Segmentation Toolbox Research Articles

Relationship between perivascular spaces and white matter hyperintensities in Alzheimer’s disease

AbstractBackgroundThere is increasing evidence that cerebrovascular and glymphatic‐related deficits may precede and influence AD pathology. In this study, we investigated the relationship between enlarged perivascular spaces (EPVS) as a measure of glymphatic clearance integrity and white matter hyperintensities (WMH) as a measure of cerebrovascular integrity.MethodParticipants from ADNI‐3 (N = 1026) who were diagnosed as cognitively unimpaired (CU) (n = 567), mild cognitively impaired (MCI) (n = 345), or AD dementia (n = 114) were included (Figure 2). Up to five visits (approximately one visit per year) were recorded for each participant. EPVS were segmented and quantified on 3T T1‐weighted images using a novel fully‐automated algorithm, while WMH were segmented on FLAIR images using the Lesion Segmentation Toolbox (Figure 1). Pairwise comparisons between EPVS and WMH metrics were tested using Mann‐Whitney U‐tests. A GLM was used to model the relationship between EPVS and WMH in each diagnostic group. LMMs were used to determine whether changes in WMH influenced EPVS changes between the groups. Age and sex were included as potential confounding factors, and data were transformed as needed.ResultIn the cross‐sectional analysis, the AD group had more EPVS count, volume, and WMH volume than the CU and MCI groups (Figure 2). A significant association was found between EPVS metrics and WMH volume (R‐squared = 0.4; p<0.001; Figure 3). MCI compared to CU presented a stronger association between EPVS count and WMH volume (p = 0.02). In the longitudinal analysis, EPVS count and volume significantly increased with time, however, no differences were found between groups (Figure 2). Changes in EPVS count were associated with changes in WMH volume, however, this association did not significantly differ between groups (Figure 2).ConclusionWe found that the association between measures of EPVS and WMH varies with different clinical stages of AD. This suggests that cerebrovascular and glymphatic‐related deficits may interact and exacerbate in all clinical stages with varying degrees and timing. Untangling the relationship between these pathophysiological changes and their influence on AD progression will further our understanding of the etiology of AD and highlight potential modifiable risk factors.

VesselBoost: A Python Toolbox for Small Blood Vessel Segmentation in Human Magnetic Resonance Angiography Data

Magnetic resonance angiography (MRA) performed at ultra-high magnetic field provides a unique opportunity to study the arteries of the living human brain at the mesoscopic level. From this, we can gain new insights into the brain’s blood supply and vascular disease affecting small vessels. However, for quantitative characterization and precise representation of human angioarchitecture to, for example, inform blood-flow simulations, detailed segmentations of the smallest vessels are required. Given the success of deep learning-based methods in many segmentation tasks, we explore their application to high-resolution MRA data and address the difficulty of obtaining large data sets of correctly and comprehensively labelled data. We introduce VesselBoost, a vessel segmentation toolbox, which utilizes deep learning and imperfect training labels for accurate vasculature segmentation. To enhance the segmentation models’ robustness and accuracy, VesselBoost employs an innovative data augmentation technique, which captures the resemblance of vascular structures across scales by zooming in or out on input image patches—virtually creating diverse scale vascular data. This approach enables detailed vascular segmentation and ensures the model’s ability to generalize across various scales of vascular structures.

NBEST: Deep-learning-based non-human primates Brain Extraction and Segmentation Toolbox across ages, sites and species

Accurate processing and analysis of non-human primate (NHP) brain magnetic resonance imaging (MRI) serves an indispensable role in understanding brain evolution, development, aging, and diseases. Despite the accumulation of diverse NHP brain MRI datasets at various developmental stages and from various imaging sites/scanners, existing computational tools designed for human MRI typically perform poor on NHP data, due to huge differences in brain sizes, morphologies, and imaging appearances across species, sites, and ages, highlighting the imperative for NHP-specialized MRI processing tools. To address this issue, in this paper, we present a robust, generic, and fully automated computational pipeline, called non-human primates Brain Extraction and Segmentation Toolbox (nBEST), whose main functionality includes brain extraction, non-cerebrum removal, and tissue segmentation. Building on cutting-edge deep learning techniques by employing lifelong learning to flexibly integrate data from diverse NHP populations and innovatively constructing 3D U-NeXt architecture, nBEST can well handle structural NHP brain MR images from multi-species, multi-site, and multi-developmental-stage (from neonates to the elderly). We extensively validated nBEST based on, to our knowledge, the largest assemblage dataset in NHP brain studies, encompassing 1,469 scans with 11 species (e.g., rhesus macaques, cynomolgus macaques, chimpanzees, marmosets, squirrel monkeys, etc.) from 23 independent datasets. Compared to alternative tools, nBEST outperforms in precision, applicability, robustness, comprehensiveness, and generalizability, greatly benefiting downstream longitudinal, cross-sectional, and cross-species quantitative analyses. We have made nBEST an open-source toolbox (https://github.com/TaoZhong11/nBEST) and we are committed to its continual refinement through lifelong learning with incoming data to greatly contribute to the research field.

