Voxel Size Research Articles

Microstructural alterations of major thalamic nuclei in the chronic pediatric spinal cord injured population

BackgroundThe brain undergoes reorganization following spinal cord injury (SCI), but little is known about how the thalamus is affected in pediatric SCIs. PurposeTo characterize microstructural alterations in the thalamus after SCI with diffusion tensor imaging (DTI) metrics. Methods18 pediatric participants with chronic SCI (8–20 years) were stratified using the American Spinal Injury Association Impairment Scale (AIS) into groups: A, B, and C/D. DTI of the brain used a 3 T Siemens Verio MRI using the parameters: 20 directions, number of averages = 3, b = 1000 s/mm2, voxel size = 1.8 mm × 1.8 mm, slice thickness = 5 mm, TE = 95 ms, TR = 4300 ms, 30 slices, FOV = 230 × 230 mm2, matrix = 128 × 128, acquisition time = 4:45 min. Diffusion data was processed to generate DTI metrics FA, MD, AD, and RD. Data analysisDTI metrics were acquired by superimposing the AAL3 thalamic atlas onto participant diffusion images registered to MNI152 space. We utilized a multiple Mann–Whitney U-test to compare between AIS groups, considering values of p ≤ 0.05 as significant. ResultsFA, AD, RD, and MD significantly differed in thalamic nuclei between AIS groups A vs B and B vs C/D. Significant nuclei include the right ventral anterior, left intralaminar, bilateral lateral pulvinar, and right lateral geniculate. ConclusionOur findings suggest the presence of microstructural alterations based on SCI severity in pediatric patients. These results are encouraging and warrant further study.

Attention-guided deep neural network with a multichannel architecture for lung nodule classification

Detecting and accurately identifying malignant lung nodules in chest CT scans in a timely manner is crucial for effective lung cancer treatment. This study introduces a deep learning model featuring a multi-channel attention mechanism, specifically designed for the precise diagnosis of malignant lung nodules. To start, we standardized the voxel size of CT images and generated three RGB images of varying scales for each lung nodule, viewed from three different angles. Subsequently, we applied three attention submodels to extract class-specific characteristics from these RGB images. Finally, the nodule features were consolidated in the model's final layer to make the ultimate predictions. Through the utilization of an attention mechanism, we could dynamically pinpoint the exact location of lung nodules in the images without the need for prior segmentation. This proposed approach enhances the accuracy and efficiency of lung nodule classification. We evaluated and tested our model using a dataset of 1018 CT scans sourced from the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI). The experimental results demonstrate that our model achieved a lung nodule classification accuracy of 90.11 %, with an area under the receiver operator curve (AUC) score of 95.66 %. Impressively, our method achieved this high level of performance while utilizing only 29.09 % of the time needed by the mainstream model.

Statistical distribution of micro and macro pores in acrylic bone cement- effect of amount of antibiotic content

Aseptic loosening due to mechanical failure of bone cement is considered to be a leading cause of revision of joint replacement systems. Detailed quantified information on the number, size and distribution pattern of pores can help to obtain a deeper understanding of the bone cement's fatigue behavior. The objective of this study was to provide statistical descriptions for the pore distribution characteristics of laboratory bone cement specimens with different amounts of antibiotic contents. For four groups of bone cement (Palacos) specimens, containing 0.3, 0.6, 1.2 and 2.4 wt/wt% of telavancin antibiotic, seven samples per group were micro computed tomography scanned (38.97 μm voxel size). The images were first preprocessed in Mimics and then analyzed in Dragonfly, with the level of threshold being set such that single-pixel pores become visible. The normalized pore volume data of the specimens were then used to extract the logarithmic histograms of the pore densities for antibiotic groups, as well as their three-parameter Weibull probability density functions. Statistical comparison of the pore distribution data of the antibiotic groups using the Mann–Whitney non-parametric test revealed a significantly larger porosity (p < 0.05) in groups with larger added antibiotic contents (2.4 and 0.6 wt/wt% vs 0.3 wt/wt%). Further analysis revealed that this effect was associated with the significantly larger frequency of micropores of 0.1–0.5 mm diameter (p < 0.05) in groups with larger antibiotic content (2.4 wt/wt% vs and 0.6 and 0.3 wt/wt%), implying that the elution of the added antibiotic produces micropores in this diameter range mainly. Based on this observation and the fatigue test results in the literature, it was suggested that micropore clusters have a detrimental effect on the mechanical properties of bone cement and play a major role in initiating fatigue cracks in highly antibiotic added specimens.

Predicted effects of image acquisition and analysis conditions on DTMRI tractography-based muscle architecture estimates.

