Neuroimaging predictors of cognitive impairment in confluent white matter lesion: volumetric analyses of 99 brain regions

Identifying Leukoaraiosis with Mild Cognitive Impairment by Fusing Multiple MRI Morphological Metrics and Ensemble Machine Learning

Leukoaraiosis (LA) is strongly associated with impaired cognition and increased dementia risk. Determining effective and robust methods of identifying LA patients with mild cognitive impairment (LA-MCI) is important for clinical intervention and disease monitoring. In this study, an ensemble learning method that combines multiple magnetic resonance imaging (MRI) morphological features is proposed to distinguish LA-MCI patients from LA patients lacking cognitive impairment (LA-nCI). Multiple comprehensive morphological measures (including gray matter volume (GMV), cortical thickness (CT), surface area (SA), cortical volume (CV), sulcus depth (SD), fractal dimension (FD), and gyrification index (GI)) are extracted from MRI to enrich model training on disease characterization information. Then, based on the general extreme gradient boosting (XGBoost) classifier, we leverage a weighted soft-voting ensemble framework to ensemble a data-level resampling method (Fusion + XGBoost) and an algorithm-level focal loss (FL)-improved XGBoost model (FL-XGBoost) to overcome class-imbalance learning problems and provide superior classification performance and stability. The baseline XGBoost model trained on an original imbalanced dataset had a balanced accuracy (Bacc) of 78.20%. The separate Fusion + XGBoost and FL-XGBoost models achieved Bacc scores of 80.53 and 81.25%, respectively, which are clear improvements (i.e., 2.33% and 3.05%, respectively). The fused model distinguishes LA-MCI from LA-nCI with an overall accuracy of 84.82%. Sensitivity and specificity were also well improved (85.50 and 84.14%, respectively). This improved model has the potential to facilitate the clinical diagnosis of LA-MCI.

Corticospinal Tract Microstructure Predicts Distal Arm Motor Improvements in Chronic Stroke

BACKGROUND AND PURPOSE

The corticospinal tract (CST) is a crucial brain pathway for distal arm and hand motor control. We aimed to determine whether a diffusion tensor imaging (DTI)-derived CST metric predicts distal upper extremity (UE) motor improvements in chronic stroke survivors.

METHODS

We analyzed clinical and neuroimaging data from a randomized controlled rehabilitation trial. Participants completed clinical assessments and neuroimaging at baseline and clinical assessments 4 months later, postintervention. Using univariate linear regression analysis, we determined the linear relationship between the DTI-derived CST fractional anisotropy asymmetry (FAasym) and the percentage of baseline change in log-transformed average Wolf Motor Function Test time for distal items (ΔlnWMFT-distal_%). The least absolute shrinkage and selection operator (LASSO) linear regressions with cross-validation and bootstrapping were used to determine the relative weighting of CST FAasym, other brain metrics, clinical outcomes, and demographics on distal motor improvement. Logistic regression analyses were performed to test whether the CST FAasym can predict clinically significant UE motor improvement.

RESULTS

lnWMFT-distal significantly improved at the group level. Baseline CST FAasym explained 26% of the variance in ΔlnWMFT-distal_%. A multivariate LASSO model including baseline CST FAasym, age, and UE Fugl-Meyer explained 39% of the variance in ΔlnWMFT-distal_%. Further, CST FAasym explained more variance in ΔlnWMFT-distal_% than the other significant predictors in the LASSO model.

DISCUSSION AND CONCLUSIONS

CST microstructure is a significant predictor of improvement in distal UE motor function in the context of an UE rehabilitation trial in chronic stroke survivors with mild-to-moderate motor impairment.Video Abstract available for more insight from the authors (see the Video, Supplemental Digital Content 1, available at: http://links.lww.com/JNPT/A350).

