Search

Data and code availability: Code used in calculating the tissue-weighting mean is available here: https://github.com/tdveale/NODDI-tissue-weighting-tool. ROI data and other scripts used in this analysis are available on request and without restriction by contacting the corresponding author. Acquired or processed NIfTI images are not available due to patient confidentiality agreements. Data used in preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf Appendix: Table A1 available at https://www.sciencedirect.com/science/article/pii/S1053811921010211?via%3Dihub#tbl0001 ; Appendix B. Supplementary materials available at https://ars.els-cdn.com/content/image/1-s2.0-S1053811921010211-mmc1.docx (Word document (3MB)). Copyright © 2021 The Authors. Neurite orientation dispersion and density imaging (NODDI) estimates microstructural properties of brain tissue relating to the organisation and processing capacity of neurites, which are essential elements for neuronal communication. Descriptive statistics of NODDI tissue metrics are commonly analyzed in regions-of-interest (ROI) to identify brain-phenotype associations. Here, the conventional method to calculate the ROI mean weights all voxels equally. However, this produces biased estimates in the presence of CSF partial volume. This study introduces the tissue-weighted mean, which calculates the mean NODDI metric across the tissue within an ROI, utilising the tissue fraction estimate from NODDI to reduce estimation bias. We demonstrate the proposed mean in a study of white matter abnormalities in young onset Alzheimer's disease (YOAD). Results show the conventional mean induces significant bias that correlates with CSF partial volume, primarily affecting periventricular regions and more so in YOAD subjects than in healthy controls. Due to the differential extent of bias between healthy controls and YOAD subjects, the conventional mean under- or over-estimated the effect size for group differences in many ROIs. This demonstrates the importance of using the correct estimation procedure when inferring group differences in studies where the extent of CSF partial volume differs between groups. These findings are robust across different acquisition and processing conditions. Bias persists in ROIs at higher image resolution, as demonstrated using data obtained from the third phase of the Alzheimer's disease neuroimaging initiative (ADNI); and when performing ROI analysis in template space. This suggests that conventional ROI means of NODDI metrics are biased estimates under most contemporary experimental conditions, the correction of which requires the proposed tissue-weighted mean. The tissue-weighted mean produces accurate estimates of ROI means and group differences when ROIs contain voxels with CSF partial volume. In addition to NODDI, the technique can be applied to other multi-compartment models that account for CSF partial volume, such as the free water elimination method. We expect the technique to help generate new insights into normal and abnormal variation in tissue microstructure of regions typically confounded by CSF partial volume, such as those in individuals with larger ventricles due to atrophy associated with neurodegenerative disease. CP and GZ were funded by the Wellcome Trust (Collaborative Award 200181/Z/15/Z). TV was funded by an Alzheimer's Research UK PhD scholarship (ARUK-PhD2018–009). MB was supported by a Fellowship award from the Alzheimer's Society, UK (AS-JF-19a-004–517). MB's work was also supported by the UK Dementia Research Institute which receives its funding from DRI Ltd, funded by the UK Medical Research Council, Alzheimer's Society and Alzheimer's Research UK. IM was supported by Alzheimer's Research UK (ARUK-PG2014–1946, ARUK-PG2017–1946) and the Wolfson Foundation (PR/ylr/18575). DLT was supported by the UCL Leonard Wolfson Experimental Neurology Centre (PR/ylr/18575), UCLH NIHR Biomedical Research Centre and the Wellcome Trust (Centre award 539208). JMS acknowledges the support of the National Institute for Health Research University College London Hospitals Biomedical Research Centre, Wolfson Foundation, Alzheimer's Research UK, Brain Research UK, Weston Brain Institute, Medical Research Council, British Heart Foundation, UK Dementia Research Institute and Alzheimer's Association. DMC was supported by the UK Dementia Research Institute which receives its funding from DRI Ltd, funded by the UK Medical Research Council, Alzheimer's Society and Alzheimer's Research UK, as well as Alzheimer's Research UK (ARUK‐PG2017‐1946) and the UCL/UCLH NIHR Biomedical Research Centre. We would also like to acknowledge Prof. Nick Fox who is a senior NIHR investigator for his role in conceiving the initial YOAD study preceding this work. The authors would like to thank all research participants who made this study possible, as well as Alzheimer's Research UK and Iceland Foods Charitable Foundation for funding the Young-Onset Alzheimer's disease study. The Dementia Research Centre is supported by Alzheimer's Research UK, Brain Research Trust, and The Wolfson Foundation. They also thank Kirsty Lu, Amelia Carton, Timothy Shakespeare, Keir Yong, Aida Suarez Gonzalez and Silvia Primativo for assistance with neuropsychology assessments.

Alzheimer’s disease is a neurodegenerative disease in which the degree of cortical atrophy in specific structures of the brain serves as a useful imaging biomarker. Recent approaches using linear mixed effects (LME) models in longitudinal neuroimaging have been powerful and flexible in investigating the temporal trajectories of cortical thickness. However, massive-univariate analysis, a simplified approach that obtains a summary statistic (e.g., a p -value) for every vertex along the cortex, is insufficient to model cortical atrophy because it does not account for spatial similarities of the signals in neighboring locations. In this article, we develop a permutation-based inference procedure to detect spatial clusters of vertices showing statistically significant differences in the rates of cortical atrophy. The proposed method, called SpLoc, uses spatial information to combine the signals adaptively across neighboring vertices, yielding high statistical power while controlling family-wise error rate (FWER) accurately. When we reject the global null hypothesis, we use a cluster selection algorithm to detect the spatial clusters of significant vertices. We validate our method using simulation studies and apply it to the Alzheimer’s Disease Neuroimaging Initiative (ADNI) data to show its superior performance over existing methods. An R package for implementing SpLoc is publicly available.