To quantify the effects of the intrinsic signal pattern, image acquisition conditions, and data analysis conditions on diffusion-tensor MRI (DTMRI) tractography-based muscle architecture estimates using a sampling-reconstruction assessment framework. Numerical models of muscles were constructed with realistic architectural properties. DTMRI signals were computed at signal-to-noise ratio (SNR) of 24-96 and common voxel sizes. Fiber tracking was performed, and the results were compared with the known architectural properties. SNR exerted the most significant impact on the outcome. The outcome variables approached asymptotes at SNR ≈ 54. Large in-plane voxel dimensions reduced the similarity between reconstructed fibers and the known architectural properties. Higher order polynomials helped reconstruct fibers with more complicated geometry but overfit noise for less complex geometries. The intrinsic fiber curvature also affected the robustness of polynomial smoothing to SNR. Other conditions, such as the fiber dimensionality, voxel aspect ratio, and slice thickness, did not affect the outcomes. SNR ≥ 54 is recommended for accurate muscle architecture characterization using DTMRI. Averaged across all simulated conditions, the greatest percent errors under SNR = 54 were -5.6% and -4.0% for the pennation angle and fiber-tract length estimates, respectively. For fiber tracts with intermediate intrinsic curvature, the greatest percent error for the curvature estimate was 9.8% for SNR = 54. Smaller in-plane voxel size (≤1.5 mm) is preferred to minimize the estimation error in architectural properties. If necessary, slice thickness may be adjusted within typical ranges to achieve sufficient SNR when slices are aligned near the fiber direction. Third-order polynomial fitting is appropriate for smoothing fiber tracts.

Usefulness of R2 maps to detect GM changes in AD continuum: a pilot study

AbstractBackgroundMRI multi‐echo sequences can be used to compute quantitative R2 maps (1/T2). This pilot study aimed to determine whether R2 maps might offer relevant information on GM changes in AD continuum.MethodThe sample consisted of 60 subjects: 20 healthy controls (HC), 20 cognitively unimpaired individuals with CSF Aβ42/40&lt;0.071 (preAD) and 20 subjects with clinical diagnosis of Alzheimer disease (AD) validated by biomarkers (Ab42/p‐Tau181 ratio &lt;10.25; Lumipulse‐Fujirebio). The MRI protocol included 3D T1W (voxel size: 1.2×1.2×1.2mm3, TR = 9.8ms, TE = 3ms, TI = 900ms) and multi‐echo (GRASE, voxel size: 2×2.2×2 mm3, TR = 2000ms, 32 echoes, TE: successive multiples of 9.3ms). R2 maps were obtained through linear regression of the logarithm of multi‐echo images. GM and R2 maps were spatially normalized to the MNI and smoothed (8mm FWHM). Voxel‐wise one‐way 3‐group ANOVA and post‐hoc paired t‐tests were performed on modulated GM (GMv) and R2 maps using age and sex as confounders and applying a common GM mask (SPM12, statistical threshold: p&lt;0.001 uncorrected; k&gt;100 voxels (∼0.33cm3) in all analyses).ResultTable 1 displays the descriptive statistics of the sample. Figure 1 shows the results from post‐hoc two‐sample t‐tests between HC and AD for GMv and R2 maps. GMv decrements were mostly located in temporal and parietal regions, as expected. Differences in R2 maps covered those regions but also others such anterior cingulate. No differences were found between HC and preAD neither in R2 maps nor in GMv.ConclusionResults might indicate that changes in R2 are not only reflecting atrophy but other undetermined tissular changes. Further research is warranted to further determine the mechanisms underlying R2 maps changes and the potential usefulness of R2 maps in AD characterization and diagnosis.

BrainAGE methods: influence of field strength, voxel size, race, and ethnicity

AbstractBackgroundThe brain‐age algorithm utilizes a regression model of the typical aging trajectory, trained on structural MRIs and chronological ages of a large population of healthy individuals spanning a wide age range (Cole et al. 2017, Neuroimage). MRIs from new populations are then compared to this regression for estimating brain age, a biomarker of brain health. Few studies have investigated stability of the brain‐age model across field strength and the influence of voxel size and race is understudied.MethodWe analyzed stability of the brain‐age model within and across field strength in the ADNI cohort (n = 75: 25 CN/25 MCI/25 AD) using intra‐class correlation coefficients (ICC) and multiple R2 values. Impact of voxel size (0.8mm, 1mm, and ADNI2 sequence 1.1×1.1×1.2mm) was compared in a sample from the Indiana ADRC (n = 48: 19 CN/16 SCD/10 MCI/3 AD). We then calculated brain age gap (BAG; estimated– chronological age) and compared results across self‐reported racial and ethnic groups matched on diagnosis, sex, age, education, and handedness using independent t‐tests. BAG results of African Americans (AAs) from the IADRC (n = 71: 30 CN/19 SCD/15 MCI/7 AD) and ADNI (n = 147: 41 CN/20 EMCI/27 LMCI/15 AD/44 SMC), and Hispanics (n = 76: 24 CN/16 EMCI/11 LMCI/11 AD/14 SMC) and Asians (n = 47: 12 CN/7 EMCI/4 LMCI/9 AD/15 SMC) from ADNI only, were compared to matched non‐Hispanic Whites (NHWs).ResultThe brain‐age method was stable across repeated scans within field strength, but less stable across field strengths (Fig.1). 3T scans consistently estimated individuals to be ∼4.43 years younger than the 1.5T scans. The method was relatively stable within and across voxel sizes (Fig.2). There was a significant difference in BAG between races (NHWs and AAs, and NHWs and Asians), but not between ethnicities (NHWs and Hispanics) (Fig.3).ConclusionDue to the model’s variability across field strength, we propose a linear offset equation for data harmonization (Fig.1): ((3T estimate/0.92)+9.55). We are also investigating ComBat harmonization to eliminate effects of field strength. While other studies show structural differences across race, it will be important to determine whether the self‐reported racial differences in BAG are due to inherent model bias or true biological differences.

