of the other ethnicities is still ongoing. In the first phase of the EDIS recruitment, Chinese participants aged >60 years underwent cognitive screening using the Abbreviated Mental Test (AMT) and a self-report of progressive forgetfulness. The AMT is a 10-item self-report scale that was initially designed to rapidly assess the elderly for the possibility of dementia. It is now used widely as a screening tool for cognitive impairment.10 This screening tool has been previously validated in Singapore.11 ,12 Screen positives were defined as AMT score ≤6 among those with ≤6 years of formal education or ≤8 among those with >6 years of formal education, or if the subject or caregiver reported progressive forgetfulness. Screen-positive subjects were invited to take part in the second phase of this study. During the second phase of this study, participants underwent extensive clinical and neuropsychological examination along with laboratory tests and MRI. The recruitment procedure, the details of the EDIS study methodology and an accompanying flowchart depicting the recruitment flow have been described elsewhere.13 Ethics approval for EDIS was obtained from the Singapore Eye Research Institute and National Healthcare Group Institutional Review Boards. Informed consent was obtained for all participants prior to recruitment.

    Neuropsychological assessment

    Patients were assessed using a formal neuropsychological battery locally validated for older Singaporeans,14 which was administered by trained research psychologists. The non-memory domains of the formal neuropsychological battery included: (1) attention (digit span test,15 visual span test15 and auditory detection test16); (2) language (15-item modified Boston Naming test and category fluency17); (3) visuomotor speed (symbol digit modalities18 and digit cancellation19); (4) visuoconstruction (visual reproduction subtest of the Wechsler Memory Scale-Revised copy task,15 clock drawing20 and the block design subtest of the Wechsler Adult Intelligence Scale-Revised21); (5) executive function (frontal assessment battery22 and maze23). The memory domains of the battery included: (1) verbal memory (word list12 and story recall15); (2) visual memory (picture recall15 and the visual reproduction subtest of the Wechsler Memory Scale-Revised15). Z-scores were derived for individual subtests. All z-scores were adapted so that a greater value reflects better performance. Z-scores for individual domains were computed by summing up the Z-scores of each subtest and dividing by the number of the subtests under that domain. Domain specific Z-scores were used to compute the final global cognitive composite score. The visual and verbal memory scores were combined into a composite memory score. Only subjects who completed all the tasks were included in statistical analysis on cognition described below.

    MRI acquisition

    Subjects underwent MRI scans on a 3 T Siemens Magnetom Trio Tim scanner using a 32-channel head coil at the Clinical Imaging Research Center, National University of Singapore. The image protocols were: (1) high-resolution T1-weighted Magnetization Prepared Rapid Gradient Recalled Echo (MPRAGE; 192 slices, 1 mm thickness, in-plane resolution 1×1 mm2, no interslice gap, sagittal acquisition, FOV=256×256 mm, matrix=256×256, TR=2300 ms, TE=1.9 ms, TI=900 ms, flip angle=9°); (2) fluid-attenuated inversion recovery (FLAIR) (TR=9000 ms, TE=82 ms, TI=2500 ms, matrix=232×256, FOV=232×256 mm, slice thickness=3 mm, no interslice gap, number of slices=48). The acquisition times of the MPRAGE and FLAIR images were respectively 5 min 20 s and 3 min 36 s.

    Assessment of subcortical SLI on MRI

    Subcortical SLI were graded on FLAIR using the University of Edinburgh Neuroimaging Stroke Scale.24 They were defined as focal parenchymal lesions ≥3 mm and <15 mm in size, with the same signal characteristics as cerebrospinal fluid (CSF) on FLAIR and T1-weighted images, and with a hyperintense rim on the FLAIR images when located supratentorially.25 There was no involvement of cortical grey matter (GM). Differentiation from Virchow-Robin (VR) spaces was based on signal intensity (absence of hyperintense rim on FLAIR images), shape (VR-spaces are more linear or lobulated in shape) and location (VR-spaces are often located around anterior commissure or near vertex of the brain).26 If there was no prior clinical history of stroke, these lacunar infarcts were considered to be ‘silent’. Only subjects with subcortical infarcts that were in the basal ganglia, thalamus, internal and external capsule, brainstem and deep white matter (WM) of the frontal, parietal, occipital and temporal regions were included in the present study.

