Aberrant Spontaneous Brain Activity in Patients with Mild Cognitive Impairment and concomitant Lacunar Infarction: A Resting-State Functional MRI Study

Abstract

Background:

Lacunar infarctions (LI) have been associated with a cognitive decline and an increased risk of dementia. Whether and how the pattern of spontaneous brain activity in patients with mild cognitive impairment (MCI) differs in subjects with and without concomitant LI remains unclear.

Objective:

To compare the pattern of spontaneous brain activity in MCI patients with versus those without LI using resting-state functional magnetic resonance imaging (rs-fMRI).

Methods:

Forty-eight MCI patients, including 22 with LI [MCI-LI] and 26 without LI [MCI-no LI], and 28 cognitive normal subjects underwent rs-fMRI post-processed using regional homogeneity (ReHo) and the amplitude of low-frequency fluctuation (ALFF) methods.

Results:

Compared with cognitively normal subjects, the MCI-LI patients had decreased ReHo in the precuneus/cuneus (Pcu/CU) and insula; decreased ALFF in the Pcu/CU and frontal lobe; and increased ALFF and ReHo in the temporal lobe. While the MCI-no LI group had increased ReHo and ALFF in the bilateral hippocampus and parahippocampal gyrus, frontal lobe, and decreased ALFF and ReHo in the temporal lobe. Compared with the MCI-no LI patients, those with MCI-LI had decreased ALFF in the frontal lobe; decreased ReHo in the Pcu/CU and insula; and increased ALFF and ReHo in the temporal lobe (p < 0.05, AlphaSim corrected). In MCI-LI patients, the MOCA scores showed a relatively weak correlation with ALFF values in the medial frontal gyrus (r = 0.432, p = 0.045) (of borderline significance after Bonferroni correction).

Conclusions:

The spontaneous brain activities in MCI-LI were distinct from MCI-no LI. The probable compensatory mechanism observed in MCI-no LI might be disrupted in MCI with LI due to vascular damage.

Keywords

Amplitude of low-frequency fluctuation lacunar infarction mild cognitive impairment regional homogeneity

INTRODUCTION

Lacunar infarction (LI) is an important predictor of cognitive decline [1] and may be one of underlying causes of mild cognitive impairment (MCI) [2]. The cognitive decline in MCI patients having concomitant LI involves all major cognitive domains including attention, working memory, executive function, memory, language, and visuospatial function [3]. Epidemiological studies have reported an increased risk to develop dementia in subjects with LI by 11%–23% [4], and even more pronounced in recurrent lacunar events [5]. The number of elderly with LI is expected to increase over the next decades; thus, better insight into how LI influences brain activity and functioning in MCI patients is critical. This may help the implementation of appropriate therapeutic interventions, since the timely treatment of vascular diseases at the stage of MCI could prevent or delay the progression to dementia [6].

Resting-state functional magnetic resonance imaging (RS-fMRI) signals reflect a spontaneous neuronal activity [7]. So far, most RS-fMRI studies have investigated low frequency (0.01–0.08 Hz) fluctuations from the perspective of temporal correlation or synchronization between distinct brain areas, i.e., seed/region based functional connectivity or independent component analysis (ICA), rather than from the perspective of regional activity. Recently, utilizing ICA method, a disrupted connectivity of the default mode network and salience network in patients with silent LI was reported [8]. Although abnormalities in functional connectivity can reveal the comprehensive and integrative characteristics of two or more remote brain areas, they cannot assess which region in particular contributes to the observed abnormality. Analysis of the amplitude of low-frequency fluctuation (ALFF) and regional homogeneity (ReHo) are two useful methods to study the various characteristics of the initial signals on the whole brain level from different perspective. ReHo indicates the temporal similarity between a single voxel and its neighboring voxels and reflects the coordinated function of the spontaneous neural activity [9], while ALFF reflects the intensity of neural activity within a single-voxel [10]. Previous studies have used either ReHo or ALFF methods to study spontaneous brain activity in MCI patients [11, 12]. However, a combination of these two methods, applied to patients with attention-deficit/hyperactivity disorder [13], diabetes mellitus [14], or primary monosymptomatic nocturnal enuresis [15], concluded that both methods are complementary to each other, i.e., ALFF yields the total power within the range of 0.01 and 0.1 Hz and ReHo characterizes the similarity of local brain activity across a region and may provide more information about the pathophysiological framework in the human brain than either method alone. In this study, we combined ALFF and ReHo analyses to investigate global spontaneous neural activity in MCI patients with LI versus those without. We hypothesized that 1) the spontaneous brain activity of MCI-LI group would differ from that of MCI-no LI group due to the concomitant lacunar infarction; 2) the abnormalities in spontaneous brain activity would be associated with cognitive performance.