BrainAGE methods: influence of field strength, voxel size, race, and ethnicity

AbstractBackgroundThe brain‐age algorithm utilizes a regression model of the typical aging trajectory, trained on structural MRIs and chronological ages of a large population of healthy individuals spanning a wide age range (Cole et al. 2017, Neuroimage). MRIs from new populations are then compared to this regression for estimating brain age, a biomarker of brain health. Few studies have investigated stability of the brain‐age model across field strength and the influence of voxel size and race is understudied.MethodWe analyzed stability of the brain‐age model within and across field strength in the ADNI cohort (n = 75: 25 CN/25 MCI/25 AD) using intra‐class correlation coefficients (ICC) and multiple R2 values. Impact of voxel size (0.8mm, 1mm, and ADNI2 sequence 1.1×1.1×1.2mm) was compared in a sample from the Indiana ADRC (n = 48: 19 CN/16 SCD/10 MCI/3 AD). We then calculated brain age gap (BAG; estimated– chronological age) and compared results across self‐reported racial and ethnic groups matched on diagnosis, sex, age, education, and handedness using independent t‐tests. BAG results of African Americans (AAs) from the IADRC (n = 71: 30 CN/19 SCD/15 MCI/7 AD) and ADNI (n = 147: 41 CN/20 EMCI/27 LMCI/15 AD/44 SMC), and Hispanics (n = 76: 24 CN/16 EMCI/11 LMCI/11 AD/14 SMC) and Asians (n = 47: 12 CN/7 EMCI/4 LMCI/9 AD/15 SMC) from ADNI only, were compared to matched non‐Hispanic Whites (NHWs).ResultThe brain‐age method was stable across repeated scans within field strength, but less stable across field strengths (Fig.1). 3T scans consistently estimated individuals to be ∼4.43 years younger than the 1.5T scans. The method was relatively stable within and across voxel sizes (Fig.2). There was a significant difference in BAG between races (NHWs and AAs, and NHWs and Asians), but not between ethnicities (NHWs and Hispanics) (Fig.3).ConclusionDue to the model’s variability across field strength, we propose a linear offset equation for data harmonization (Fig.1): ((3T estimate/0.92)+9.55). We are also investigating ComBat harmonization to eliminate effects of field strength. While other studies show structural differences across race, it will be important to determine whether the self‐reported racial differences in BAG are due to inherent model bias or true biological differences.

Age‐ and Sex‐Related Morphological Changes in Cerebral Blood Vessels: a 7T TOF MRA Study

AbstractBackgroundIn this study we evaluated age‐ and sex‐associated morphological changes of cerebral blood vessels using 7T Time‐of‐flight (TOF) magnetic resonance angiography (MRA).MethodOur study consisted of 25 cognitively normal older adults (20 female, mean age 68.2 years). Whole‐Brain TOF MRA images were collected on a 7T Siemens scanner at the University of Pittsburgh (slice number = 354, voxel size = 0.38*0.38*0.38 mm, 12 mins). An iterative vessel segmentation algorithm, VesselMapper, was used to segment blood vessels on TOF MRA images (Li et al., 2023). For each subject, vessels are further divided into large and small vessels by median split based on vessel diameters. Mean tortuosity of large vessels, small vessels and overall vessels was computed. Linear regression analyses were performed to evaluate the main effects of age and sex, and the age*sex interaction on vessel tortuosity and diameter.ResultAs shown in Figure 1, we found age‐related increase in mean vessel tortuosity (greater age is associated with greater tortuosity, p&lt;0.02). This age‐related difference is driven by small vessels (p = 0.002) rather than large vessels (p = 0.7). We also found that age is positively associated with mean voxel tortuosity (p&lt;0.001, black line in Figure 2). Furthermore, women showed greater age‐related increase in mean arterial tortuosity, compared to men (age*sex interaction p = 0.04, Figure 2).ConclusionOur findings validated that morphological markers of cerebral blood vessels are sensitive to aging. Small vessels in the brain are particularly vulnerable. Further, compared to men, older women are at a higher risk for cerebral small vessel disease, which may contribute to elevated SVD‐related AD risk in women.Limitation: The study has an unbalanced sample (more women than men) and a relatively modest sample size (N = 25). A cross‐sectional instead of longitudinal design was used in this analysis. Future study with a large balanced sample and a longitudinal design are required to confirm our findings.Reference: Li, Jiatai, et al. “VesselMapper ‐ A Robust Vessel Segmentation Algorithm for 3D Images.” Annual Meeting of the Organization for Human Brain Mapping. 2023

Age‐ and Sex‐Related Morphological Changes in Cerebral Blood Vessels: a 7T TOF MRA Study