Modified cerebral small vessel disease score is associated with vascular cognitive impairment after lacunar stroke

We conducted a cross-sectional study to characterize the relationship between total and modified small vessel disease (SVD) score with vascular cognitive impairment (VCI). Patients (n = 157) between the ages of 50 and 85 years old who had suffered their first lacunar infarction were analyzed prospectively. Brain magnetic resonance imaging was performed to identify SVD manifestations, which were used to calculate total or modified SVD scores. Neuropsychological assessments measured cognitive function. Spearman correlation analysis demonstrated that the total and modified SVD scores were associated with overall cognition as well as with function in the executive and visuospatial domains. The associations remained significant in linear regression after adjusting for age, sex, education and vascular risk factors. Binary logistic regression and chi-squared trend tests revealed that VCI risk increased significantly with SVD burden based on the modified SVD score. Subsequent chi-squared testing demonstrated that the VCI rate was significantly higher in patients with a modified SVD score of 5-6 than in patients without any SVD burden. Our results suggest that both the total and modified SVD scores show a negative association with cognitive function, but the modified SVD score may be better at identifying patients at high VCI risk.

Cerebral Small Vessel Disease

Cerebral small vessel disease (CSVD) represents a cluster of various vascular disorders with different pathological backgrounds. The advanced vasculature net of cerebral vessels, including small arteries, capillaries, arterioles and venules, is usually affected. Processes of oxidation underlie the pathology of CSVD, promoting the degenerative status of the epithelial layer. There are several classifications of cerebral small vessel diseases; some of them include diseases such as Binswanger’s disease, leukoaraiosis, cerebral microbleeds (CMBs) and lacunar strokes. This paper presents the characteristics of CSVD and the impact of the current knowledge of this topic on the diagnosis and treatment of patients.

Use of Multimodal Magnetic Resonance Imaging Techniques to Explore Cognitive Impairment in Leukoaraiosis

Leukoaraiosis, also referred to as white-matter hyperintensities (WMHs) or age-related white matter changes, is the most frequently seen lesion on brain magnetic resonance images (MRI) in the elderly. LA is a subject of intense research interest, and is correlated with stroke, cognitive impairment or dementia, disturbances, affective disorders, and poor prognoses. Rapid advances in neuroimaging have enabled greater understanding of LA associated with aging-related cognitive decline or dementia. Recently, the techniques of multimodal MRI, such as structural MRI (sMRI), resting-state functional MRI (rs-MRI), cerebrovascular reactivity (CVR), diffusion tensor imaging (DTI), magnetic resonance spectroscopy (MRS), and dynamic contrast-enhanced MRI (DCE-MRI), have been used to explore the underlying mechanism of cognitive impairment in patients with LA. These multimodal MRI techniques may provide further insights into the structural and functional changes of LA with cognitive dysfunction.

Disconnections of Cortico-Subcortical Pathways Related to Cognitive Impairment in Patients with Leukoaraiosis: A Preliminary Diffusion Tensor Imaging Study

BACKGROUND

We aimed to explore the relation between the microstructural integrity of white matter using the technique of diffusion tensor imaging (DTI) and changes of cognition in leukoaraiosis (LA).

METHODS

Fifty patients with LA and 50 age- and gender-matched controls were recruited consecutively. The average values of mean diffusivity (MD) and fractional anisotropy (FA) were quantified both within white matter lesions (WMLs) and normal-appearing white matter (NAWM) from the regions of interest (ROIs).

RESULTS

We found significantly decreased FA and increased MD in WMLs at the 5 ROIs than that in NAWM and controls (p < 0.05). The values of FA in NAWM were significantly lower at centrum semiovale and posterior periventricular white matter than those of controls (p < 0.05). The values of MD in NAWM were significantly higher at the anterior periventricular white matter and corpus callosum than those of controls (p < 0.05). The values of FA in NAWM located at anterior periventricular white matter correlated inversely with the Z scores of executive function (r = -0.420, p = 0.028).

CONCLUSIONS

DTI may provide some important information about the cognitive dysfunction in patients with LA, which may largely attribute to the "disconnection" of cortico-subcortical pathways, with the evidence of reduced FA and increased MD.

Predictors for cortical gray matter volume in stroke patients with confluent white matter changes

BACKGROUND AND PURPOSE

Our previous study found that cortical gray matter (cGM) volume predicted vascular cognitive impairment independent of age-related white matter changes (WMC). We aimed to investigate predictors for cGM volume in ischemic stroke patients with confluent WMC.

METHODS

One-hundred post-stroke patients with confluent WMC were recruited into the study. All volumetric measures were standardized by intracranial volume as volume ratio. Univariate analyses and multivariate linear regression models were used to test relationship of cGM volume with basic demography, vascular risk factors, APOE status, WMC volume (periventricular and deep WMC), infarct measures (volume, number and location) and microbleed (number, presence and location).