Early detection of white matter hyperintensities using SHIVA-WMH detector.

White matter hyperintensities (WMHs) are well-established markers of cerebral small vessel disease, and are associated with an increased risk of stroke, dementia, and mortality. Although their prevalence increases with age, small and punctate WMHs have been reported with surprisingly high frequency even in young, neurologically asymptomatic adults. However, most automated methods to segment WMH published to date are not optimized for detecting small and sparse WMH. Here we present the SHIVA-WMH tool, a deep-learning (DL)-based automatic WMH segmentation tool that has been trained with manual segmentations of WMH in a wide range of WMH severity. We show that it is able to detect WMH with high efficiency in subjects with only small punctate WMH as well as in subjects with large WMHs (i.e., with confluency) in evaluation datasets from three distinct databases: magnetic resonance imaging-Share consisting of young university students, MICCAI 2017 WMH challenge dataset consisting of older patients from memory clinics, and UK Biobank with community-dwelling middle-aged and older adults. Across these three cohorts with a wide-ranging WMH load, our tool achieved voxel-level and individual lesion cluster-level Dice scores of 0.66 and 0.71, respectively, which were higher than for three reference tools tested: the lesion prediction algorithm implemented in the lesion segmentation toolbox (LPA: Schmidt), PGS tool, a DL-based algorithm and the current winner of the MICCAI 2017 WMH challenge (Park et al.), and HyperMapper tool (Mojiri Forooshani et al.), another DL-based method with high reported performance in subjects with mild WMH burden. Our tool is publicly and openly available to the research community to facilitate investigations of WMH across a wide range of severity in other cohorts, and to contribute to our understanding of the emergence and progression of WMH.

Cholinergic white matter hyperintensity volume, cognitive decline, and incident dementia in older adults

AbstractBackgroundWhite matter hyperintensities (WMHs) are frequently observed in the elderly and their severity tends to increase with age. Previous studies have revealed that WMHs are closely related to cognitive dysfunction. However, as inter‐individual variations and a clinicoradiological discrepancy exist, not all individuals with severe WMH burden experience worse cognitive performance. Furthermore, the location of WMHs appears to be a critical factor in the relationship between WMHs and cognitive function. Few studies take both volume and location of white matter hyperintensities (WMHs) into account to explore the association between WMH burden and cognitive impairment. We therefore investigated associations of cholinergic WMH volume (WMHV) with global cognitive function, cognitive decline, and incident dementia in older adults.MethodA total of 752 older adults (mean age 60 years, 50% female) from the Taizhou Imaging Study were involved. Brain magnetic resonance imaging data at baseline and repeated measures of cognition over 5 years were collected and analyzed. WMHV in whole brain, the cholinergic pathways, and different tract classes in the Montreal Neurologic Institute standard space was obtained using the Lesion Segmentation Toolbox and tools from FMRIB Software Library. Multivariable linear regression, Cox regression, and partial correlation tests were performed.ResultOver a median 5.67 (IQR: 2.88‐6.84) years of follow‐up, cholinergic WMHV was associated with annual decline of Mini‐Mental State Examination (MMSE) (β = ‐0.206; P = 0.006), and incident dementia (HR = 3.45; 95%CI: 2.08‐5.74). Within the cholinergic pathways, WMHs in commissural fibers were related to incident dementia. However, greater global WMHV only slightly increased the risk of incident dementia (HR = 1.03; 95%CI: 1.01‐1.06). Neither global nor cholinergic WMHV was associated with MMSE in cross‐sectional analysis.ConclusionCholinergic WMHV is associated with longitudinal cognitive decline and incident dementia in older adults. Within the cholinergic pathways, commissural fiber tracts were strategic locations correlated to incident dementia. Our findings suggest cholinergic WMHV might be a potential indicator of cognitive deterioration.

Cognitively healthy Indians with higher Framingham Cardiovascular Risk Scores have higher WMH burden with lesser grey matter, white matter and hippocampal volumes