AbstractBackgroundIn this study we evaluated age‐ and sex‐associated morphological changes of cerebral blood vessels using 7T Time‐of‐flight (TOF) magnetic resonance angiography (MRA).MethodOur study consisted of 25 cognitively normal older adults (20 female, mean age 68.2 years). Whole‐Brain TOF MRA images were collected on a 7T Siemens scanner at the University of Pittsburgh (slice number = 354, voxel size = 0.38*0.38*0.38 mm, 12 mins). An iterative vessel segmentation algorithm, VesselMapper, was used to segment blood vessels on TOF MRA images (Li et al., 2023). For each subject, vessels are further divided into large and small vessels by median split based on vessel diameters. Mean tortuosity of large vessels, small vessels and overall vessels was computed. Linear regression analyses were performed to evaluate the main effects of age and sex, and the age*sex interaction on vessel tortuosity and diameter.ResultAs shown in Figure 1, we found age‐related increase in mean vessel tortuosity (greater age is associated with greater tortuosity, p&lt;0.02). This age‐related difference is driven by small vessels (p = 0.002) rather than large vessels (p = 0.7). We also found that age is positively associated with mean voxel tortuosity (p&lt;0.001, black line in Figure 2). Furthermore, women showed greater age‐related increase in mean arterial tortuosity, compared to men (age*sex interaction p = 0.04, Figure 2).ConclusionOur findings validated that morphological markers of cerebral blood vessels are sensitive to aging. Small vessels in the brain are particularly vulnerable. Further, compared to men, older women are at a higher risk for cerebral small vessel disease, which may contribute to elevated SVD‐related AD risk in women. Limitation: The study has an unbalanced sample (more women than men) and a relatively modest sample size (N = 25). A cross‐sectional instead of longitudinal design was used in this analysis. Future study with a large balanced sample and a longitudinal design are required to confirm our findings. Reference: Li, Jiatai, et al. “VesselMapper ‐ A Robust Vessel Segmentation Algorithm for 3D Images.” Annual Meeting of the Organization for Human Brain Mapping. 2023

Hippocampal subfield volumes and memory performances: associations with stimulus modality and free recall or recognition

AbstractBackgroundAge‐related memory decline is well associated with hippocampal atrophy, although the specific role of hippocampal subfields in memory function has not been thoroughly investigated yet. In this study, we investigated the associations between hippocampal subfield volumes and free recall and recognition performances in verbal and visual memory tasks in older adults without dementia.MethodWe selected 97 right‐handed participants without dementia aged 60 to 85, 42 males, from the Translational Biomarkers in Aging and Dementia (TRIAD) cohort. From T2‐weighted MRIs with 0.7*0.7*1mm voxel size, the hippocampi were segmented five subfields: 1) DG (DG); 2) CA2 with CA3 (CA2/CA3); 3) CA1; 4) strata radiatum, lacunosum and moleculare (SRLM); and 5) subiculum by using the MAGeT‐Brain algorithm. Memory was assessed with Rey Auditory Learning test for verbal free recall (RAVLT7) and recognition (RAVLT‐R), and with Aggie Figure Learning test for visual free recall (AFLT7) and recognition (AFLT‐R). Linear models were used to test the association between hippocampal subfield volumes and memory performances with age, sex and total intracranial volume as covariates (p&lt;0.05, corrected for false discovery rate).ResultThe volumes of bilateral CA1 and SRLM were significantly associated with RAVLT7 and RAVLT‐R. The right DG volume was significantly associated with AFLT7 and RAVLT7. AFLT7 was also associated with the volume of right CA2/CA3 while AFLT‐R was not associated with any hippocampal subfield (see Figures 1 &amp; 2).ConclusionOur results are consistent with the view that hippocampal subfields contribute differently to memory traces: CA1 for recollection that requires contextual information; the DG for pattern separation (creating new distinct representations of each stimulus); and CA3 for pattern completion (the recollection of a stimulus from a partial memory). Contextual information is more important when recollecting known words, while distinct and efficient encoding may be especially useful for search strategies inherent to free recall. Pattern completion could be particularly useful to efficiently recollect visual abstract figures based on visual features during free recall. Overall, this shows that hippocampal subfield segmentation offers a better understanding of their distinctive roles in cognition.

Validation of high resolution subcortical MRI signal using voxel to voxel histology to neuroimaging correlation in AD continuum: advancing neuroimaging biomarker tools for early AD staging

AbstractBackgroundStudies suggest a decrease in LC volume by 8% each Braak stage, starting at Braak 0‐1, while in normal aging the LC volume remains intact. Thus, LC volumetry is an attractive potential biomarker to monitor AD progression from precognitive stages. However, precise LC volumetry neuroimaging is challenging due to the LC’s small size and location in the brainstem that’s prone to artifacts. To validate the precision/utility of neuroimaging sequences designed for LC volumetry, we developed a pipeline allowing voxel‐to‐voxel histology to neuroimaging comparisons. We tested this pipeline to analyze the neuro‐histological counterpart of an LC volume obtained from manual segmentation from a MRI neuromelanin(NM)sensitive GRE image.MethodWhole brain high resolution(0.7mm)3D‐T1 images and 2D‐2mm thick NM‐sensitive axial GRE images of the brainstem were obtained postmortem in situ before brain removal from a single subject using a 7T MRI(Siemens Magnetom). We combined celloidin‐based histological processing and 3D reconstructions of the histology volume at microscopic resolution which was then aligned to T1. We used structures visible in structural MRI(T1)(brainstem surface, red nucleus, and 4th ventricle) to calculate Dice coefficient to measure registration accuracy. Next, we performed an overlap percentage analysis to investigate how well GRE sequence detected LC borders compared to histology.ResultWe ran Dice coefficient analysis for brainstem structures to measure histology to MRI registration accuracy(Table1). The LC GRE to histology masks overlap percentage was 35.29% and 10.77% compared to the MRI and histology mask total voxels(voxel size = 0.34mm3). The overlap percentage obtained for the rostral and middle LC thirds ranged from 9.43% to 75%when compared to total voxels in MRI/histology masks, while the caudal third had zero overlap(Table2,Fig.1).ConclusionOur pipeline enables voxel‐to‐voxel correlation analysis between histology and high resolution LC MRIs for neuroimaging validation. The caudal third of the LC in our case has shown a clear lack of overlap potentially due to lack of signal specificity in MRI when compared to the gold standard that is histology. Current analysis of another five other cases is ongoing. Validation of LC MRI signal will inform interpretations of signal results in living patients with the potential to act as a diagnostic tool for ADRD.