    Subcortical and lateral ventricular shape and cortical thickness analysis

    The lateral ventricles, subcortical structures (amygdala, hippocampus, caudate, putamen, globus pallidus, thalamus), cortical GM, WM and CSF were automatically segmented from the intensity-inhomogeneity corrected T1-weighted MRI.27 ,28 For shape analysis, the subcortical and lateral ventricular shapes were generated using the prior shape information of an atlas that were created from 41 manually labelled individual structures via a LDDMM atlas generation procedure.29 Shape variations of individual subjects relative to the atlas were characterised by the Jacobian determinant of the deformation in the logarithmic scale, where the deformation transformed the atlas shape to be similar to subjects. This measure, termed as the ‘deformation map’, represents the ratio of each subject's structural volume to the atlas volume in the logarithmic scale: that is, positive values correspond to expansion, while negative values correspond to compression of the subject's structure relative to the atlas at each anatomical location.

    For cortical thickness, an inner surface was constructed at the boundary between WM and GM and then propagated to the outer surface at the boundary between GM and CSF. The cortical thickness was measured as the distance between the corresponding points on the inner and outer surfaces.30 A brain mapping algorithm, multimanifold LDDMM (MM-LDDMM), that incorporates the cortical surface and eight gyral curves per cortical hemisphere (including the precentral gyrus, postcentral gyrus, superior temporal gyrus, medial temporal gyrus, lingual gyrus, cuneus gyrus, precuneus and paracentral gyrus) was then applied to align individual cortical surfaces to an atlas cortical surface for group analysis of cortical thickness.31

    Statistical analysis

    Comparisons of demographic variables between those with and without SLI used independent sample t tests for continuous variables, assuming inhomogeneity of variance due to large differences in sample size between the SLI group and controls, and χ2 tests for categorical variables.

    Analysis of covariance was used to investigate differences in volume between the SLI group and controls for the subcortical structures (amygdala, caudate, globus pallidus, hippocampus, putamen, thalamus) and lateral ventricles. Models were adjusted for age, sex, years of education, total brain volume (TBV), diabetes, intracranial stenosis (defined as >50% intraluminal narrowing in at least one intracranial artery through visual inspection of MRA) and history of smoking. The two groups had similar rates of hypertension and hyperlipidaemia; thus these variables were not included in the model.

    SurfStat Toolbox in Matlab32 was used to examine group differences in cortical thickness and subcortical shapes between SLI subjects and controls. Briefly, the cortical thickness and deformation maps were first smoothed using heat kernel smoothing detailed in reference.33 Linear regression, in which the diagnostic group was a main factor and age, sex, years of education, TBV, history of diabetes, intracranial stenosis and smoking history were entered as covariates, was then performed at each vertex of the cortical or subcortical surface. Finally, the statistical results were corrected for multiple comparisons at the level of significance (corrected p<0.05).

    The associations between SLI and cognition as well as between brain atrophy and cognition were examined using a series of multiple regressions. Brain atrophy was computed as average cortical thickness or shape deformation of subcortical structures in the regions with significant associations with SLI in each cerebral lobe (or subcortical structures). Separate models evaluated each cognitive domain score as dependent variables. SLI or brain atrophy was entered as the primary independent variable. Subsequently, a series of multiple regression analysis examined whether SLI and brain atrophy had independent contributions to cognitive decline or if brain atrophy mediated cognitive decline in SLI. For this, SLI and brain atrophy were simultaneously entered as the primary independent variables. All models were adjusted for age, sex, years of education, TBV, diabetes, intracranial stenosis and smoking history as these factors were significantly different between subjects with and without SLI. All variables were entered into the equation in one step, and all analyses were performed on SPSS V.19 for Windows unless otherwise specified.