AbstractBackgroundIncreasing evidence indicates the important role of cardiovascular risk factors in dementia risk, particularly in Indians. It is known that cardiovascular risk factors tend to cluster together, and a widely used multivariable risk scoring is the Framingham Cardiovascular Risk Score (FRS). However, very little is known about the association between FRS and structural brain MRI parameters in cognitively healthy Indians.MethodBaseline clinical and brain MRI data of 487 cognitively healthy Indians (Clinical Dementia Rating score of ‘0’), who are part of two ongoing, large, aging cohort studies in India – Srinivaspura Aging, Neuro Senescence and COGnition study (SANSCOG) study (n = 200) and Tata Longitudinal Study of Aging (TLSA, n = 289) were utilized. FRS was calculated using clinical and blood biochemistry data, whereas brain MRI scans were performed on 3T MRI scanners. White matter hyperintensity (WMH) lesion segmentation was performed using Lesion Segmentation Toolbox (LST) and brain volumes (total grey matter, total white matter, 3rd ventricle, right and left hippocampus) were derived using FreeSurfer software and normalized according to total intracranial volumes of the respective subjects. Linear regression analysis was performed with FRS as independent variable and various structural MRI parameters as dependent variables. A p value of <0.05 was considered significant.ResultUrban subjects with higher FRS had significantly higher WMH volume (β = 0.00011, p<0.001, R2 = 0.130), lower grey matter (β = ‐0.00285, p<0.001, R2 = 0.206), white matter (β = ‐0.00180, p<0.001, R2 = 0.106), right hippocampal (β = ‐0.00004, p<0.001, R2 = 0.180), left hippocampal (β = ‐0.00003, p<0.001, R2 = 0.156) volumes and higher 3rd ventricle volumes (β = 0.00003, p<0.001, R2 = 0.150). Rural subjects higher FRS had lower grey matter (β = ‐0.00166, p<0.001, R2 = 0.118, white matter (β = ‐0.0011, p<0.001, R2 = 0.060), right hippocampal (β = ‐0.00002, p<0.001, R2 = 0.087), left hippocampal (β = ‐0.00002, p<0.001, R2 = 0.084) volumes and higher 3rd ventricle volumes (β = 0.00004, p<0.001, R2 = 0.264) but no significant association was seen with WMH volume.ConclusionThese findings indicate that cardiovascular risk is associated with structural brain changes in cognitively healthy Indians. Since cardiovascular risk can be reduced through cost‐effective community level strategies, such as lifestyle‐based interventions, promptly addressing this risk in Indians could be an important part of dementia prevention strategies.

Comparison of white matter hyperintensities between Latin American and American individuals with Alzheimer’s disease and cognitively normal older adults

AbstractBackgroundWhite matter hyperintensities (WMH) are Flair signal abnormalities frequently seen in periventricular and subcortical brain regions in the elderly. They are presumed to be, at least in part, a sign of cardiovascular disease (CVD). WMHs are associated with stroke, cognitive decline, and dementia. Given that most of the literature on dementia comes from high‐income countries, little is known about dementia in Latin American countries (LAC), which have an increased risk of dementia and CVD due to unique demographic, socioeconomic, cultural and ethnic factors. Our study aims to characterize the effects of white matter hyperintensities (WMH) in patients with AD within LAC and to compare these findings with a sample from the USA.MethodA total of 491 latin american subjects (235 cognitively normal (CN), 256 AD) and 271 american subjects (165 CN, 106 AD), underwent a brain MRI. WMH were extracted from T2‐FLAIR images using the Lesion Segmentation Toolbox and were divided into periventricular (pWMH) or deep (dWMH) depending on their distance from the ventricles. Linear mixed‐effects were used to model the volumes of WMH as a function of age, gender, years of education, intracranial volume, average cortical thickness, and region (LAC or USA).ResultAD groups from both regions showed larger total, dWMH and pWMH volumes compared to HC. Higher pWMH from AD and HC was associated with age, female sex and inversely associated with years of education. On the other hand, the increase in dWMH was found to be associated with female sex for the HC group and with age for AD. Cortical thinning was associated with an increase in total, subcortical and periventricular WMH for both groups. No statistically significant association was found between the region (USA or LAC) and WMH.ConclusionAD showed higher WMH compared to HC. Our findings suggest that pWMH and dWMH have different clinical correlates. No associations between region and WMH load or distribution were found. Future studies need to elucidate if there is an increase in other signs of CVD (such as cerebral microbleeds, brain lacunae, enlarged perivascular spaces) in Latin America and if they are linked to cognition and social determinants of health.