MATERIALS AND METHODS

Subjects

The study was approved by the ethical committee of the Affiliated Drum Tower Hospital of Nanjing University Medical School, and the methods were carried out in accordance with the approved guidelines. All subjects or their proxies signed an informed consent form. From August 2012 to July 2014, 55 MCI patients with and without LI were included in the study. All participants were recruited from the Memory clinic of Neurology Department in the Affiliated Drum Tower Hospital, Nanjing, China.

The MCI patients were diagnosed based on criteria by Petersen et al. [16] for the broad definition of MCI: (a) memory complaint, preferably confirmed by an informant; (b) objective memory impairment, adjusted for age and education; (c) normal or near-normal performance on general cognitive functioning and no or minimum impairment of daily life activities; (d) the Clinical Dementia Rating (CDR) score of 0.5; and (e) not meeting the criteria for dementia according to the DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4rd edition, revised). Patients with MCI were diagnosed by experienced neurologists.

Assessment of lacunar infarctions

Presence of chronic subcortical LI was derived from fluid attenuated inversion recovery (FLAIR) sequence using the University of Edinburgh Neuroimaging Stroke Scale [17] with corresponding hypointensity on diffusion weighted imaging. That is, a focal parenchymal lesions (diameter between 3 and 15 mm), with the same signal characteristics as cerebrospinal fluid on FLAIR, and with a hyperintense rim on the FLAIR images when located supratentorially [18]. 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) [19]. Only subjects with subcortical infarcts that were in the basal ganglia, thalamus, internal and external capsule, corona radiate, and deep white matter of the frontal, parietal, occipital, and temporal regions were included in the present study. Of the 22 patients with lacunar infarctions, six had multiple lacunar infarcts and two of them reported a history of clinical stroke or transient ischemic attack. Lesions distribution was depicted using BrainNet Viewer [20].

Two experienced radiologists blinded to the diagnosis performed the lacunar infarction assessment separately. Consensus was obtained through discussion between the two raters.

Among the 55 MCI patients, those with concomitant lacunar infarction were defined as MCI-LI, and the rest without lacunar infarction were defined as MCI-no LI.

For comparison, we recruited 28 cognitively normal subjects (CNs) through advertisements in the hospital area. The CNs were identified as individuals who (a) had no cognitive complaints, (b) were scored normally on global cognitive scales including Montreal Cognitive Assessment (MoCA), Mini-Mental State Examination (MMSE) and CDR, and (c) had no evidence of any structural abnormality on a conventional MRI.

All subjects were self-identified as right-handed.

We excluded subjects with (a) primary neurological, psychiatric, or systemic disease associated with a cognitive impairment such as brain tumor, Parkinson’s disease, trauma, epilepsy, HIV, metabolic and endocrine disorders, (b) severe depression (Hamilton Depression Rating Scale (HAMD) score >17 points), (c) any brain lesions other than lacunar infarcts assessed on basis of medical history, (d) white matter disease/hyperintensities (Fazekas scale ≥2) assessed on both FLAIR and T2-weighted images, (e) excessive head motion within the scanner (more than 3.0 mm maximum translation in any of the x, y or z directions or 3.0° of maximum rotation about three axes), (f) subcortical infarction >15 mm in diameter, infratentorial or cortical infarction, and acute infarction. Seven patients were excluded for cerebral microbleeds (n = 3), head motion more than 3.0 mm or 3.0° (n = 3), and cavernous hemangioma (n = 1). Finally, 48 MCI patients, n = 22 with lacunar infarction (MCI-LI) and n = 26 without LI (MCI-no LI) were included in the final analysis.

Neuropsychological assessment

Cognitive performance was assessed using global cognitive or functioning scales such as MMSE, MoCA, and CDR. The MoCA was used to screen subjects with MCI and to assess their general cognition. Recorded results from the MMSE and the global score of the CDR scale were used to match the disease groups for severity of cognitive impairment [21]. The patient’s functional status was evaluated using the activities of daily living (ADL). The HAMD was used to rate the severity of a patient’s major depression. All participants completed the tests within 3 days before or after MR examination.

Image acquisition

Images were acquired on a 3 Tesla MR scanner (Achieva 3.0 T TX dual Medical Systems; Philips Medical Systems, Eindhoven, Netherlands). All subjects were placed in an 8-channel phased array coil and fitted to foam padding to reduce head motion, and a pair of earplugs was used to reduce scanner noise. They were instructed to hold still, keep eyes closed but stay awake during the scanning time. Compliance with these instructions was verified as part of the exit interview.

