Human Brain Connectomics In Ischemic Leukoaraiosis:based On Multi-Modal MRI

BackgroundVascular cognitive impairment(VCI)is a kind of syndrome from mild cognitive impairment to dementia caused by cerebrovascular diseases.Vascular dementia(Va D),in particular,seriously damages the daily living ability and social function of patients.Subcortical vascular cognitive impairment(SVCI)is the most important subtype of VCI.Importantly,cerebral small vessel disease(CSVD)is the most important cerebrovascular pathological cause of SVCI in the elderly,especially in patients with white matter hyperintensities(WMHs)and multiple subcortical lacunar infarction(LI);it is also the most common cause of clinical Va D(account for 36% to 67%).Notably,cognitive impairment caused by CSVD is more likely to progress Va D,with an average annual dementia progression rate of 2%-10%.For example,LI and WMHs can increase the incidence of ischemic stroke by 4 times and 2 times,respectively;and then increase the incidence of Va D by 5 times.But so far,the etiology and pathogenesis of cognitive dysfunction caused by CSVD are still unknown.Therefore,early diagnosis and timely intervention are the key measures to delay the progression of CSVD and SVCI.At present,it has become a hotspot and focus of this field to explore the neural circuit mechanism of CSVD patients with cognitive dysfunction and the biological predictive markers of their progression to SVCI.SVCI has been considered as a "disconnection syndrome" due to the extensive damage to the white-matter tracts caused by CSVD.Many previous studies have shown that CSVD subtypes can influence and promote each other.This suggests that researchers need to study the neurobiological mechanism of SVD-induced cognitive dysfunction based on a whole-brain level.Importantly,recent years,researchers have proposed "brain network dysfunction" as the best explanatory model for understanding the biological mechanism of SVCI,which suggested that abnormal brain network function may be involved in the pathogenesis and progression of SVCI,and lead to cognitive impairment.Therefore,it is necessary to explore the association between cognitive impairment and neural circuit changes in patients at high risk of SVCI(e.g.,CSVD patients)based on the brain network level,which is of great significance for the early diagnosis and prevention of SVCI.In recent years,multi-modal magnetic resonance imaging(multi-modal MRI)technology has become an ideal tool for studying brain networks.Moreover,numerous neuroimaging studies have revealed abnormal brain structure and functional connectivity in patients with SVCI and CSVD.However,it is worth noting that the human brain is a complex network of interconnected and interacting systems that can continuously integrate information from different brain regions.Neuroimaging studies in the past decade have shown that multimodal MRI data can be used to construct human brain structural and functional networks,and these networks can be measured by graph theory analysis to quantitatively describe the attributes of brain networks.Using these measures,human brain networks have been found to have many important organizational characteristics,including "small-world" properties,modular structures,and highly connected core nodes(hubs).Importantly,many studies have found abnormalities in the topological organization of the whole brain structural and functional network in SVCI patients,including loss of "small-world" attributes and redistribution of core brain regions.However,the study of brain structure and functional network topological organization structure based on CSVD population is still in its infancy.Therefore,to explore the changes in the topological and organizational structure of the whole brain structure and functional network based on CSVD population not only provides a new perspective for understanding the pathophysiological mechanism of SVCI,but also may find potential biological early warning markers for the early diagnosis of SVCI.Importantly,LI and WMHs have consistently been demonstrated to be associated with converted neurological and cognitive symptoms,when compared with the other features.Previous studies have indicated that ILA is characterized by cognitive impairment mainly involved in attention,processing speed and executive function.Therefore,in this study,we selected ILA patients,which is clinically homogeneous,and normal controls.For all subjects,we assessed cognitive function using multi-domain neurocognitive neuropsychological tests and constructed their whole-brain structural and functional networks by using the diffusion tensor imaging(DTI)and resting-state functional MRI(R-fMRI)data respectively.(1)To investigate whether the topological organization structure of the whole brain functional network was abnormal in ILA patients,and to further investigate whether the abnormal topological properties were associated with the reduced network connection integrity and neuropsychological tests.