Longitudinal white matter hyperintensities in Down’s syndrome

AbstractBackgroundDown’s syndrome (DS) is a genetically determined form of Alzheimer’s disease (AD). Typically not complicated by systemic vascular risk factors, the DS population represents a unique opportunity to unravel the complex relationship between AD and cerebrovascular disease. Here, we aim at studying the evolution over time of white matter hyperintensities (WMHs), a key hallmark of cerebrovascular disease frequently found in AD, and to assess their topographical progression and associations with AD biomarkers in adults with DS.MethodThis longitudinal study included 43 DS adults (age = 39.8±10.4y; females = 41.9%; n = 37 asymptomatic AD, n = 6 symptomatic AD) and 81 euploid controls (age = 53.7±6.2y; females = 67.3%) from the Down‐Alzheimer Barcelona Neuroimaging Initiative, with at least 2 MRI visits with T1 and FLAIR acquisitions (time delay = 3.4±1.6y). WMHs were segmented using the Lesion Prediction Algorithm implemented in the Lesion Segmentation Toolbox (SPM12). WMH volumes were extracted in 36 regions of interest and used to compute total and regional annual volume changes. Non‐parametric statistical tests were performed to assess the effect of disease severity on WMH volume changes over time and define associations between longitudinal WMHs and baseline demographic and genetic data, and fluid biomarkers.ResultTotal annual WMH volume change increased in symptomatic DS compared to both asymptomatic DS and controls (Figure 1A). Regionally, WMH load increased in regions showing WMH at baseline in all groups (Figure 2). These changes were greater and more distributed, involving layers more distant from the ventricles (especially in occipito‐parietal regions) in symptomatic DS. Age, but not sex, intellectual disability, or APOEε4 status, was associated with longitudinal WMH in DS (Figure 1B‐F): the annual WMH volume changes in DS differed from controls at age ∼45y. Finally, WMH volume changes were negatively related to CSF‐Aβ42/40 and positively related to plasma and CSF ptau‐181 and neurofilament light (Figure 3).ConclusionPathological increases in WMH volume start ∼10 years before the mean age at onset for AD (53.8y) in DS and relate to underlying AD pathology and neurodegeneration. Regional analyses suggest that WMH volume changes increase with disease progression and progressively involve deeper white matter areas. Together, these results support that WMHs are tightly related to AD pathogenesis.

Baseline cognitive dispersion is related to increases in white matter hyperintensities in cognitively preserved older individuals

AbstractBackgroundCognitive dispersion (CD), understood as intra‐individual variability in performance, is considered a sensitive marker of prodromal stages of Alzheimer’s Disease (AD; Halliday et al. J. Intell 2018;6(1):12), and it is thought to complement cognitive performance (CP) in predicting age‐related changes (Costa et al. ClinicNeuropsychol 2019;33(2)). Moreover, white matter hyperintensities (WMH) are a common age‐associated finding that especially impacts processing speed (Boutzoukas et al. Front in AgingNeurosci 2021;13) and executive function (Papp et al. NeuropsycholDev 2014;21(2)). Therefore, we aimed to investigate (i) associations between CD, CP, and WMH, and (ii) whether WMH age‐related changes can be better predicted by baseline CD or CP in a cognitively normal aging sample.MethodOne hundred and three healthy elders (age: 68.61±3.01 years; 72 females) underwent a neuropsychological assessment and a magnetic resonance image acquisition and repeated them after 2 years. Based on processing speed and executive function scores, we computed a baseline CD (i.e., CD‐tp1; using the intraindividual standard deviation method, Costa et al. Clinic Neuropsychol 2019;33(2)) and a CP (calculated with principal component analysis) at both time‐points (i.e., CP‐tp1, CP‐tp2). 3D‐FLAIR sequence was used to estimate WMH volumes (i.e., WMH‐tp1, WMH‐tp2) using the SPM12 Lesion Segmentation Toolbox (Schmidt et al. Neuroimage 2012;59(4)). The baseline volumes were subtracted from the follow‐up ones to obtain the difference (i.e., CP‐diff, WMH‐diff). Statistical analyses were linear regressions adjusted by age, sex, years of education, and intracranial volume for WMH analyses. Paired t‐tests were used for the longitudinal approach.ResultThere was a negative cross‐sectional relationship between CD‐tp1 and CP‐tp1 (B = ‐.240, t = ‐3.034, p = .004; Fig1), while those cognitive measures were unrelated to WMH‐tp1 volumes. Longitudinally, we detected a significant increase in WMH (t = 7.244, p<.001), while CP remained stable and there were no significant associations between CP‐diff and WMH‐diff. The main finding was the significant relationship found between WMH volume increases (i.e., WMH‐diff) and CD‐tp1 (B = ‐.320, t = ‐3.060, p = .003; Fig2), but not with CP‐tp1.ConclusionThe results show that in healthy elderly individuals who evidenced cognitive stability in processing speed and executive function, the CD is a sensitive marker to predict significant age‐related increases in WMH, which would be beneficial in the AD‐continuum.

Clinical Manifestations.