RESULTS

After controlling for significant variables in the univariate analyses, multivariate linear regression models found that old age (β=-0.288, p=0.001), low triglyceride (β=0.194, p=0.027), periventricular WMC (PVWMC) (β=-0.392, p<0.001) and presence of thalamic microbleed (β=-0.197, p=0.041) were independently predictive of less cGM volume ratio.

CONCLUSIONS

Age, PVWMC and left thalamic microbleed predict less cGM volume.

A quantitative systematic review of domain-specific cognitive impairment in lacunar stroke

OBJECTIVE

To quantitatively characterize domain-specific cognition in individuals with symptomatic lacunar stroke in a systematic review.

METHODS

Systematic searches of MEDLINE and EMBASE were conducted. Inclusion criteria were all articles published prior to December 2011 evaluating domain-specific cognitive status in individuals with a symptomatic lacunar infarct. Data extraction identified cognitive domains with reported impairment and effect size calculations and heterogeneity analyses were completed to assess the magnitude of this impairment for all studies with control group data.

RESULTS

Results of the search yielded 12 cross-sectional and 5 longitudinal studies that met inclusion criteria. Effect size calculations revealed small to medium effect sizes (ES) estimations for impairment after stroke in the domains of executive function (ES -0.44, 95% confidence interval [CI] -0.83, -0.50), memory (ES -0.55, 95% CI -0.96, -0.13), language (ES -0.63, 95% CI -0.92, -0.33), attention (ES -0.37, 95% CI -0.67, -0.07), and visuospatial abilities (ES -0.61, 95% CI -1.03, 0.19), and large effect sizes for global cognition (ES -0.90, 95% CI -1.48, -0.31) and information processing speed (ES -0.93, 95% CI -1.63, -0.23). Heterogeneity analyses revealed that a subset of these domains were heterogeneous and identified moderating factors accounting for this heterogeneity.

CONCLUSIONS

Results of this systematic review are consistent with previous characterizations of cognitive impairment associated with lacunar strokes. However, impaired cognition in this stroke subtype appears less selective than previously thought, involving all major cognitive domains.

Atrophy of the left dorsolateral prefrontal cortex is associated with poor performance in verbal fluency in elderly poststroke women

This study aimed to investigate the association between atrophy in the prefrontal cortex with executive function and verbal fluency in elderly male and female patients poststroke. Thirty elderly female patients with non-aphasic ischemic stroke aged ≥ 60 years and 30 age-matched non-aphasic male patients with ischemic stroke were recruited. Automatic magnetic resonance imaging segmentation was used to assess the volume of the whole prefrontal cortex, along with its subdivisions: anterior cingulate cortex, orbitofrontal cortex and dorsolateral prefrontal cortex. The Semantic Verbal Fluency Test was administered at 3 and 15 months poststroke. At 3 months poststroke, left dorsolateral prefrontal cortex volume was significantly correlated with Verbal Fluency Test score in female patients only (partial coefficient = 0.453, P = 0.045), after controlling for age, education, diabetes, neurological deficit, white matter lesions volume, as well as the location and volume of infarcts. At 15 months poststroke, there remained a significant association between the left dorsolateral prefrontal cortex volume and Verbal Fluency Test (partial coefficient = 0.661, P = 0.001) and between the left prefrontal cortex volume and Verbal Fluency Test (partial coefficient = 0.573, P = 0.004) in female patients after the same adjustments. These findings indicate that atrophy of the left dorsolateral prefrontal cortex contributes to the impairment of verbal fluency in elderly female patients with stroke. Sex differences may be present in the neuropsychological mechanisms of verbal fluency impairment in patients with stroke.

Neurochemical correlates of cognitive dysfunction in patients with leukoaraiosis: a proton magnetic resonance spectroscopy study

OBJECTIVES

Leukoaraiosis (LA) is a common radiological finding in the elderly and may reflect cerebral small vessel disease (SVD). Although SVD has been identified as a major cause of vascular cognitive impairment or vascular dementia, the mechanisms for this association remain unclear. We therefore aimed to measure brain metabolites in LA using proton magnetic resonance spectroscopy ((1)H-MRS) as to determine the relationship between cognitive function and neurochemical white matter profile.