Visualizing Braak stages with deformation‐based morphometry in super‐sampled MRI

AbstractBackgroundUnlike MRI data, histological data collected post‐mortem can’t be used to study the longitudinal formation of neurofibrillary tangles (NFT) and corresponding brain changes over time for individuals. Although structural MRI offers the possibility of non‐invasive longitudinal antemortem imaging, its resolution is a limiting factor. Using image processing and super‐sampling techniques, we examined the possibility of employing structural MRI to observe the NFT footprints in the brain.MethodData included baseline MRI scans of 1473 subjects from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset including ADNI1, 2, and GO. Voxel size was super‐sampled to 0.5×0.5×0.5mm (Manjón et al. 2010) before non‐linear registration to an ADNI‐based unbiased symmetric MRI template. The resulting deformation fields were used to compute the Jacobian determinant map for each subject as a proxy for local volume. For each subject, the sex‐ and age‐corrected z‐scored Jacobian maps were then calculated and averaged for each diagnostic group: normal subjects(CN), early MCI, late MCI, and AD. The average map for CN group was subtracted from the other average maps to show the AD‐stage differences.ResultThe Braak tau staging pattern (Braak &amp; Braak, 1995) is seen in average Jacobian maps as the disease progresses. Subjects in the MCI groups showed on average more pronounced neurodegeneration in the transentorhinal and entorhinal region (Braak Stages 1‐2), followed by degeneration in the hippocampus (Stages 3‐4). Later on, this neurodegeneration progresses toward the neocortex (Stages 5‐6). Figure 1 illustrates these results: A) the template, B)EMCI vs CN, C) LMCI vs EMCI, D) LMCI vs CN, E) AD vs CN. Red color shows an increase and blue indicates a decrease in local volume.ConclusionUsing novel image processing techniques, we were able to verify the staging of AD‐related neurofibrillary pathology progression in MRI data. This shows that structural MRI data, although limited in resolution, can still offer in‐depth insight into the brain antemortem which will in return allow the longitudinal and possible lifespan studies with an accuracy close to histological data.

Resting‐state network connectivity changes in prodromal stage of Alzheimer’s disease

AbstractBackgroundAlzheimer Disease (AD), the most common type of dementia, is accompanied by a generalized loss of neurons and cause severe changes in brain’s structure and function. Mild Cognitive Impairment is associated with a higher risk of developing dementia. This clinical classification is particularly relevant because each subtype is linked to a presumed etiology, with the amnestic subtypes (aMCI) considered as a prodromal form of AD (Petersen et al., 2009, 2014). Resting state fMRI functional connectivity (FCrs‐fMRI) analysis is broadly used to detect brain network differences in neurodegenerative diseases such as AD.MethodForty‐five participants were evaluated: patients with aMCI (n = 17), patients with dementia due to AD (n = 14) and healthy controls (HC) (n = 14). All participants received an extensive neuropsychological battery and brain MRI scanning. Resting state functional scans consisted of 240 T2*‐weighted echo planar imaging volumes (repetition time 3000 ms; echo time 60 ms; ip angle 90°; 32 axial slices; matrix 64×64; voxel size 3.3×3.4×4 mm). Additionally, a high‐resolution T1‐weighted magnetization prepared rapid acquisition gradient echo image was acquired. Data analyses were performed using the Functional Connectivity Toolbox (CONN) for Matlab (https://www.nitrc.org/frs/) to identify large‐scale patterns of temporal signal‐intensity coherence, interpreted as functional connectivity (Beckmann et al., 2005). Regions of interests (ROIs) analysis was performed for each participants and among group. Spatial maps of the group independent component analysis were used in a linear model t against each individual fMRI data set.ResultThe Default Mode Network (DMN) showed two regions of lower functional connectivity (FC) in AD when comparing with HC within the precuneus (p&lt;0.001 uncorrected). No regions of FC changes were found in the DMN when comparing aMCI with HC or AD.ConclusionThe changes in FC rsfMRI in adults with AD are specifically related to regions of the DMN.