A three-dimensional turbo fast echo (3D-TFE) T₁WI sequence with high resolution with TR/TE = 9.8/4.6 ms; flip angle, 8°; in-plane resolution of 1.0 mm², FOV = 256×256 mm², matrix = 256×256, and a slice thickness of 1 mm was performed for spatial normalization in the preprocessing step. A gradient-echo echo-planar (GRE-EPI) sequence sensitive to blood oxygenation level dependent (BOLD) contrast was used to acquire functional images (TR = 2000 ms, TE = 30 ms, flip angle = 90°, FOV = 192×192 mm², matrix = 64×64, slice thickness = 4 mm). Each brain volume comprised 35 axial slices and each functional image contained 240 volumes. The sections were approximately aligned with the anterior commissure–posterior commissure line and covered about –30 to 60 mm in the inferior-superior direction. Each fMRI scan lasted 487 s.

Functional image pre-processing

Testing state data were analyzed using Data Processing Assistant for Resting State fMRI Advanced edition (DPARSFA_V2.2; http://www.restfmri.net) and Resting-State fMRI Data Analysis Toolkit (REST1.8; http://www.restfmri.net). Programs were run using Statistical Parametric Mapping (SPM8; http://www.fil.ion.ucl.ac.uk/spm), based on the Matlab platform (The Mathworks Inc., USA).

The first ten volumes were discarded for the signal equilibrium and participants’ adaptation to the scanning. The remaining 230 volumes were further analyzed. Standard steps of slice timing, head motion correction, and spatial normalization to the standard Montreal Institute (MNI) template using T₁ segment information with a resampled voxel size of 3*3*3 mm³ were performed. No participant had head motion of more than 3.0 mm maximum translation in any of the x, y or z directions or 3.0 degree of maximum rotation about three axes during scanning. We also evaluated the group differences in translation and rotation of head motion according to the formula [19]:

$\begin{matrix} Head Motion / Rotation = \frac{1}{L - 1} \sum_{i = 2}^{L} \sqrt{{| x_{i} - x_{i - 1} |}^{2} + {| y_{i} - y_{i - 1} |}^{2} + {| z_{i} - z_{i - 1} |}^{2}} \end{matrix}$ where L is the length of the time series (L = 230 in this study), xi,yi and zi are translations/rotations at the ith time point in the x, y and z directions, respectively. The results showed that the three groups had no significant differences (one - way analysis of variance, F = 1.61, p = 0.2077 for translational motion and F = 1.67, P = 0.1953 for rotational motion). Spatial smoothing was performed with a Gaussian kernel of 6 mm full-width at half-maximum (FWHM) before ALFF but after ReHo calculation, since smoothing before ReHo calculation substantially increases the regional similarity [22]. Finally, the REST toolkit (http://resting-fmri.sourceforge.net) was used to remove the linear trend of time courses and for temporally band-pass filtering (0.01–0.08 Hz) that reduces low-frequency drift and physiological high frequency respiratory and cardiac noise.

ReHo analysis

Both the ALFF and ReHo analyses were performed for each subject using REST toolkit. The ReHo value (also named Kendall’s coefficient of concordance; KCC) was calculated to measure the similarity of the ranked time series of a given voxel in relation to its nearest [23] neighbor voxels. Through calculating the ReHo (KCC) value of every voxel in the whole brain, an individual ReHo map was obtained for each subject. The intracranial voxels were extracted to make a mask [23]. For standardization purposes, each individual ReHo map was divided by its own global mean ReHo value within the respective mask.

ALFF analysis

For a given voxel, the filtered time course was first converted to the frequency domain using fast Fourier transform to acquire the power spectrum. Since the power at a given frequency is proportional to the square of the amplitude of this frequency component, the square root at each frequency of the power spectrum was computed, and the mean square root was obtained across 0.01–0.08 Hz within each voxel. This mean square root represented the ALFF [10]. For standardization purposes, the ALFF of each voxel was divided by the global mean ALFF value.

Statistics analysis

Statistical analysis was performed using SPSS software version 16.0 (statistical program for social sciences, SPSS Inc. Chicago, IL) for demographic and clinical data, and SPM8 (statistical parametric mapping, http://www.fil.ion.ucl.ac.uk/spm) for fMRI data.

To assess the ReHo and ALFF differences across the groups (MCI-LI, MCI-no LI patients, and CNs), a one-way analysis of variance (ANCOVA) was performed on the individual normalized ReHo and ALFF maps in a voxel-wise manner within a gray matter mask. Age, gender, years of education, head motion parameters, and modulated grey matter maps obtained from T₁ segmentation were included as covariates in all functional data analyses. The result was corrected using the Alphasim program, which setting at p < 0.01 and cluster size >999 mm³, which corresponded to a corrected p < 0.05. If statistical difference was present, post hoc t-tests were performed to detect the inter-group differences within respective brain regions.