The Hachinski ischemic score (HIS) scale is used to exclude as much as possible, individuals with dominant cerebrovascular pathology from those with dominant AD pathology. White matter hyperintensities (WMH) are the gold-standard for evaluating cerebrovascular pathology and linked to worsening cognitive impairment. HIS ≤ 4 are used to minimize non-AD pathology in ADNI participants. Here we sought to understand whether HIS scores for hypertension (HMHYPERT) in ADNI3 were associated with greater WMH burden (and hence cerebrovascular pathology) even within individuals with HIS ≤ 4. We compared HMHYPERT scores between individuals with normal cognition (CN) and with mild cognitive impairment (MCI), and participants and correlated with percent WMH burden. A total of 263 subjects screened for ADNI3 (92 MCI [38 F], age = 73 ± 6; 146 CN [94 F], age = 74 ± 7) who had HIS scores and FLAIR MRI acquisitions within a year of each other. 25 All AD participants were excluded. WMH burden was calculated by segmenting them using the Lesion Segmentation Toolbox and normalizing their volume by total intracranial volume. We used the HMHYPERT as a vascular risk factor, excluding those with any focal neurological symptoms. For analysis, we dichotomized participants into those with HMHYPERT = 0 and HMHYPERT>0. To examine whether a non-zero HMHYPERT score corresponds to greater WMH burden, we examined the prevalence of WMH in between these 2 groups and within the MCI and CN groups. Although not significant, HMHYPERT scores were higher in MCI. (Fig. 1A, MCI = 0.48±0.50, CN = 0.39±0.49). We found a greater proportion of MCI subjects had HMHYPERT>0 than CN subjects (Fig. 1B, HMHYPERT = 0: CN = 61.0% and MCI = 52.2%; HMHYPERT > 0: CN = 40.0% and MCI = 48.1%). WMH burden was greater in MCI (0.33±0.53) than CN (0.20±0.32) (Fig. 2A). MCI participants with HMHYPERT>0 had the highest WMH burden followed by MCI participants with HMHYPERT = 0, CN with HMHYPERT >0, and CN with HMHYPERT = 0 had the lowest WMH burden. (Fig. 2B). Cerebrovascular pathology increases significantly with diagnosis, but its severity increases further with findings of hypertension even when total HIS ≤ 4.

The Influence of White Matter Lesion Volume on Hippocampal Shape in Patients Along the Alzheimer’s Disease Continuum

AbstractBackgroundAlzheimer’s Disease (AD) is a progressive neurocognitive disorder with irreversible hippocampal atrophy and degradation of cognitive function. Hippocampal atrophy can be assessed by volume and shape changes on MRI, and hippocampal volume is a sensitive marker for AD diagnosis and progression; however, it does not reflect regional changes within the hippocampus. Interest in shape analysis of the hippocampus is growing, as it provides more detailed assessment of morphological changes that highlight regional atrophic differences. Cerebral small vessel disease (CSVD) burden quantified as white matter lesion volume from MRI is strongly associated with hippocampal volume loss in AD and mild cognitive impairment (MCI). Our goal was to examine relationships between hippocampal morphometric features and white matter hyperintensity (WMH) volume, a measure of CSVD burden, across the AD continuum.MethodWe analyzed baseline data from 186 participants from the Wake Forest Alzheimer’s Disease Research Center (ADRC) Clinical Core cohort whose cognitive diagnosis was adjudicated by an expert panel in accordance with NIA‐AA guidelines (without reference to AD biomarkers). Participants ranged from cognitively normal (CN) (n = 62, 72±8y), to MCI (n = 62, 73±7y), and dementia (n = 62, 73±9y). We focused on the left hippocampus and used T1‐weighted images from 3T brain MRI to generate left hippocampal segmentations and total intracranial volume estimates. Total WMH volume was estimated from FLAIR images using the lesion segmentation toolbox for SPM, it was then normalized to intracranial volume and log transformed to address skewness. Hippocampal shape analysis was performed using SlicerSALT, and a multivariate linear model was implemented to relate WMH volume to hippocampal shape features while adjusting for diagnosis, age, sex, and years of education.ResultOn the superior aspect of the hippocampus, greater WMH volume was associated with more surface contraction of the anterior portions of the CA1, CA2, CA3 and subiculum as shown in Figure 1A. Similar effects were seen on the inferior aspect as shown in Figure 1B.ConclusionThese finding suggest that higher WMH volume is associated with selective hippocampal subfield degeneration in the anterior portions of the hippocampus which may be more vulnerable to CSVD buildup.

Developing Topics.

The novel diffusion-based Neurite Orientation and Dispersion Imaging (NODDI) model may provide complementary information to Diffusion Tensor Imaging (DTI) to increase sensitivity in characterizing tissue microarchitecture. We examined if one or a combination of both approaches can improve our understanding of at-risk tissue in and around white matter hyperintensities (WMH). From ADNI3, we identified individuals [n = 54, age = 78 ± 7, 46% F) with concurrent measures of multi-shell DTI, 3D FLAIR, and 3D MPRAGE. We segmented WMH from structural data using Lesion Segmentation Toolbox. We derived DTI (FA, MD) and NODDI parameters (NDI, ODI, Fiso) using QSIprep. We dilated up to 5 mm outside the WMH, applied an FA threshold of ≥ 0.2 to exclude gray matter, calculated DTI and NODDI parameters at 1 mm intervals within this region. We performed Pearson correlations between all parameters. Assuming normal white matter at 5 mm, we normalized these parameters with their respective values at 5 mm and plotted the spatial variation of each parameter in and outside the WMH. Normalized DTI and NODDI parameters representing spatial variations within the 5 mm region are shown in Fig 1A. Table 1 shows p-values comparing DTI and NODDI parameters between WMH the subsequent regions outside it. Compared to WMH, Fiso was significantly different at all distances. FA showed similar spatial characteristics with NDI and was negatively correlated with ODI. Fiso correlated very strongly with MD. In the WMHs and its immediate vicinity, low FA, NDI, ODI suggest lower cell density, while low Fiso and MD suggest low extracellular fluid volume. Together these observations may reflect cytotoxic edema, cell loss, and injury. Normal appearing white matter (3-5 mm) showed high FA, NDI, ODI, and low MD consistent with high cell density and normal extracellular water. In the transition between WMH vicinity and normal white matter, low FA, NDI, ODI, high MD and Fiso suggest cell loss only. This transition region could be at risk of converting to WMH. Future work will include longitudinal evaluation of WMH growth in the at-risk tissue. However, final validation of WMH pathology will need histological confirmation.