    Results

    Clinical and demographic characteristics

    Among screened positives (n=642), 330 ethnic Chinese subjects agreed to participating in phase II of this study. Among the 330 subjects, the present study further excluded forty-five subjects due to history of stroke or transient ischaemic attack (n=26), silent cortical or cerebellar infarct (n=6), silent haemorrhagic stroke (n=5), evidence of gliotic changes on MRI due to head injuries (n=2), brain surgery (n=1) and incomplete MRI data (n=5). Hence, the present study only consisted of 285 subjects. The excluded subjects were similar in age (t329=0.54, p=0.593), years of education (t329=1.17, p=0.246), Mini Mental Status Examination (MMSE) score (t329=1.48, p=0.146), AMT score (t329=1.14, p=0.257) and sex (χ2=3.44, p=0.064) to those included in the present study.

    Demographic information for the study sample is shown in table 1. There were a total of 34 subjects with SLI and 251 controls without SLI. Compared with controls, subjects with SLI were similar in terms of gender (χ2=0.171, p=0.728), but were significantly older (t284=3.46, p=0.001), and had fewer years of education (t284=3.14, p=0.003), lower MMSE scores (t284=4.79, p<0.001), lower AMT scores (t284=2.49, p<0.017) and smaller TBV (t284=2.23, p=0.031). There were no differences between groups for history of cardiovascular disease (χ2=0.042, p=0.838), history of hypertension (χ2=3.79, p=0.052) and history of hyperlipidaemia (χ2=1.01, p=0.314). However, those with SLI had higher rates of diabetes (χ2=6.30, p=0.012), intracranial stenosis (χ2=27.32, p<0.001) and smoking history (χ2=4.51, p=0.034).

    View this table:
    Table 1

    Clinical and demographic characteristics of the study sample

    Subcortical and lateral ventricular volumes and shapes

    After controlling for age, sex, years of education, TBV, diabetes, intracranial stenosis and smoking history, SLI subjects had significantly smaller amygdala (F1277=10.17, p=0.002), hippocampus (F1277=8.00, p=0.003) and globus pallidus volumes (F1277=7.89, p=0.005) and significantly larger lateral ventricle volumes (F1277=17.67, p<0.001) compared with controls (table 2). There were no differences in the thalamus (F1277=0.44, p=0.510), caudate (F1277=1.26, p=0.263) and putamen volumes (F1277=2.35, p=0.126).

    View this table:
    Table 2

    Volumes of subcortical structures and the lateral ventricles

    Relatively symmetric shape differences in the lateral ventricles and multiple subcortical structures were observed between SLI subjects and controls (figure 1). Compared with controls, SLI subjects showed volume loss bilaterally in the amygdala, globus pallidus, anterior hippocampus and the caudate head, as well as in the left putamen and the right posterior thalamus. Furthermore, the ventricular enlargement was seen bilaterally in SLI subjects when compared with controls. Interestingly, our shape analysis revealed that the enlargement in the inferior lateral ventricles filled the gap where the anterior hippocampus and amygdala shrunk.

    Figure 1
    Figure 1

    Statistical maps of subcortical compression and ventricular expansion in subjects with silent lacunar infarcts. Warm colours denote regions with compression in silent lacunar infarcts relative to controls, while cool colours denote regions of expansion. Voxelwise analyses were corrected for multiple comparisons (p<0.05 corrected). All analyses were adjusted for age, sex, total intracranial volume, diabetes, stenosis and smoking history. Am, amygdala; Cd, caudate; Hp, hippocampus; iLV, inferior lateral ventricles; Pa, globus pallidus; Pu, putamen; Th, thalamus; V, ventricles.

    Cortical thickness

    Compared with controls, SLI subjects showed relatively asymmetric patterns of cortical thickness abnormalities (figure 2). In the left hemisphere, the SLI group showed widespread cortical thinning in the posterior regions of the cortex and small regions of the frontal lobe. The thinning was particularly dominant in the precuneus, posterior cingulate cortex, lingual gyrus and fusiform cortex. In the right hemisphere, areas affected were primarily in the temporal pole, fusiform, parahippocampal, middle temporal and lingual gyri, insula and inferior parietal lobule. Smaller regions of thinning were observed in the lateral occipital cortex, supramarginal, superior temporal and superior frontal gyri. In contrast, the SLI group had greater cortical thickness in the frontal areas, namely, the left medial orbitofrontal cortex and bilateral dorsolateral prefrontal cortex.