Quantification of the Brain Iron Metabolism in the Cognitively Declined Adults Using 7T MRI

AbstractBackgroundThe extracellular aggregation of beta‐amyloid (AB) plaques is one of the hallmark pathology of Alzheimer’s diseases (AD) that develops years prior to clinical symptoms. However, inconsistency in the rate of cognitive decline among subjects with AB pathology and ineffectiveness of anti‐amyloid therapeutics to mitigate cognitive impairment suggest that other parameters, such as patterns of brain iron, may impact the risk of cognitive decline. In this study we leveraged improved MRI signal at 7T to explore susceptibility changes in the iron‐enriched regions among individuals with AD/MCI and healthy controls (HCs) through quantitative susceptibility mapping (QSM) with unprecedented resolution.MethodIn vivo MRI experiment was performed in AD/MCI and age‐matched healthy subjects. Scanning was conducted on a 7T MRI scanner using 1Tx/32Rx Nova head coil. To calculate QSM, a 3D high‐resolution spoiled gradient echo (GRE) sequence was used (voxels size = 0.3×0.3×1.5 mm3). QSM was used to measure susceptibility in the major iron‐enriched structures (Fig. 1), including red nucleus (RN), substantia nigra (SN), globus pallidus (GP), Caudate nucleolus (CN), putamen (PT), and dentate nucleus (DN).ResultWe found a significantly higher magnetic susceptibility in GP and DN in patients with AD/MCI compared with HCs (Fig. 2). Overall mean susceptibility values for the RN, SN, CN, GP, PT, and DN in AD/MCI and HCs is summarized in Table 1. In AD/MCI the highest susceptibility values were found in the GP (233.5±12.7) and DN (214.7±10.4), verses 174.6±11.8 and 168.9±11.6 for GP and DN in the HCs, respectively. There were small differences in the susceptibility values between the other ROIs.ConclusionWe reported an initial result for QSM at 7T in the AD/MCI subjects and appropriate age matched HCs. The main aim of the study was to investigate whether susceptibility values measured using ultra‐high resolution QSM differ in the iron‐enriched regions between these two populations. A significant susceptibility value was reported in the AD/MCI compared with the HC in GP and DN regions. This observation clearly deserves further investigations in larger cohorts.

Hippocampal subfield volumes and memory performances: associations with stimulus modality and free recall or recognition

AbstractBackgroundAge‐related memory decline is well associated with hippocampal atrophy, although the specific role of hippocampal subfields in memory function has not been thoroughly investigated yet. In this study, we investigated the associations between hippocampal subfield volumes and free recall and recognition performances in verbal and visual memory tasks in older adults without dementia.MethodWe selected 97 right‐handed participants without dementia aged 60 to 85, 42 males, from the Translational Biomarkers in Aging and Dementia (TRIAD) cohort. From T2‐weighted MRIs with 0.7*0.7*1mm voxel size, the hippocampi were segmented five subfields: 1) DG (DG); 2) CA2 with CA3 (CA2/CA3); 3) CA1; 4) strata radiatum, lacunosum and moleculare (SRLM); and 5) subiculum by using the MAGeT‐Brain algorithm. Memory was assessed with Rey Auditory Learning test for verbal free recall (RAVLT7) and recognition (RAVLT‐R), and with Aggie Figure Learning test for visual free recall (AFLT7) and recognition (AFLT‐R). Linear models were used to test the association between hippocampal subfield volumes and memory performances with age, sex and total intracranial volume as covariates (p&lt;0.05, corrected for false discovery rate)ResultThe volumes of bilateral CA1 and SRLM were significantly associated with RAVLT7 and RAVLT‐R. The right DG volume was significantly associated with AFLT7 and RAVLT7. AFLT7 was also associated with the volume of right CA2/CA3 while AFLT‐R was not associated with any hippocampal subfield (see Figures 1 &amp; 2).ConclusionOur results are consistent with the view that hippocampal subfields contribute differently to memory traces: CA1 for recollection that requires contextual information; the DG for pattern separation (creating new distinct representations of each stimulus); and CA3 for pattern completion (the recollection of a stimulus from a partial memory). Contextual information is more important when recollecting known words, while distinct and efficient encoding may be especially useful for search strategies inherent to free recall. Pattern completion could be particularly useful to efficiently recollect visual abstract figures based on visual features during free recall. Overall, this shows that hippocampal subfield segmentation offers a better understanding of their distinctive roles in cognition.

Cross vendor test‐retest validation of diffusion tensor analysis along the perivascular space (DTI‐ALPS) method for evaluating glymphatic system function in vascular cognitive impairment and dementia (VCID)