As the hippocampus and parahippocampal gyrus were consistent early prominent involvement in mild cognitive impairment and dementia [24], we would like to further compare the exact ReHo and ALFF values of these two regions among the three groups. We extracted the values of these ROIs resulted from ANCOVA analysis and analyzed group differences using SPSS, at the alpha level p < 0.05, and corrected for multiple comparison using the Bonferroni correction. Finally, regions showing significant differences in the ANCOVA were extracted as regions of interest (ROIs). ReHo and ALFF values were subsequently extracted from these seed regions within each subject, Pearson correlations were calculated to measure the association between the group mean ReHo and ALFF values within ROIs and MoCA, MMSE, and HAMD scores at alpha p < 0.05 (Bonferroni corrected).

RESULTS

Demographics and clinical data for MCI patients and CNs are listed in Table 1. The groups did not differ on gender (p = 0.172), but the MCI-LI was significantly older than MCI-no LI (p = 0.001) and CNs (p < 0.001) groups, and the education level in MCI-LI group was lower than that in MCI-no LI (p = 0.001) and CNs (p = 0.012) groups. None of the subjects showed a depressive state evaluated by the HAMD. On global cognitive and functional measures of MOCA, MMSE, and ADL, both the MCI with and without LI scored lower than CNs group (p < 0.001), whereas the MCI with LI did not differ from those without LI group on these scores. Locations of lacunar infarctions are displayed in Fig. 1 and in brief as follows: insular cortex (n = 1), frontal lobe (n = 2), basal ganglia (n = 10); thalamus (n = 7), and corona radiate (n = 8). Because of the small sample size, we did not separate the subgroup due to the locations of LI, to discuss the impact of locations of LI on cognition.

The MCI patients with and without LI had different patterns of spontaneous brain activity alteration. Compared with CNs, the MCI-LI patients had decreased ReHo values in the precuneus/cuneus and insula, decreased ALFF values in the precuneus/cuneus and medial frontal gyrus (MFG), and increased ALFF and ReHo in the inferior temporal lobe (ITL). Compared with the CNs, the MCI-no LI group had increased ReHo and ALFF values in the bilateral hippocampus and parahippocampal gyrus, superior frontal gyrus (SFG), anterior cingulate cortex (ACC), and MFG, increased ALFF values in the left caudate, and decreased ALFF and ReHo in the superior temporal gyrus (STG). Compared with the MCI-no LI patients, those with MCI-LI had decreased ALFF values in the SFG, ACC, and MFG, decreased ReHo in the precuneus/cuneus and insula, and increased ALFF and ReHo in the ITL (Fig. 2, Tables 2 and 3) (p < 0.05, AlphaSim corrected). Since the MCI-LI group was significantly older than MCI-no LI (p = 0.001) and CNs, the effect of age on the clusters of significant differences was investigated with and without age as covariate. The ALFF and ReHo results were approximately consistent with those with age correction (see Supplementary Figs. 1 and 2).

The ALFF and ReHo in the bilateral hippocampus and parahippocampal gyrus (Hip&PHG) showed an increasing trend from CN to MCI-LI to MCI-no LI, only the difference between CN and MCI-no LI reached the statistical significance (p < 0.05, Bonferroni corrected) (Fig. 3 and Table 4).

In MCI-LI patients, the MOCA scores showed a relatively weak correlation with ALFF values in the MFG (r = 0.432, p = 0.045) (of borderline significance after Bonferroni corrected) (Fig. 4). No other significant correlations were found between ReHo or ALFF values and cognitive or HAMD scores within other patient groups.

DISCUSSION

In the present study, we found a unique pattern of brain activity alterations in MCI patients with LI. At the same level of global cognition, the MCI-LI showed mainly diffusely decreased spontaneous brain activity in the temporal, frontal, and occipital lobe, while the MCI-no LI group showed mainly increased activity in frontal lobe and bilateral Hip&PHG. The lower ALFF in MFG in MCI-LI patients showed weak correlation with worse global cognitive performance. Moreover, in terms of the analytic method, both the ReHo and ALFF detected spontaneous brain activity alterations in Hip&PHG, with ReHo mainly in the occipital area and ALFF in the frontal area.