    Figure 2
    Figure 2

    Statistical maps of cortical thickness in subjects with silent lacunar infarcts. Warm colours denote regions with cortical thinning in silent lacunar infarcts relative to controls, while cool colours denote regions with cortical thickening. Voxelwise analyses were corrected for multiple comparisons (p<0.05 corrected). All analyses were adjusted for age, sex, total intracranial volume, diabetes, stenosis and smoking history.

    Associations of SLI and brain atrophy with cognition

    Among 285 subjects, only two controls were not able to complete the neuropsychological testing and hence were excluded in statistical analysis for examining associations of SLI and brain atrophy with cognition.

    When SLI was the primary independent variable, multiple regression analyses revealed that SLI was negatively associated with all cognitive domains (table 3, Part I). This suggested that subjects with SLI performed worse in all cognitive domains in comparison with controls. On the other hand, when brain atrophy was the primary independent variable, multiple regression analyses revealed that distinct patterns of cortical thinning and hippocampal shape shrinkage were associated with worse performance in specific cognitive domains (table 3, Part II). In details, frontal and occipital thinning were both associated with attention and memory; occipital thinning was also associated with visuoconstruction; temporal thinning was associated with attention, language, memory and visuoconstruction; and parietal cortical thinning was only associated with memory. Furthermore, hippocampal atrophy was associated with worse performance in memory, language and attention. Lateral ventricular enlargement was associated with worse performance in all cognitive domains except executive function. However, no association of brain atrophy or the hippocampus was found with executive function.

    View this table:
    Table 3

    Unstandardised β weights for the association of silent lacunar infarctions (SLIs) and brain atrophy with scores of individual cognitive domains

    Interestingly, our results also showed that (1) brain atrophy and SLI independently contribute to declines in attention, language and memory, (2) only SLI contributes to declines in executive function and visuomotor speed and (3) temporal and occipital cortical thinning, as well as lateral ventricular enlargement mediate the contribution of SLI to declines in visuoconstruction, once both SLI and atrophy are considered (table 4). Furthermore, when atrophy and SLI were simultaneously entered into the multiple regressions, frontal cortical thinning was no longer associated with declines in attention and memory, hippocampal atrophy was no longer related to attention, and ventricular enlargement was no longer associated with poorer visuomotor speed. These results emphasise the contribution of SLI in these cognitive domains.

    View this table:
    Table 4

    Unstandardised β weights for examining contributions of silent lacunar infarctions (SLIs) and brain atrophy with scores of individual cognitive domains

    The amygdala, thalamus, caudate, putamen and globus pallidus atrophy were not associated with scores on any cognitive domains and are hence not shown in tables 3 and 4.

    Discussion

    This study found associations of SLI with diffuse cortical thinning, ventricular enlargement and shape shrinkage in multiple subcortical structures. Also, distinct patterns of cortical thinning and subcortical shape shrinkage that were observed in subjects with SLI were associated with declines in numerous cognitive domains. Interestingly, SLI independently contributed to impairment in all cognitive domains except the visuoconstruction. Our findings lend weight to the emerging concept that SLI contributes to widespread brain atrophy as well as to impairment in multiple cognitive domains.

    Our findings corroborate a recent study that found reductions in the hippocampal volume in the elderly with SLI4 while extending the findings to the amygdala, thalamus, caudate, putamen and globus pallidus. We also observed a previously unreported finding: cortical thinning in the SLI group involved many posterior areas of the brain. This is compatible with findings from patients with symptomatic lacunar infarcts where volume reductions in the thalamus and hippocampus, as well as widespread atrophy in multiple posterior cortical regions, including the middle temporal gyrus and posterior cingulate cortex, have previously been shown.34 Moreover, volume reductions in the amygdala were also found in patients with a stroke or transient ischaemic attack.35 Finally, our findings on cortical thickness in SLI are in line with existing evidence of cortical GM loss related to vascular pathology.36

    The shape analysis employed in this study allowed for the precise localisation of differences within the individual subcortical and lateral ventricular structures above and beyond what has been examined in prior MRI studies on stroke patients. Indeed, the shape analysis revealed that the subcortical and ventricular shape abnormalities due to SLI were similar to what was observed in a study of patients with mild cognitive impairment and Alzheimer's disease (AD) that used the same method of shape analysis.37 CSF in the inferior lateral ventricles filled the gap where the anterior hippocampus and amygdala shrunk. In addition, we found cortical thinning in the medial temporal lobe in SLI, which has been identified as a structural hallmark of AD.38 Hence, these new observations in SLI subjects hint at the possibility of a neurodegenerative component that is active in SLI.