AbstractBackgroundGlymphatic system (GS) is a recently discovered brain‐wide perivascular fluid transport system in the cerebral nervous system (CNS)(Benveniste H et al., 2019) The impairment of glymphatic transport has been demonstrated to be associated with several neurological diseases, in particular the vascular cognitive impairment and dementia (VCID)(Tang J et al., 2022). The diffusion tensor analysis along the perivascular space (DTI‐ALPS) was proposed to non‐invasively evaluate the GS clearance dysfunction which has been used to investigate the GS in various pathologies(Taoka T et al., 2017, Hsu JL et al., 2022). However, studies on the cross‐vendor and test‐retest reliability of DTI‐ALPS method are lacking. The current study aimed to perform cross‐vendor, inter‐rater and test‐retest validations of DTI‐ALPS method by using a cohort from the MarkVCID consortium.MethodFifty participants’ DTI data from the MarkVCID consortium (consists of 7 sites) were included in this study: 15 participants were scanned on four MRI scanners; 35 participants underwent two MR scans within 14 days. The DTI data were acquired used a single shell, b = 1000s/mm2, 40‐direction with a voxel size of 2.02.02.0mm3 and six b = 0 s/mm2 on 3T MR scanners from seven sites, including two Siemens systems, one Philips system and one GE system. Two pipelines by using DSI studio and FSL software were developed for data processing and ALPS index calculation (Figure 1). The mean ALPS (mALPS) index was obtained by the average of bilateral ALPS index and was used for testing the inter‐rater, cross‐vendor and test‐retest reliability. The association between mALPS index and Montreal Cognitive assessment (MoCA) scores were evaluated using a general linear model with age and gender as covariate.ResultThe Bland‐Altman plot and scatterplot illustrated the validation results of the mALPS index analyzed using DSI studio and FSL pipelines (Figure 2). The mALPS index demonstrated favorable inter‐scanner reproducibility (ICC = 0.77 to 0.95, P&lt; 0.001), inter‐rater reliability (ICC = 0.96 to 1, P&lt; 0.001) and test‐retest repeatability (ICC = 0.89 to 0.95, P&lt; 0.001). A significantly positive correlation was observed between mALPS index and MoCA scores (r = 0.502, P&lt; 0.01) (Figure 3).ConclusionThe mALPS index offers a robust potential biomarker for evaluating GS clearance function in VCID.

Body Mass Index, but not Visceral Abdominal Obesity and Insulin Resistance, Negatively Impacts White Matter Microstructure in Midlife

AbstractBackgroundMidlife obesity, as evidenced by high body mass index (BMI), is increasingly understood as a risk factor for development of Alzheimer’s disease. Importantly, visceral fat, by producing adipokines, contributes to inflammation as well as insulin resistance, mechanisms involved in development of Alzheimer’s disease. Herein, we aimed to assess association of white matter microstructure, measured by Diffusion tensor imaging (DTI), with obesity and abdominal adipose tissue composition at midlife.MethodA total of 34 middle‐aged (age: 51.27 ± 6.12 years, BMI: 32.28 ± 6.39 kg/m2), cognitively normal participants, underwent bloodwork, brain and abdominal MRI. Insulin resistance status was assessed using Homeostatic Model Assessment for Insulin Resistance. Abdominal visceral and subcutaneous adipose tissue (VAT, SAT) were automatically segmented using VOXel Analysis Suite (Voxa) and VAT/SAT ratio was calculated. A diffusion spectrum imaging scheme with a total of 98 diffusion samplings and a voxel size of 2 mm was acquired. Using DSI studio (https://dsi‐studio.labsolver.org/), the diffusion data were reconstructed using generalized q‐sampling imaging and the tensor metrics were calculated. Fractional anisotropy (FA), mean, axial, and radial diffusivity (MD, AD, RD, respectively) for white matter tracts were compared across study groups using one‐way ANOVA. Multiple regression models with sex, age, and education as covariates were used to assess the association of BMI, VAT/SAT ratio, and HOMA‐IR with DTI metrics.ResultOur data showed that BMI was negatively associated with AD values in right inferior fronto‐occipital fasciculus, tapetum and forceps major of Corpus Callosum (CC), superior and middle cerebellar peduncle, and bilateral reticular tract. The negative association of BMI with AD values in tapetum of CC was significant only in women. Right reticular tract MD and superior cerebellar peduncle MD and RD values were also significantly associated with BMI. There were no significant associations between neither of diffusion metrics and VAT/SAT ratio or HOMA‐IR.ConclusionAssociation of midlife obesity with lower axonal diffusivity, reflective of lower white matter integrity, with differential patterns in men and women especially in CC, is in line with previous findings in healthy individuals. Reduced axonal diffusivity in reticular and brain stem pathways could be indicative of early dysregulations observed in Alzheimer’s disease.

Ex‐vivo microbleed detection in community‐based older adults using confidence‐aware learning

AbstractBackgroundCerebral microbleed (CMB) detection on MRI scans of autopsied brains of community‐based older adults is crucial for MR‐pathology studies of cerebral small vessel disease (SVD). However, training an ex‐vivo CMB detection model is difficult due to the low prevalence of CMBs in the brains of community‐based older adults, and the fact that CMB mimics greatly outnumber CMBs on ex‐vivo MRI. We demonstrate that confidence‐aware deep learning, a technique for enforcing correct ranking of examples during neural network training, significantly improves ex‐vivo CMB detection performance and enhances the interpretability of model predictions.MethodCMBs from 286 participants from two longitudinal cohort studies of aging were included in this work. T2*‐weighted gradient echo scans of autopsied brains with a voxel size of 1×1×1 mm were used. Confidence‐aware learning was used to regularize deep learning model predictions such that they reflect the degree of confidence the detection model has in a prediction. Model confidence was estimated during training by correctness, the number of correct prediction events across epochs divided by the total number of epochs. The objective loss function was modified to include a correctness ranking loss (CRL) term that encourages correspondence between correctness and predicted probabilities. Models are trained with and without the CRL loss term to evaluate the impact of confidence‐aware learning on model performance. An end‐to‐end CMB detection framework that combines data synthesis, candidate selection, false positive reduction, and full scan evaluation was used as the backbone for this work.ResultCMB detection models trained using correctness ranking loss achieved a significantly higher ensemble average precision (0.2915) compared with models trained without correctness ranking loss (0.2379). The CRL trained model achieved a 25% higher sensitivity at 0.5 false positives per subject and 12% higher sensitivity at 3 false positives per subject when compared to models trained without CRL.ConclusionThis work demonstrates that training with confidence‐aware learning can improve the performance and interpretability of ex‐vivo CMB detection algorithms in community‐based cohorts. The use of a confidence‐aware CMB detection algorithm for assisted CMB annotation in ex‐vivo MRI will aid in future MR‐pathology studies into the relationships between CMBs and neuropathology.