Previously, ReHo and ALFF have been used to investigate the intrinsic neuropathology of various neurological disorders [25, 26]. In healthy subjects, Zang et al. found that regions showing a higher ReHo [9] and ALFF [10, 27] are topically consistent with location of the default mode network, and have the highest metabolic rate as measured by positron emission tomography. These two methods are based on different neurophysiology mechanisms, with ALFF demonstrating neural intensity and ReHo demonstrating the coherence of neural activity regardless intensity. Since the two methods found some changes in common cerebral functional regions, both were adopted to reduce inaccuracies and to provide reliable and comprehensive conclusions. In our study, ALFF and ReHo findings were consistent in showing activity alterations in bilateral Hip&PHG in MCI-no LI patients compared with CNs. The coexisting abnormalities of functional intensity and activity coherence in these two regions might represent more severe functional changes than those brain areas reflected by a single method. In keeping with that, ReHo and ALFF analysis showed different regional involvement between the MCI-LI and MCI-no LI patients in this study, suggesting they are complementary to each other in depicting regional spontaneous neural activity and provide different perspective in the understanding of the physiopathological effect of lacunar infarction in MCI patients.

Compared with CNs, the MCI-LI patients showed relatively spared and even decreased ALFF values in frontal area mainly in the MFG, whereas the MCI-no LI patients showed increased both ReHo and ALFF values in frontal areas including SFG, MFG, and ACC. The increased spontaneous brain activity in the frontal regions of MCI-no LI patients was partially consistent with previous findings. For example, several task-related [28, 29] and rs-fMRI [30, 31] studies have demonstrated increased prefrontal activation during specific cognitive tasks or increased frontal brain activities during resting state in AD or MCI patients, all of which have been proposed as a compensatory reallocation or recruitment of cognitive resources in AD or MCI patients. The relatively spared ReHo and even decreased ALFF in the MFG might indicate that the compensatory or recruitment capacity was lessened or even impaired in MCI-LI group. In our study, only 2 of 22 LIs located in the frontal lobe, we speculated that the decreased brain activity in MFG might be related to the disruption of fronto-subcortical circuits but not the LI per se. It has been established that fronto-subcortical circuits are damaged in subcortical vascular dementia, resulting in cognitive impairment (e.g., executive dysfunction, inertia, disinhibition) [32]. Alteration of brain activity in MFG may be associated with cognitive decline in MCI-LI patients, and the weak positive correlation between the ALFF of the MFG and the scores of MoCA partially supports this notion.

Medial temporal lobe (MTL), including the hippocampus, parahippocampal gyrus, and entorhinal cortex, is critical in the process of information storage and retrieval [33]. Using both ReHo and ALFF analyses, hyperactivity of the Hip&PHG in MCI-no LI patients was identified. We can interpret this as a compensatory process. The compensatory hypothesis is thought to take place in patients with AD dementia, as well as in individuals at risk for AD [34]. However, previous studies have reported inconsistent results on brain activity in MTL, i.e., decreased activity, hypometabolism [35 –37], or no significant changes compared with CN [38]. These differences could be due to the different population characteristics, different MCI subtypes, different location of infarction lesions, and technical methodological issues. To the best of our knowledge, this study was the first to combine ALFF and ReHo methods to investigate the differences in spontaneous brain activity in MCI-LI, MCI-no LI, and CNs. Although no significant difference between MCI-LI versus MCI-no LI and versus CNs was found, the ReHo and ALFF values of Hip&PHG showed an increasing trend from CNs to MCI-LI to MCI-no LI group, suggesting the limbic system impairment was not only specific for MCI and AD, but also present in MCI of vascular origin and might be milder than in AD [39]. Our findings are at least partially consistent with those showing structural atrophy of hippocampus in vascular dementia [21, 40], i.e., the finding of thinner cortex in the medial temporal lobe in subjects with clinically silent lacunar infarction [41] can represent a link between neurodegenerative and cerebrovascular disease. Associations or interactions between cerebrovascular disease and neurodegenerative factors have been previously reported and may underlie these findings [42].

In addition to the frontal area, MCI-LI patients showed decreased regional coherence or synchronization in the insula, precuneus/cuneus, and lingual gyrus. Previously, insular atrophy has been observed in patients with LI [8], and the insula has been functionally integrated in memory process in MCI and AD patients [43]. The findings reveal that vascular dysfunction including LI might increase the risk for the development of AD. Cognitive impairment in subcortical vascular dementia is probably related to ischemic interruption of frontal subcortical circuits [44] or disruption of cholinergic pathways that traverse the subcortical white matter [45]. Cholinergic pathway dysfunction is known to adversely affect short-term episodic memory, attention, and executive functions [45]. Moreover, the perisylvian division of the lateral cholinergic pathway is known to originate from the basal forebrain and project to opercular areas including superior temporal gyrus and insula cortex [46]. Thus, decreased regional synchronization of the insular area in our MCI-LI patients might be related to this disruption of the lateral cholinergic pathway. The PCu and adjacent PCC are mainly involved in episodic memory processing [47] and are critical nodes in human brain structural and functional networks [48]. Many neuroimaging studies have consistently demonstrated parallel structural and functional abnormalities in both AD and MCI patients, such as cortical thinning [49], metabolic pathologies [37], and disruptions in spontaneous or intrinsic brain activity [36]. Thus, our findings of decreased ALFF and ReHo values in MCI-LI subjects were in accordance with those previous studies.