    SLI has previously been linked to declines in the global cognition (ie, MMSE), memory, executive function and perceptual speed,2 ,3 ,39 which were also seen in our study. Beyond these findings, we additionally provided evidence that the SLI-related cortical and subcortical atrophy was associated with impairment in specific cognitive domains. For instance, frontal cortical thinning was associated with poorer attention and memory. This is consistent with the role of the frontal cortex. Also, occipital thinning was linked with declines in the attention, memory and visuoconstruction. The affected area was primarily the lateral occipital cortex, a key structure in object recognition, visual perception, information capture and visualisation.40 ,41 Furthermore, there is a well-established relationship between performance on memory tasks and the size of the hippocampus and medial temporal lobe,42 which was also shown in our study.

    Our findings suggest that SLI and atrophy may independently contribute to cognitive impairment in the elderly. This is consistent with the findings of Blum and colleagues who reported that hippocampal atrophy and infarcts both independently contribute to memory decline in the elderly.4 We also showed that numerous cortical areas were associated with memory. Even though cortical thinning (especially in the medial temporal lobe) is generally thought to be a result of neurodegeneration related to AD or aging,43 ,44 these cortical regions may also be affected by vascular disease. Associations or interactions between cerebrovascular disease and AD have been previously reported and may underlie these findings.45

    In our sample, SLI was located in the subcortical regions distal from the frontal cortex. Despite this, we observed that only SLI was associated with executive function deficits. Additionally, occipital thinning was directly linked with poorer performance in attention, memory and visuoconstruction; temporal cortical thinning was related to poorer language performance and no associations were found between cognitive performance and the thalamus. One explanation for these findings is that the observed cognitive impairment may be due to disruptions in WM tracts connecting brain areas that facilitate higher-order cognition rather than direct damage to, or atrophy in, the cortical areas responsible for cognitive impairment. Thus, SLI and small-vessel disease that contribute to WM disruptions may be affecting connectivity between neural networks necessary for facilitating cognition. Nevertheless, it is not possible to fully elucidate this hypothesis in the current study and future work investigating this possibility is recommended.

    Of note is the asymmetric pattern of cortical atrophy observed in this study. Atrophy was more severe and widespread in the right hemisphere compared with the left hemisphere. A possible explanation for the observed pattern was that there were a greater number of infarcts in the right hemisphere; nevertheless, the number of infarcts in each hemisphere was roughly equal in number and in average size in our SLI sample, ruling out this possibility. Here, a parallel may be drawn from the literature on aging. It has been suggested that the right hemisphere is more sensitive to damage from aging compared with the left.46 Vascular pathology, including silent infarctions, may well have greater impact on the right hemisphere than on the left hemisphere. However, it is not possible to elucidate this in the current study and hence future investigations are recommended.

    The observation of cortical thickening in the frontal cortex in SLI subjects is unexpected. Speculatively, this may be due to functional reorganisation of the brain, as the prefrontal regions compensate for cognitive decline arising from ‘challenges’ to the brain (in this case, SLI and atrophy).47 However, we did not find any association of frontal cortical thickening with cognitive impairment. More research is required in order to confirm this finding and to fully clarify the mechanisms.

    The strength of this study is that we employed advanced brain mapping techniques that revealed novel evidence of structural abnormalities associated with SLI. Also, the neuropsychological assessments performed were comprehensive and provided information on multiple cognitive domains. Several limitations require discussion. First, the SLI subjects were older than controls by 5 years. Although we used age-adjusted statistical models, our findings may be still influenced by the age difference. Second, the number of the SLI subjects was smaller than controls. To address these two issues, we conducted supplementary analyses comparing a subsample of age-matched controls to the SLI group. Our findings remained largely unchanged (see the online supplementary material). Additionally, while rates of hypertension were not significantly different between those with and without SLI, it is known to be an important risk factor for SLI.1 We hence ran additional analyses including hypertension as a covariate, but the results remained unchanged. Another limitation was the lack of a healthy control group of similar demographics against which we could benchmark cognitive performance in the SLI and control groups. Without this additional control group it is difficult to evaluate the ecological significance of the differences in cognition between SLI and control subjects found in this study. Furthermore, the number of SLI subjects (N=34) was far from ideal. This limits the generalisation of our findings as the sample may not have been sufficiently representative of the general population of SLI subjects. In addition to this, we may not have been able to detect more subtle changes in the brain or in cognition that would only manifest in a larger sample.