Compressed SENSE accelerated 3D single-breath-hold late gadolinium enhancement cardiovascular magnetic resonance with isotropic resolution: clinical evaluation

AimThe purpose of this study was to investigate the clinical application of Compressed SENSE accelerated single-breath-hold LGE with 3D isotropic resolution compared to conventional LGE imaging acquired in multiple breath-holds.Material &amp;amp; MethodsThis was a retrospective, single-center study including 105 examinations of 101 patients (48.2 ± 16.8 years, 47 females). All patients underwent conventional breath-hold and 3D single-breath-hold (0.96 × 0.96 × 1.1 mm3 reconstructed voxel size, Compressed SENSE factor 6.5) LGE sequences at 1.5 T in clinical routine for the evaluation of ischemic or non-ischemic cardiomyopathies. Two radiologists independently evaluated the left ventricle (LV) for the presence of hyperenhancing lesions in each sequence, including localization and transmural extent, while assessing their scar edge sharpness (SES). Confidence of LGE assessment, image quality (IQ), and artifacts were also rated. The impact of LV ejection fraction (LVEF), heart rate, body mass index (BMI), and gender as possible confounders on IQ, artifacts, and confidence of LGE assessment was evaluated employing ordinal logistic regression analysis.ResultsUsing 3D single-breath-hold LGE readers detected more hyperenhancing lesions compared to conventional breath-hold LGE (n = 246 vs. n = 216 of 1,785 analyzed segments, 13.8% vs. 12.1%; p &amp;lt; 0.0001), pronounced at subendocardial, midmyocardial, and subepicardial localizations and for 1%–50% of transmural extent. SES was rated superior in 3D single-breath-hold LGE (4.1 ± 0.8 vs. 3.3 ± 0.8; p &amp;lt; 0.001). 3D single-breath-hold LGE yielded more artifacts (3.8 ± 1.0 vs. 4.0 ± 3.8; p = 0.002) whereas IQ (4.1 ± 1.0 vs. 4.2 ± 0.9; p = 0.122) and confidence of LGE assessment (4.3 ± 0.9 vs. 4.3 ± 0.8; p = 0.374) were comparable between both techniques. Female gender negatively influenced artifacts in 3D single-breath-hold LGE (p = 0.0028) while increased heart rate led to decreased IQ in conventional breath-hold LGE (p = 0.0029).ConclusionsIn clinical routine, Compressed SENSE accelerated 3D single-breath-hold LGE yields image quality and confidence of LGE assessment comparable to conventional breath-hold LGE while providing improved delineation of smaller LGE lesions with superior scar edge sharpness. Given the fast acquisition of 3D single-breath-hold LGE, the technique holds potential to drastically reduce the examination time of CMR.

Control of brine composition over reactive transport processes in calcium carbonate rock dissolution: Time-lapse imaging of evolving dissolution patterns

This study investigates the impact of brine composition—specifically calcium ions and NaCl-based salinity—on the development of dissolution features in Ketton, a porous calcium carbonate rock. Utilizing a laboratory XMT (X-ray microtomography) scanner, we captured time-lapse in situ images of Ketton samples throughout various dissolution experiments, conducting four distinct flow-through experiments with differing brine solutions at a flow rate of 0.26 ml min⁻1. The scans yielded a voxel size of 6 μm, enabling the assessment of the temporal evolution of porosity and pore structure through image analysis and permeability evaluations via single-phase fluid flow simulations employing direct numerical solutions and network modeling, as opposed to direct measurement.Time-lapse imaging technique has delineated the extent to which the concentrations of CaCl₂ and NaCl in the injecting solution control the structural evolution of dissolution patterns, subsequently triggering the development of characteristic dissolution pattern. The inflow solution with no Ca2+ ions and with the minimal salt content manifested maximum dissolution near the sample inlet, coupled with the formation of numerous dissolution channels, i.e., wormholes. Conversely, solutions with a trace amount of Ca2⁺ ions induced focused dissolution, resulting in the formation of sparsely located channels. Inflow solutions with high concentrations of both Ca2⁺ ions and salt facilitated uniformly dispersed dissolution, primarily within microporous domains, initiating particle detachment and displacement and leading to localized pore-clogging. The relative increase in permeability, in each experiment, was correlated with the developed dissolution pattern. It was discerned that varying ratios of salt and calcium concentrations in the injected solution systematically influenced image-based permeability simulations and porosity, allowing for the depiction of an empirical porosity-permeability relationship.