(2)To further explore the structural integrity of the whole brain white matter fiber tracts and the whole brain structural network in ILA patients.Multi-scale network analysis was performed at the level of whole brain network and local brain region to explore the coupling changes of functional and structural connections in ILA patients.(3)Graph theory analysis based on voxel level was used to comprehensively study ILA related changes in the core regions of the global functional network.We tried to determine two questions: i)whether the connection pattern of central nodes of the global functional network in ILA patients is broken,and whether this disruption depends on the distance of connection;ii)whether such topological changes in functional core areas are significantly associated with behavioral characteristics of ILA patients.Part 1 Altered Topological Patterns of Whole-Brain Functional Networks in ILAObjective: We aimed to investigate the topological organization of whole-brain networks in patients with ILA combing R-fMRI and graph-theory approaches,and further to examine the relationship between topological aberrations and cognitive performances.Methods: A total of 36 patients with ILA(Fazekas rating score ≥ 2)and 31 healthy controls underwent comprehensive neuropsychological assessments(covering 4 cognitive domains,i.e.,information processing speed,episodic memory,executive function and visuospatial function)and R-fMRI scans.To construct the brain functional network,the images of each brain was parcellated into 90 regions of interest(ROIs)using the automated anatomically labeling atlas.To measure interregional resting-state functional connectivity,Pearson correlation coefficients between any pair of ROIs were calculated.Each absolute correlation matrix was then thresholded into a binary matrix with a fixed sparsity level.Finally,for the constructed brain networks at each sparsity threshold,we calculated the global network measures(e.g.,small-world,network efficiency).For statistical analysis,nonparametric permutation tests first were used for group comparisons of topological metrics.To investigate the clinical relevance of altered brain network topologies in the ILA group,we then conducted multiple linear regression analyses to examine the relationships between the neuropsychological measures and the topological properties.Results:(1)Compared with the controls,the ILA patients showed abnormal global topology in their brain functional networks(i.e.,increased shortest path length and decreased network efficiency).(2)Moreover,network-based statistic(NBS)analysis revealed a functional-disconnected network in ILA,which is comprised of functional connections linking different brain modules(i.e.,default-mode,frontoparietal,ventral attention and limbic systems)and connections within single modules(i.e.,ventral attention and limbic systems).Intriguingly,the network disconnections showed significant associations with the topological disorganization of the brain connectome in ILA patients.(3)Finally,we showed that the abnormal network metrics correlated with cognitive deficits in the ILA patients.Conclusion: Our findings provide further evidence to support the concept that ILA pathologies could disrupt brain connections,impairing network functioning,and cognition via a “disconnection syndrome”.Part 2 Aberrant Structural-Functional Coupling of Multiscale Brain Networks in ILAObjective: We collected R-fMRI and d MRI data from the same cohort of ILA patients and healthy controls,followed by graph-theory approaches to systematically examine the topological properties of the whole-brain structural connectivity networks(SCN),functional connectivity networks(FCN),and the SCN-FCN coupling.Methods: A total of 36 patients with ILA(Fazekas rating score ≥ 3)and 34 healthy controls(Fazekas rating score = 0)underwent comprehensive neuropsychological assessments and multi-modal MRI scans(including 3D,d MRI and R-fMRI).To construct the brain structural network,the images of each brain was parcellated into 90 regions of interest using the automated anatomically labeling atlas.Two regions were considered structurally connected if at least 3 fiber bundles with two endpoints were located in these two regions.Then,for the constructed fiber-number(FN)weighted networks,we calculated both global and regional network measures(e.g.,small-world,network and nodal efficiency).For statistical analysis,nonparametric permutation tests were first conducted for group comparisons of topological metrics.To investigate the clinical relevance of altered brain network topologies in the ILA group,we then conducted multiple linear regression analyses to examine the relationships between the neuropsychological measures and the topological properties.Finally,to evaluate the structural-functional coupling of multiscale brain