    In conclusion, our study showed that SLI is associated with widespread cortical thinning, subcortical atrophy and lateral ventricular enlargement. In turn, these brain structural abnormalities are associated with poorer performance on many cognitive domains in addition to memory. These findings have implications for the pathogenesis of cognitive impairment and dementia and support the idea of a vascular contribution to neurodegeneration and cognitive impairment.

    References

    1. ↵
      1. Vermeer SE,
      2. Longstreth WT Jr.,
      3. Koudstaal PJ
      . Silent brain infarcts: a systematic review. The Lancet Neurology 2007;6:611–19.
      OpenUrlCrossRef
    2. ↵
      1. Vermeer SE,
      2. Hollander M,
      3. van Dijk EJ,
      4. et al
      . Silent brain infarcts and white matter lesions increase stroke risk in the general population. Stroke 2003;34:1126–29.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Wright CB,
      2. Festa JR,
      3. Paik MC,
      4. et al
      . White matter hyperintensities and subclinical infarction. Stroke 2008;39:800–5.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Blum S,
      2. Luchsinger JA,
      3. Manly JJ,
      4. et al
      . Memory after silent stroke. Neurology 2012;78:38–46.
      OpenUrlCrossRef
    5. ↵
      1. Baron JC,
      2. D'Antona R,
      3. Pantano P,
      4. et al
      . Effects of thalamic stroke on energy metabolism of the cerebral cortex. Brain 1986;109:1243–59.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Kwan LT,
      2. Reed BR,
      3. Eberling JL,
      4. et al
      . Effects of subcortical cerebral infarction on cortical glucose metabolism and cognitive function. Arch Neurol. 1999;56:809–14.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Pappata S,
      2. Mazoyer B,
      3. Tran Dinh S,
      4. et al
      . Effects of capsular or thalamic stroke on metabolism in the cortex and cerebellum: a positron tomography study. Stroke 1990;21:519–24.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Farkas E,
      2. Luiten PGM
      . Cerebral microvascular pathology in aging and Alzheimer's disease. Prog Neurobiol 2001;64:575–611.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Miller MI,
      2. Qiu A
      . The emerging discipline of computational functional anatomy. NeuroImage 2009;45(1, Suppl 1):S16–39.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Hodkinson HM
      . Evaluation of a mental test score for assessment of mental impairment in the elderly. Age Ageing 1972;1:233–8.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Sahadevan S,
      2. Lim PiPJ,
      3. Tan NJL,
      4. et al
      . Diagnostic performance of two mental status tests in the older Chinese: influence of education and age on cut-off values. Int J Geriatr Psychiatry 2000;15:234–41.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Sahadevan S,
      2. Tan NJ,
      3. Tan T,
      4. et al
      . Cognitive testing of elderly Chinese people in Singapore: influence of education and age on normative scores. Age Ageing 1997;26:481–6.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Hilal S,
      2. Ikram MK,
      3. Saini M,
      4. et al
      . Prevalence of cognitive impairment in Chinese: epidemiology of dementia in Singapore study. J Neurol Neurosurg Psychiatry 2013.
    14. ↵
      1. Yeo D,
      2. Gabriel C,
      3. Chen C,
      4. et al
      . Pilot validation of a customized neuropsychological battery in elderly Singaporeans. Neurol J South East Asia 1997;2:123.
      OpenUrl
    15. ↵
      1. Wechsler D
      . Wechsler memory scale-revised. San Antonio, TX: Harcourt Brace Jovanovich, 1997.
    16. ↵
      1. Lewis R,
      2. Rennick PM
      . Manual for the repeatable cognitive perceptual-motor battery. Clinton Township, MI: Axon, 1979.
    17. ↵
      1. Isaacs B,
      2. Kennie AT