Maternal pre-pregnancy obesity affects the uncinate fasciculus white matter tract in preterm infants.

A growing body of evidence suggests an association between a higher maternal pre-pregnancy body mass index (BMI) and adverse long-term neurodevelopmental outcomes for their offspring. Despite recent attention to the effects of maternal obesity on fetal and neonatal brain development, changes in the brain microstructure of preterm infants born to mothers with pre-pregnancy obesity are still not well understood. This study aimed to detect the changes in the brain microstructure of obese mothers in pre-pregnancy and their offspring born as preterm infants using diffusion tensor imaging (DTI). A total of 32 preterm infants (born to 16 mothers with normal BMI and 16 mothers with a high BMI) at <32 weeks of gestation without brain injury underwent brain magnetic resonance imaging at term-equivalent age (TEA). The BMI of all pregnant women was measured within approximately 12 weeks before pregnancy or the first 2 weeks of gestation. We analyzed the brain volume using a morphologically adaptive neonatal tissue segmentation toolbox and calculated the major white matter (WM) tracts using probabilistic maps of the Johns Hopkins University neonatal atlas. We investigated the differences in brain volume and WM microstructure between preterm infants of mothers with normal and high BMI. The DTI parameters were compared among groups using analysis of covariance adjusted for postmenstrual age at scan and multiple comparisons. Preterm infants born to mothers with a high BMI showed significantly increased cortical gray matter volume (p = 0.001) and decreased WM volume (p = 0.003) after controlling for postmenstrual age and multiple comparisons. We found a significantly lower axial diffusivity in the uncinate fasciculus (UNC) in mothers with high BMI than that in mothers with normal BMI (1.690 ± 0.066 vs. 1.762 ± 0.101, respectively; p = 0.005). Our study is the first to demonstrate that maternal obesity impacts perinatal brain development patterns in preterm infants at TEA, even in the absence of apparent brain injury. These findings provide evidence for the detrimental effects of maternal obesity on brain developmental trajectories in offspring and suggest potential neurodevelopmental outcomes based on an altered UNC WM microstructure, which is known to be critical for language and social-emotional functions.

Cross-city matters: A multimodal remote sensing benchmark dataset for cross-city semantic segmentation using high-resolution domain adaptation networks

Artificial intelligence (AI) approaches nowadays have gained remarkable success in single-modality-dominated remote sensing (RS) applications, especially with an emphasis on individual urban environments (e.g., single cities or regions). Yet these AI models tend to meet the performance bottleneck in the case studies across cities or regions, due to the lack of diverse RS information and cutting-edge solutions with high generalization ability. To this end, we build a new set of multimodal remote sensing benchmark datasets (including hyperspectral, multispectral, SAR) for the study purpose of the cross-city semantic segmentation task (called C2Seg dataset), which consists of two cross-city scenes, i.e., Berlin-Augsburg (in Germany) and Beijing-Wuhan (in China). Beyond the single city, we propose a high-resolution domain adaptation network, HighDAN for short, to promote the AI model's generalization ability from the multi-city environments. HighDAN is capable of retaining the spatially topological structure of the studied urban scene well in a parallel high-to-low resolution fusion fashion but also closing the gap derived from enormous differences of RS image representations between different cities by means of adversarial learning. In addition, the Dice loss is considered in HighDAN to alleviate the class imbalance issue caused by factors across cities. Extensive experiments conducted on the C2Seg dataset show the superiority of our HighDAN in terms of segmentation performance and generalization ability, compared to state-of-the-art competitors. The C2Seg dataset and the semantic segmentation toolbox (involving the proposed HighDAN) will be available publicly at https://github.com/danfenghong/RSE_Cross-city.

Cocaine use is associated with cerebral white matter hyperintensities in HIV disease.