Our study had some limitations including the relatively small sample size. Due to the small sample size, we could not separate the subgroups focusing on cognitive domain, such as aMCI and non-aMCI, to discuss whether fMRI changes are the result of LI or combined LI and neurodegenerative burden. Second, cognitive impairment in MCI might also be related to other factors such as depression state and premorbid IQ, quantified white matter lesions in MCI patients with LI, since the HAMD in our study did not show correlations with brain activity, further studies were needed to assess these factors on cognitive impairment. Last, different topography of lacunar infarctions may have different effects on brain activity, so we excluded the cortical and infratentorial infarctions, the relationship between location and onset time with the brain activity were needed with larger sample in the future.

In conclusion, fMRI measuring different patterns of brain activity alterations has a potential to reflect different underlying pathology of subjects with MCI syndrome. Findings of increased activity in the MCI-no LI group and diffused decreased activity in the MCI-LI group indicates that the possible compensatory mechanism observed in MCI-no LI was impaired in MCI-LI due to vascular damage. In terms of the analysis methods, the ReHo and ALFF methods provide different views on neurophysiology mechanisms and they are complementary to each other in depicting regional spontaneous neural activity. Therefore, combination of these two fMRI methods has a potential help to better understand underlying pathophysiological and compensatory mechanism.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the National Science Foundation of China, (2014–2016, 81300925, B.Z.), the Provincial Natural Science Foundation of Jiangsu (2014–2016, BK20131085, B.Z.), and Jiangsu Province Medical Key talent people and “the 12th five years plan for China development” (2011-2-16, RC2011013, B.Z.).

Authors’ disclosures available online ().

Appendix

References

Vermeer

, Prins

, den Heijer

, Hofman

, Koudstaal

, Breteler

(2003) Silent brain infarcts and the risk of dementia and cognitive decline. N Engl J Med 348, 1215–1222.

Grau-Olivares

, Arboix

, Bartres-Faz

, Junque

(2007) Neuropsychological abnormalities associated with lacunar infarction. J Neurol Sci 257, 160–165.

Edwards

, Jacova

, Sepehry

, Pratt

, Benavente

(2013) A quantitative systematic review of domain-specific cognitive impairment in lacunar stroke. Neurology 80, 315–322.

Barba

, Martinez-Espinosa

, Rodriguez-Garcia

, Pondal

, Vivancos

, Del

(2000) Poststroke dementia: Clinical features and risk factors. Stroke 31, 1494–1501.

Morris

, Waite

, Williams

, Neal

, Blake

(2012) Recent advances in the genetics of the ALS-FTLD complex. Curr Neurol Neurosci Rep 12, 243–250.

Chang

, Park

, Ohn

, Park

, Lee

, Kim

(2011) Neural correlates of donepezil-induced cognitive improvement in patients with right hemisphere stroke: A pilot study. Neuropsychol Rehabil 21, 502–514.

Raichle

, Snyder

(2007) A default mode of brain function: A brief history of an evolving idea.1083-1090. Neuroimage 37, 1097–1099.

Chen

, Wang

, Zhang

, Chen

, Fleisher

, Wang

, Zhang

(2014) Aberrant functional networks connectivity and structural atrophy in silent lacunar infarcts: Relationship with cognitive impairments. J Alzheimers Dis 42, 841–850.

Zang

, Jiang

, Lu

, He

, Tian

(2004) Regional homogeneity approach to fMRI data analysis. Neuroimage 22, 394–400.

10.

Zang

, He

, Zhu

, Cao

, Sui

, Liang

, Tian

, Jiang

, Wang

(2007) Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI. Brain Dev 29, 83–91.

11.

Zhang

, Liu

, Jiang

, Zhou

, An

, Dai

, Wang

, Niu

, Wang

, Zhang

(2012) Altered spontaneous activity in Alzheimer’s disease and mild cognitive impairment revealed by regional homogeneity. Neuroimage 59, 1429–1440.