      . The set test as an aid to the detection of dementia in old people. Br J Psychiatry 1973;123:467–70.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Smith A
      . Symbol digit modalities test. Los Angeles, CA: Western Psychological Services, 1973.
    19. ↵
      1. Diller L
      . Studies in cognition and rehabilitation in hemiplegia. New York: Behavioral Science, Institute of Rehabilitation Medicine, New York University Medical Center, 1974.
    20. ↵
      1. Sunderland T,
      2. Hill JL,
      3. Mellow AM,
      4. et al
      . Clock drawing in Alzheimer's disease. A novel measure of dementia severity. J Am Geriatr Soc 1989;37:725–9.
      OpenUrlPubMedWeb of Science
    21. ↵
      1. Wechsler D
      . Wechsler adult intelligence scale-revised. New York, NY: Harcourt Brace Jovanovich, 1981.
    22. ↵
      1. Dubois B,
      2. Slachevsky A,
      3. Litvan I,
      4. et al
      . The FAB: a frontal assessment battery at bedside. Neurology 2000;55:1621–6.
      OpenUrlCrossRef
    23. ↵
      1. Porteus S
      . The maze test and clinical psychology. Palo Alto, CA: Pacific Books, 1959.
    24. ↵
      1. Mead GE,
      2. Lewis SC,
      3. Wardlaw JM,
      4. et al
      . How well does the Oxfordshire Community Stroke Project classification predict the site and size of the infarct on brain imaging? J Neurol Neurosurg Psychiatry 2000;68:558–62.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Vermeer SE,
      2. Koudstaal PJ,
      3. Oudkerk M,
      4. et al
      . Prevalence and risk factors of silent brain infarcts in the population-based Rotterdam scan study. Stroke 2002;33:21–5.
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Zhu Y-C,
      2. Tzourio C,
      3. Soumaré A,
      4. et al
      . Severity of dilated Virchow-Robin spaces is associated with age, blood pressure, and MRI markers of small vessel disease. Stroke 2010;41:2483–90.
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Sied JG,
      2. Zijdenbos AP,
      3. Evans AC
      . A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans Med Imaging 1998;17:87–97.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Dale AM,
      2. Fischl B,
      3. Sereno MI
      . Cortical surface-based analysis. I. Segmentation and surface reconstruction. Neuroimage 1999;9:179–94.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Qiu A,
      2. Brown T,
      3. Fischl B,
      4. et al
      . Atlas generation for subcortical and ventricular structures with its applications in shape analysis. Image Process IEEE Trans 2010;19:1539–47.
      OpenUrlCrossRef
    30. ↵
      1. Fischl B,
      2. Dale AM
      . Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci USA 2000;97:11050–5.
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Zhong J,
      2. Qiu A
      . Multi-manifold diffeomorphic metric mapping for aligning cortical hemispheric surfaces. NeuroImage 2010;49:355–65.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Chung MK,
      2. Worsley KJ,
      3. Nacewicz BM,
      4. et al
      . General multivariate linear modeling of surface shapes using SurfStat. Neuroimage 2010;53:491–505.
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Chung MK,
      2. Robbins SM,
      3. Dalton KM,
      4. et al
      . Cortical thickness analysis in autism with heat kernel smoothing. Neuroimage 2005;25:1256–65.
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Grau-Olivares M,
      2. Bartrés-Faz D,
      3. Arboix A,
      4. et al
      . Mild cognitive impairment after lacunar infarction: voxel-based morphometry and neuropsychological assessment. Cerebrovasc Dis 2007;23:353–61.
      OpenUrlCrossRefPubMed
    35. ↵
      1. Sachdev PS,
      2. Chen X,
      3. Joscelyne A,
      4. et al
      . Amygdala in stroke/transient ischemic attack patients and its relationship to cognitive impairment and psychopathology: the Sydney Stroke Study. Am J Geriatr Psych 2007;15:487–96
      OpenUrlCrossRef
    36. ↵
      1. Leritz EC,
      2. Salat DH,
      3. Williams VJ,