networks,we further used a sparsity-based thresholding approach.Each absolute correlation matrix or FN-matrix was thresholded into a binary matrix with a fixed sparsity level(i.e.,sparsity = 30%).We explored the relationship between SC and FC matrices and networks at the following 3 levels(i.e.,whole-brain connectivity level,small-worldness level,and nodal level).Nonparametric permutation tests were also used for group comparisons of SC-FC coupling.Results:(1)The ILA patients performed significantly worse than the controls in episodic memory,executive function,and information processing speed.Differences in information processing speed and executive function remained significant after covarying for episodic memory and visuospatial function.(2)The ILA patients exhibited abnormal global topology in their structural brain networks(i.e.,increased shortest path length and decreased network efficiency)when compared with the healthy controls.(3)The ILA patients showed decreased nodal efficiencies,predominately in the prefrontal-subcortical and default-mode regions.(4)Intriguingly,we showed that the topological aberrations correlated with the neuropsychological performances in the ILA patients.(5)The brain structural networks of ILA were equally as robust to random failures as those of healthy controls,but more vulnerable against targeted attacks.(6)Finally,at the whole-brain connectivity level,the strength of the whole-brain SC-FC coupling showed significant decrease in the ILA patients than the controls.At the nodal level,we found that a significantly decreased coupling degree located in the prefrontal and default-mode regions.Importantly,the coupling degree of prefrontal regions correlated with the information processing speed,and executive function separately.Conclusion: These results demonstrate that ILA is reflected in a selective aberrant topological organization in multiscale brain networks and may improve our understanding of the pathophysiological mechanisms underlying ILA-related cognitive impairment.Part 3 Identifying and Mapping Connectivity Patterns of Brain Functional Network Hubs in ILAObjective: We employed R-fMRI data and voxel-based graph-theory analysis to systematically investigate intrinsic functional connectivity patterns of whole-brain networks on ILA patients.And,we further sought to determine whether this distrution is connection-distance-dependent.Methods: A total of 36 patients with ILA(Fazekas rating score ≥ 3)and 34 healthy controls(Fazekas rating score = 0)underwent comprehensive neuropsychological assessments and R-MRI scans.We employed two voxel-wise network centrality measures,degree centrality(DC)and eigenvector centrality(EC),to quantify locally and globally functional integrity of the brain connectome.In this current study,we further calculated the Euclidean distance,and then divided whole-brain functionall connectivity maps into 18 bins with Euclidean distances binned into 10 mm steps,ranging from 0 to 180 mm.Thus,for each subject,a new DC map was produced for each distance bin.Voxel-wise one-way analyses of covariance were separately performed to examine between-group differences in DC and EC maps,with age,gender and years of education as covariates.Finally,we performed multiple linear regression analyses to examine the relationships between cognitibe performances and network centrality values(i.e.,DC and EC)in brain areas showing significant between-group differences in the ILA patients.Results:(1)The ILA patients showed lower DC values in the prefrontal-subcortical,default-mode and occipital regions.In addition,comparted with the healthy controls,the ILA patients have lower EC values in the regions mainly located in the default-mode network(i.e.,ventromedial prefrontal cortex and posterior cingulate cortex).(2)We noted that the DC maps showed similar patterns at the neighboring distance bins,but were very different between very short and long distances.For example,the between-group differences results showed decreased DC in ILA were primarily located in the ventromedial prefrontal cortex and precuneus at 30-40 mm distance,but in the prefrontal-subcortical,ventromedial prefrontal and occipital regions at the 130-140 mm distance.Notably,the most sigficant ILA-related DC decreases apperared in the 130-140 mm range,suggesting that ILA was mainly associated with longer distance disconnections.(3)Voxel-based multiple linear regression analyses revealed that the long-range DC values in the left prefrontal cortex and short-range DC values in the right precuneus positively correlated with the information processing speed scores in ILA patients.Conclusion: These findings indicated that the disrupted brain hubs patterns in ILA are connection-distance-dependent,being characterized by the disruptions of longer distance connections,and providing novel insights into the pathophysiological mechanisms of connectivity dysfunction in ILA.