12.

Wang

, Yan

, Zhao

, Qi

, Zhou

, Lu

, He

, Li

(2011) Spatial patterns of intrinsic brain activity in mild cognitive impairment and Alzheimer’s disease: A resting-state functional MRI study. Hum Brain Mapp 32, 1720–1740.

13.

, Cao

, Sui

, Sun

, Zou

, Zang

, Wang

(2013) Local synchronization and amplitude of the fluctuation of spontaneous brain activity in attention-deficit/hyperactivity disorder: A resting-state fMRI study. Neurosci Bull 29, 603–613.

14.

Cui

, Jiao

, Chen

, Wang

, Gao

, Wen

, Ju

, Teng

(2014) Altered spontaneous brain activity in type 2 diabetes: A resting-state functional MRI study. Diabetes 63, 749–760.

15.

Lei

, Ma

, Du

, Shen

, Tian

, Li

(2012) Spontaneous brain activity changes in children with primary monosymptomatic nocturnal enuresis: A resting-state fMRI study. Neurourol Urodyn 31, 99–104.

16.

Petersen

, (2004) Mild cognitive impairment as a diagnostic entity. J Intern Med 256, 183–194.

17.

Sharma

, Hasson

, Butcher

(2001) How well does the Oxfordshire Community Stroke Project classification predict the site and size of infarct of brain imaging? J Neurol Neurosurg Psychiatry 70, 567.

18.

Vermeer

, Koudstaal

, Oudkerk

, Hofman

, Breteler

(2002) Prevalence and risk factors of silent brain infarcts in the population-based Rotterdam Scan Study. Stroke 33, 21–25.

19.

Bokura

, Kobayashi

, Yamaguchi

(1998) Distinguishing silent lacunar infarction from enlarged Virchow-Robin spaces: A magnetic resonance imaging and pathological study. J Neurol 245, 116–122.

20.

Xia

, Wang

, He

(2013) BrainNet Viewer: A network visualization tool for human brain connectomics. PLoS One 8, e68910.

21.

van de Pol

, Gertz

, Scheltens

, Wolf

(2011) Hippocampal atrophy in subcortical vascular dementia. Neurodegener Dis 8, 465–469.

22.

Chao-Gan

, Yu-Feng

(2010) DPARSF: A MATLAB Toolbox for “Pipeline” Data Analysis of Resting-State fMRI. Front Syst Neurosci 4, 13.

23.

Smith

(2002) Fast robust automated brain extraction. Hum Brain Mapp 17, 143–155.

24.

Apostolova

, Dutton

, Dinov

, Hayashi

, Toga

, Cummings

, Thompson

(2006) Conversion of mild cognitive impairment to Alzheimer disease predicted by hippocampal atrophy maps. Arch Neurol 63, 693–699.

25.

, Qi

, Zhang

, Zhong

, Zheng

, Zhang

, Zhong

, Xu

, Liao

, Jiao

, Wu

, Fan

, Lu

(2012) Altered regional homogeneity in the development of minimal hepatic encephalopathy: A resting-state functional MRI study. PLoS One 7, e42016.

26.

, Zhang

, Wu

, Zhong

, Zhang

, Zhong

, Ni

, Zhang

, Li

, Jiao

, Wu

, Fan

, Liu

, Lu

(2012) Altered resting-state brain activity at functional MR imaging during the progression of hepatic encephalopathy. Radiology 264, 187–195.

27.

Yang

, Long

, Yang

, Yan

, Zhu

, Zhou

, Zang

, Gong

(2007) Amplitude of low frequency fluctuation within visual areas revealed by resting-state functional MRI. Neuroimage 36, 144–152.

28.

Gould

, Arroyo

, Brown

, Owen

, Bullmore

, Howard

(2006) Brain mechanisms of successful compensation during learning in Alzheimer disease. Neurology 67, 1011–1017.

29.

Grady

, McIntosh

, Beig

, Keightley

, Burian

, Black

(2003) Evidence from functional neuroimaging of a compensatory prefrontal network in Alzheimer’s disease. J Neurosci 23, 986–993.

30.

, Wang

, Zang

, Tian

, Zhang

, Li

, Jiang

(2007) Regional coherence changes in the early stages of Alzheimer’s disease: A combined structural and resting-state functional MRI study. Neuroimage 35, 488–500.

31.

Wang

, Yan

, Zhao

, Qi

, Zhou

, Lu

, He

, Li

(2011) Spatial patterns of intrinsic brain activity in mild cognitive impairment and Alzheimer’s disease: A resting-state functional MRI study. Hum Brain Mapp 32, 1720–1740.

32.