      4. et al
      . Thickness of the human cerebral cortex is associated with metrics of cerebrovascular health in a normative sample of community dwelling older adults. Neuroimage 2011;54:2659–71.
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Qiu A,
      2. Fennema-Notestine C,
      3. Dale AM,
      4. et al
      . Regional shape abnormalities in mild cognitive impairment and Alzheimer's disease. NeuroImage 2009;45:656–61.
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Desikan RS,
      2. Sabuncu MR,
      3. Schmansky NJ,
      4. et al
      . Selective disruption of the cerebral neocortex in Alzheimer's disease. PLoS ONE 2010;5:e12853.
      OpenUrlCrossRefPubMed
    39. ↵
      1. Aggarwal NT,
      2. Wilson RS,
      3. Bienias JL,
      4. et al
      . The association of magnetic resonance imaging measures with cognitive function in a biracial population sample. Arch Neurol 2010;67:475–82.
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Grill-Spector K,
      2. Kourtzi Z,
      3. Kanwisher N
      . The lateral occipital complex and its role in object recognition. Vision Res 2001;41:1409–22.
      OpenUrlCrossRefPubMedWeb of Science
    41. ↵
      1. Stokes M,
      2. Thompson R,
      3. Cusack R,
      4. et al
      . Top-down activation of shape-specific population codes in visual cortex during mental imagery.  J Neurosci 2009;29:1565–72.
      OpenUrlAbstract/FREE Full Text
    42. ↵
      1. Prull MW,
      2. Gabrieli JD,
      3. Bunge SA
      . Age-related changes in memory: a cognitive neuroscience perspective. In: Craik FI, Salthouse TA, eds. The handbook of aging and cognition. Mahwah, NJ: Lawrence Erlbaum Associates Publishers, 2000:91–153.
    43. ↵
      1. Fjell AM,
      2. Westlye LT,
      3. Amlien I,
      4. et al
      . High consistency of regional cortical thinning in aging across multiple samples. Cereb Cortex 2009;19:2001–12.
      OpenUrlAbstract/FREE Full Text
    44. ↵
      1. Frisoni GB,
      2. Fox NC,
      3. Jack CR,
      4. et al
      . The clinical use of structural MRI in Alzheimer disease. Nat Rev Neurol 2010;6:67–77. http://www.nature.com/nrneurol/journal/v6/n2/suppinfo/nrneurol.2009.215_S1.html
      OpenUrlCrossRefPubMedWeb of Science
    45. ↵
      1. Iadecola C
      . The overlap between neurodegenerative and vascular factors in the pathogenesis of dementia. Acta Neuropathol (Berl). 2010;120:287–96.
      OpenUrlCrossRefPubMedWeb of Science
    46. ↵
      1. Dolcos F,
      2. Rice HJ,
      3. Cabeza R
      . Hemispheric asymmetry and aging: right hemisphere decline or asymmetry reduction. Neurosci Biobehav Rev 2002;26:819–25.
      OpenUrlCrossRefPubMedWeb of Science
    47. ↵
      1. Park DC,
      2. Reuter-Lorenz P
      . The adaptive brain: aging and neurocognitive scaffolding. Annu Rev Psychol 2009;60:173–96.
      OpenUrlCrossRefPubMedWeb of Science

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:

    Footnotes

    • Contributors JYJT and AQ: study design, data analysis, interpretation of the data and drafting of the manuscript; SH: data acquisition and review on silent infarcts; YW, HWS, TTA: MRI data analysis; YD and SLC: cognitive study design, data acquisition and analysis; MKI, TYW, NV and CC: study design.

    • Funding This work was supported by: a centre grant from the National Medical Research Council (NMRC/CG/NUHS/2010), the Young Investigator Award at the National University of Singapore (NUSYIA FY10 P07), the National University of Singapore MOE AcRF Tier 1 and Singapore Ministry of Education Academic Research Fund Tier 2 (MOE2012-T2-2-130).

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval Ethics approval for EDIS was obtained from the Singapore Eye Research Institute and National Healthcare Group Institutional Review Boards. Informed consent was obtained for all participants prior to recruitment.

    • Provenance and peer review Not commissioned; externally peer reviewed.