White matter hyperintensities (WMH), a marker of cerebral small vessel disease and predictor of cognitive decline, are observed at higher rates in persons with HIV (PWH). The use of cocaine, a potent central nervous system stimulant, is disproportionately common in PWH and may contribute to WMH. The sample included of 110 PWH on antiretroviral therapy. Fluid-attenuated inversion recovery (FLAIR) and T1-weighted anatomical MRI scans were collected, along with neuropsychological testing. FLAIR images were processed using the Lesion Segmentation Toolbox. A hierarchical regression model was run to investigate predictors of WMH burden [block 1: demographics; block 2: cerebrovascular disease (CVD) risk; block 3: lesion burden]. The sample was 20% female and 79% African American with a mean age of 45.37. All participants had persistent HIV viral suppression, and the median CD4+ T-cell count was 750. Nearly a third (29%) currently used cocaine regularly, with an average of 23.75 (SD = 20.95) days in the past 90. In the hierarchical linear regression model, cocaine use was a significant predictor of WMH burden (β = .28). WMH burden was significantly correlated with poorer cognitive function (r = -0.27). Finally, higher WMH burden was significantly associated with increased serum concentrations of interferon-γ-inducible protein 10 (IP-10) but lower concentrations of myeloperoxidase (MPO); however, these markers did not differ by COC status. WMH burden is associated with poorer cognitive performance in PWH. Cocaine use and CVD risk independently contribute to WMH, and addressing these conditions as part of HIV care may mitigate brain injury underlying neurocognitive impairment.

Longitudinal characterization of white matter degeneration in early stages of AD

AbstractBackgroundWhite matter (WM) degeneration in older adults is often characterized unimodally by diffusion tensor imaging (DTI), T1w and T2‐FLAIR MRI. However, approaches integrating multimodal MRI information will be necessary for better understanding WM degeneration in early stages of AD.MethodThis study characterizes the longitudinal changes of WM structure among 66 participants with marked baseline white matter hyperintensity volume (WMHv &gt; 2000 mm3). 42 cognitively normal [CN, 72±7.2y, F = 25] and 24 mild cognitive impairment [MCI, 74±5.7y, F = 14]) adults underwent longitudinal MRI (scan interval: 2.8±0.85 years) within the Wake Forest AD Research Center. DTI metrics were estimated using FMRIB’s Diffusion Toolbox, voxel‐wise volumetric changes of WM (Jacobian Determinant [JD]) were estimated from fractional anisotropy (FA) maps using SPM CAT12, and WMHv were quantified using SPM’s Lesion Segmentation Toolbox. We segmented WM into WMH and normal‐appearing WM (NAWM) to characterize different degeneration patterns by integrating multimodal MRI information. We performed a linear regression analysis of %WMHv change with cognitive status, scan interval, and baseline WMH volume as covariates.ResultDecreased FA and increased mean diffusivity (MD) were observed in WMH regions compared with NAWM. Within WMH, interestingly, volume and FA increased, while MD decreased (p &lt;0.001) during the longitudinal follow‐up (Figs. 1 and 2). Baseline WMHv was not significantly associated with cognitive status, but trended toward higher WMHv in MCI compared to CN (p = 0.07). Adjusted for baseline WMHv, cognitive status was associated with WMHv change (p = 0.036, MCI faster than CN, Fig. 3). While FA (p = 0.99) and MD (p = 0.16) did not change in the NAWM, volumetric change rate (JD) indicated volumetric contraction (p&lt;0.001, Figs. 4 and 5). WMH progresses with time in older adults at risk of AD dementia by occupying neighboring NAWM regions, resulting in relative increases in FA and decreases in MD.ConclusionIn NAWM, WM degeneration can be better characterized by volumetric contraction than white matter microstructural integrity. This observation could implicate that secondary or peripheral WM fibers may reveal more vulnerable tracts in early stages of AD. Further investigation incorporating imaging makers of blood perfusion and metabolism could better explain structural mechanisms of WM degeneration.

YKL-40 as a novel biomarker related to white matter damage and cognitive impairment in patients with cerebral small vessel disease.

YKL-40 is a novel neuroinflammatory marker associated with white matter damage and cognitive dysfunction. 110 CSVD patients, including 54 with mild cognitive impairment (CSVD-MCI), 56 with no cognitive impairment (CSVD-NCI), and 40 healthy controls (HCs) underwent multimodal magnetic resonance examination, serum YKL-40 level detection and cognitive function assessment to investigate the association between YKL-40 and white matter damage and cognitive impairment in cerebral small vessel disease (CSVD) patients. White matter hyperintensities volume was calculated using the Wisconsin White Matter Hyperintensity Segmentation Toolbox (W2MHS) for white matter macrostructural damage evaluation. For white matter microstructural damage evaluation, fractional anisotropy (FA) and mean diffusivity (MD) indices of the region of interest were analyzed based on diffusion tensor imaging (DTI) images using the Tract-Based Spatial Statistics (TBSS) pipeline. The serum YKL-40 level of CSVD was significantly higher than those of HCs, and the CSVD-MCI was higher than in HCs and CSVD-NCI. Furthermore, serum YKL-40 provided high diagnostic accuracy for CSVD and CSVD-MCI. The macroscopic and microstructure of white matter in CSVD-NCI and CSVD-MCI patients indicated different degrees of damage. Disruption of white matter macroscopic and microstructure was significantly associated with YKL-40 levels and cognition deficits. Moreover, the white matter damage mediated the associations between the increased serum YKL-40 levels and cognitive impairment. Our findings demonstrated that YKL-40 might be a potential biomarker of white matter damage in CSVD, whereas white matter damage was associated with cognitive impairment. Serum YKL-40 measurement provides complementary information regarding the neural mechanism of CSVD and its associated cognitive impairment.