Seo

, Ahn

, Yoon

, Im

, Lee

, Tae

, Ahn

, Chin

, Jeong

, Na

(2010) Cortical thinning in vascular mild cognitive impairment and vascular dementia of subcortical type. J Neuroimaging 20, 37–45.

33.

Buckner

, Andrews-Hanna

, Schacter

(2008) The brain’s default network: Anatomy, function, and relevance to disease. Ann N Y Acad Sci 1124, 1–38.

34.

Han

, Wang

, Zhao

, Min

, Lu

, Li

, He

, Jia

(2011) Frequency-dependent changes in the amplitude of low-frequency fluctuations in amnestic mild cognitive impairment: A resting-state fMRI study. Neuroimage 55, 287–295.

35.

, Zhao

, Wang

, Guo

, Jiang

, Cao

, He

, Yan

(2012) Spontaneous brain activity in mild cognitive impairment revealed by amplitude of low-frequency fluctuation analysis: A resting-state fMRI study. Radiol Med 117, 865–871.

36.

Han

, Wang

, Zhao

, Min

, Lu

, Li

, He

, Jia

(2011) Frequency-dependent changes in the amplitude of low-frequency fluctuations in amnestic mild cognitive impairment: A resting-state fMRI study. Neuroimage 55, 287–295.

37.

Seo

, Cho

, Park

, Chin

, Na

(2009) Subcortical vascular versus amnestic mild cognitive impairment: Comparison of cerebral glucose metabolism. J Neuroimaging 19, 213–219.

38.

Bai

, Zhang

, Yu

, Shi

, Yuan

, Zhu

, Zhang

, Qian

(2008) Default-mode network activity distinguishes amnestic type mild cognitive impairment from healthy aging: A combined structural and resting-state functional MRI study. Neurosci Lett 438, 111–115.

39.

, Schuff

, Laakso

, Zhu

, Jagust

, Yaffe

, Kramer

, Miller

, Reed

, Norman

, Chui

, Weiner

(2002) Effects of subcortical ischemic vascular dementia and AD on entorhinal cortex and hippocampus. Neurology 58, 1635–1641.

40.

Scher

, Xu

, Korf

, Hartley

, Witter

, Scheltens

, White

, Thompson

, Toga

, Valentino

, Launer

(2011) Hippocampal morphometry in population-based incident Alzheimer’s disease and vascular dementia: The HAAS. J Neurol Neurosurg Psychiatry 82, 373–376.

41.

Thong

, Hilal

, Wang

, Soon

, Dong

, Collinson

, Anh

, Ikram

, Wong

, Venketasubramanian

, Chen

, Qiu

(2013) Association of silent lacunar infarct with brain atrophy and cognitive impairment. J Neurol Neurosurg Psychiatry 84, 1219–1225.

42.

Iadecola

(2010) The overlap between neurodegenerative and vascular factors in the pathogenesis of dementia. Acta Neuropathol 120, 287–296.

43.

Xie

, Bai

, Yu

, Shi

, Yuan

, Chen

, Li

, Chen

, Zhang

, Li

(2012) Abnormal insula functional network is associated with episodic memory decline in amnestic mild cognitive impairment. Neuroimage 63, 320–327.

44.

Tierney

, Black

, Szalai

, Snow

, Fisher

, Nadon

, Chui

(2001) Recognition memory and verbal fluency differentiate probable Alzheimer disease from subcortical ischemic vascular dementia. Arch Neurol 58, 1654–1659.

45.

Bocti

, Swartz

, Gao

, Sahlas

, Behl

, Black

(2005) A new visual rating scale to assess strategic white matter hyperintensities within cholinergic pathways in dementia. Stroke 36, 2126–2131.

46.

Selden

, Gitelman

, Salamon-Murayama

, Parrish

, Mesulam

(1998) Trajectories of cholinergic pathways within the cerebral hemispheres of the human brain. Brain 121(Pt 12), 2249–2257.

47.

Buckner

, Andrews-Hanna

, Schacter

(2008) The brain’s default network: Anatomy, function, and relevance to disease. Ann N Y Acad Sci 1124, 1–38.

48.

Buckner

, Sepulcre

, Talukdar

, Krienen

, Liu

, Hedden

, Andrews-Hanna

, Sperling

, Johnson

(2009) Cortical hubs revealed by intrinsic functional connectivity: Mapping, assessment of stability, and relation to Alzheimer’s disease. J Neurosci 29, 1860–1873.

49.

, Chen

, Evans

(2008) Structural insights into aberrant topological patterns of large-scale cortical networks in Alzheimer’s disease. J Neurosci 28, 4756–4766.