METHODS

We recruited 23 patients with LA and 23 age- and sex-matched healthy controls consecutively. Multi-voxel (1)H-MRS was performed with a volume of interest located in centrum semiovale that contained mainly white matter voxels. Three main ratios of N-acetyl aspartate (NAA)/Cr, choline (Cho)/Cr and NAA/Cho were obtained. Spearman rank correlation coefficients were calculated between the cognitive function and the measured metabolite ratios.

RESULTS

We found significantly lower levels of NAA/Cho and NAA/Cr ratios in lesioned white matter in patients with LA than healthy controls (P<0.05). The ratios of NAA/Cho and NAA/Cr in normal appearing white matter (NAWM) were higher than lesioned white matter and lower than controls, but this difference was not significant (P>0.05). There was a positive relationship between Mini-Mental State Examination (MMSE) and NAA/Cho in NAWM (r = 0.417, P = 0.048), and also a positive relationship between MMSE and NAA/Cr in lesioned white matter (r = 0.551, P = 0.006) in patients with LA. A positive relationship between the Z scores of the executive function and NAA/Cho in lesioned white matter (r = 0.557, P = 0.006) was also found.

CONCLUSION

The main finding of this study was a significant reduction in the ratios of NAA/Cr and NAA/Cho in lesioned white matter, which indicates a marker of neuronal loss or dysfunction in patients with LA, which was correlated with cognitive function. This relationship between cognitive function and metabolic changes suggests that (1)H-MRS can be explored as a marker for cognitive dysfunction in patients with LA.

White matter lesions and brain gray matter volume in cognitively normal elders

Cerebral white matter lesions (WMLs) reflect small vessel disease, are common in elderly individuals, and are associated with cognitive impairment. We sought to determine the relationships between WMLs, age, gray matter (GM) volume, and cognition in the Cardiovascular Health Study (CHS). From the Cardiovascular Health Study we selected 740 cognitively normal controls with a 1.5 T magnetic resonance imaging (MRI) scan of the brain and a detailed diagnostic evaluation. WML severity was determined using a standardized visual rating system. GM volumes were analyzed using voxel-based morphometry implemented in the Statistical Parametric Mapping software. WMLs were inversely correlated with GM volume, with the greatest volume loss in the frontal cortex. Age-related atrophy was observed in the hippocampus and posterior cingulate cortex. Regression analyses revealed links among age, APOE*4 allele, hypertension, WMLs, GM volume, and digit symbol substitution test scores. Both advancing age and hypertension predict higher WML load, which is itself associated with GM atrophy. Longitudinal data are needed to confirm the temporal sequence of events leading to a decline in cognitive function.

Absence of Cerebral Microbleeds Predicts Reversion of Vascular ‘Cognitive Impairment No Dementia’ in Stroke

Background

Cerebral microbleeds may contribute to cognitive deficits in stroke. Cognitive impairment that does not meet the criteria for dementia (cognitive impairment no dementia) is common in stroke, and patients with such impairment can revert to normal cognition.

Aims and hypothesis

This study examined the association between cerebral microbleeds and the reversion of cognitive impairment no dementia.

Method

A total of 328 Chinese patients with acute ischemic stroke admitted to the acute stroke unit of a university-affiliated regional hospital in Hong Kong participated in the study. All subjects were assessed for cognitive impairment no dementia with a neuropsychological test battery at three- and 15 months following the index stroke. Of the 180 patients with cognitive impairment no dementia at three-months poststroke, 143 (79·4%) attended the 15-month follow-up. Twenty-nine subjects had reverted from cognitive impairment no dementia to normal cognitive status (reverters), 98 were nonreverters and 16 had progressed to dementia.

Results

In univariate analysis, the reverters were found to be younger, less likely to have hypertension and cerebral microbleeds, and to have smaller white matter hyperintensity volumes. In multivariate analysis, the absence of cerebral microbleeds remained an independent predictor of reversion with an odds ratio of 4·3. Absence of deep cerebral microbleeds predicted the reversion of the language domain, whereas the absence of lobar cerebral microbleeds predicted the reversion of the visuomotor speed domain.

Conclusions

The results suggest that the absence of cerebral microbleeds may be associated with a higher likelihood of a reversible cognitive impairment in stroke patients. The mechanism of how this occurs is not well understood.