Association of white matter lesions and brain atrophy with the development of dementia in a community: the Hisayama Study.

To investigate the association of white matter lesions volume (WMLV) levels with dementia risk and the association between dementia risk and the combined measures of WMLV and either total brain atrophy or dementia-related gray matter atrophy in a general older population. One thousand one hundred fifty-eight Japanese dementia-free community-residents aged ≥65 years who underwent brain magnetic resonance imaging were followed for 5.0 years. WMLV were segmented using the Lesion Segmentation Toolbox. Total brain volume (TBV) and regional gray matter volume were estimated by voxel-based morphometry. The WMLV-to-intracranial brain volume ratio (WMLV/ICV) was calculated, and its association with dementia risk was estimated using Cox proportional hazard models. Total brain atrophy, defined as the TBV-to-ICV ratio (TBV/ICV), and dementia-related regional brain atrophy defined based on our previous report were calculated. The association between dementia risk and the combined measures of WMLV/ICV and either total brain atrophy or the number of atrophied regions was also tested. During the follow-up, 113 participants developed dementia. The risks of dementia increased significantly with higher WMLV/ICV levels. In addition, dementia risk increased additively both in participants with higher WMLV/ICV levels and lower TBV/ICV levels and in those with higher WMLV/ICV levels and a higher number of dementia-related brain regional atrophy. The risk of dementia increased significantly with higher WMLV/ICV levels. An additive increment in dementia risk was observed with higher WMLV/ICV levels and lower TBV/ICV levels or a higher number of dementia-related brain regional atrophy, suggesting the importance of prevention or control of cardiovascular risk factors.

Performance evaluation of automated white matter hyperintensity segmentation algorithms in a multicenter cohort on cognitive impairment and dementia.

White matter hyperintensities (WMH), a biomarker of small vessel disease, are often found in Alzheimer's disease (AD) and their advanced detection and quantification can be beneficial for research and clinical applications. To investigate WMH in large-scale multicenter studies on cognitive impairment and AD, appropriate automated WMH segmentation algorithms are required. This study aimed to compare the performance of segmentation tools and provide information on their application in multicenter research. We used a pseudo-randomly selected dataset (n = 50) from the DZNE-multicenter observational Longitudinal Cognitive Impairment and Dementia Study (DELCODE) that included 3D fluid-attenuated inversion recovery (FLAIR) images from participants across the cognitive continuum. Performances of top-rated algorithms for automated WMH segmentation [Brain Intensity Abnormality Classification Algorithm (BIANCA), lesion segmentation toolbox (LST), lesion growth algorithm (LGA), LST lesion prediction algorithm (LPA), pgs, and sysu_media] were compared to manual reference segmentation (RS). Across tools, segmentation performance was moderate for global WMH volume and number of detected lesions. After retraining on a DELCODE subset, the deep learning algorithm sysu_media showed the highest performances with an average Dice's coefficient of 0.702 (±0.109 SD) for volume and a mean F1-score of 0.642 (±0.109 SD) for the number of lesions. The intra-class correlation was excellent for all algorithms (>0.9) but BIANCA (0.835). Performance improved with high WMH burden and varied across brain regions. To conclude, the deep learning algorithm, when retrained, performed well in the multicenter context. Nevertheless, the performance was close to traditional methods. We provide methodological recommendations for future studies using automated WMH segmentation to quantify and assess WMH along the continuum of cognitive impairment and AD dementia.

NeuroConstruct: 3D Reconstruction and Visualization of Neurites in Optical Microscopy Brain Images.

We introduce NeuroConstruct, a novel end-to-end application for the segmentation, registration, and visualization of brain volumes imaged using wide-field microscopy. NeuroConstruct offers a Segmentation Toolbox with various annotation helper functions that aid experts to effectively and precisely annotate micrometer resolution neurites. It also offers an automatic neurites segmentation using convolutional neuronal networks (CNN) trained by the Toolbox annotations and somas segmentation using thresholding. To visualize neurites in a given volume, NeuroConstruct offers a hybrid rendering by combining iso-surface rendering of high-confidence classified neurites, along with real-time rendering of raw volume using a 2D transfer function for voxel classification score versus voxel intensity value. For a complete reconstruction of the 3D neurites, we introduce a Registration Toolbox that provides automatic coarse-to-fine alignment of serially sectioned samples. The quantitative and qualitative analysis show that NeuroConstruct outperforms the state-of-the-art in all design aspects. NeuroConstruct was developed as a collaboration between computer scientists and neuroscientists, with an application to the study of cholinergic neurons, which are severely affected in Alzheimer's disease.