Prediction of subject age from brain anatomical MRI has the potential to provide a sensitive summary of brain changes, indicative of different neurodegenerative diseases. However, existing studies typically neglect the uncertainty of these predictions. In this work we take into account this uncertainty by applying methods of functional data analysis. We propose a penalised functional quantile regression model of age on brain structure with cognitively normal (CN) subjects in the Alzheimer’s Disease Neuroimaging Initiative (ADNI), and use it to predict brain age in Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD) subjects. Unlike the machine learning approaches available in the literature of brain age prediction, which provide only point predictions, the outcome of our model is a prediction interval for each subject., Highlights • The workflow proposed returns brain age prediction intervals from 3D images. • Quantile regression can be used to create distribution-free prediction intervals. • Functional regression is an efficient way to fit a global model for brain images. • The predictive accuracy is similar to more sophisticated approaches. • The gap between predicted and chronological age correlates with cognitive decline.

It is challenging to compare amyloid PET images obtained with different radiotracers. Here, we introduce a new approach to improve the interchangeability of amyloid PET acquired with different radiotracers through image-level translation. Deep generative networks were developed using unpaired PET datasets, consisting of 203 [11C]PIB and 850 [18F]florbetapir brain PET images. Using 15 paired PET datasets, the standardized uptake value ratio (SUVR) values obtained from pseudo-PIB or pseudo-florbetapir PET images translated using the generative networks was compared to those obtained from the original images. The generated amyloid PET images showed similar distribution patterns with original amyloid PET of different radiotracers. The SUVR obtained from the original [18F]florbetapir PET was lower than those obtained from the original [11C]PIB PET. The translated amyloid PET images reduced the difference in SUVR. The SUVR obtained from the pseudo-PIB PET images generated from [18F]florbetapir PET showed a good agreement with those of the original PIB PET (ICC = 0.87 for global SUVR). The SUVR obtained from the pseudo-florbetapir PET also showed a good agreement with those of the original [18F]florbetapir PET (ICC = 0.85 for global SUVR). The ICC values between the original and generated PET images were higher than those between original [11C]PIB and [18F]florbetapir images (ICC = 0.65 for global SUVR). Our approach provides the image-level translation of amyloid PET images obtained using different radiotracers. It may facilitate the clinical studies designed with variable amyloid PET images due to long-term clinical follow-up as well as multicenter trials by enabling the translation of different types of amyloid PET.

Longitudinal MRI studies are often subjected to mid-study scanner changes, which may alter image characteristics such as contrast, signal-to-noise ratio, contrast-to-noise ratio, intensity non-uniformity and geometric distortion. Measuring brain volume loss under these conditions can render the results potentially unreliable across the timepoint of the change. Estimating and accounting for this effect may improve the reliability of estimates of brain atrophy rates.We analyzed 237 subjects who were scanned at 1.5 T for the Alzheimer's Disease Neuroimaging Initiative (ADNI) study and were subject to intra-vendor or inter-vendor scanner changes during follow-up (up to 8 years). Sixty-three subjects scanned on GE Signa HDx and HDxt platforms were also subject to a T1-weighted sequence change from Magnetization Prepared Rapid Gradient Echo (MP-RAGE) to Fast Spoiled Gradient Echo with IR Preparation (IR-FSPGR), as part of the transition from ADNI-1 to ADNI-2/GO. Two-timepoint percentage brain volume changes (PBVCs) between the baseline "screening" and the follow-up scans were calculated using SIENA. A linear mixed-effects model with subject-specific random slopes and intercepts was applied to estimate the fixed effects of scanner hardware changes on the PBVC measures. The same model also included a term to estimate the fixed effects of the T1-weighted sequence change.Different hardware upgrade or change combinations led to different offsets in the PBVC (SE; p): Philips Intera to Siemens Avanto, -1.81% (0.30; p 0.0001); GE Genesis Signa to Philips Intera, 0.99% (0.47, p = 0.042); GE Signa Excite to Signa HDx, 0.33% (0.095, p = 0.0005); GE Signa Excite to Signa HDxt, -0.023% (0.23, p = 0.92); GE Signa Excite to Signa HDx to Signa HDxt, 0.25% (0.095, p = 0.010) and 0.27% (0.16, p = 0.098), respectively; GE Signa HDx to Signa HDxt, -0.24% (0.25, p = 0.34); Siemens Symphony to Symphony TIM, -0.39% (0.16; p = 0.019). The sequence change from MP-RAGE to IR-SPGR was associated with an average -1.63% (0.12; p 0.0001) change.Inter-vendor scanner changes generally led to greater effects on PBVC measurements than did intra-vendor scanner upgrades. The effect of T1-weighted sequence change was comparable to that of the inter-vendor scanner changes. Inclusion of the corrective fixed-effects terms for the scanner hardware and T1-weighted sequence changes yielded better model goodness-of-fits, and thus, potentially more reliable estimates of whole-brain atrophy rates.

According to the latest Braak staging of Alzheimer's disease (AD), tau pathology occurs earliest in the brain in the locus coeruleus (LC) of the brainstem, then propagates to the transentorhinal cortex (TEC), and later to other neocortical regions. Recent animal and in vivo human brain imaging research also support the trans-axonal propagation of tau pathology. In addition, neurochemical studies link norepinephrine to behavioral symptoms in AD. It is thus critical to examine the integrity of the LC-TEC pathway in studying the early development of the disease, but there has been limited work in this direction. By leveraging the high-resolution and multi-shell diffusion MRI data from the Human Connectome Project (HCP), in this work we develop a novel method for the reconstruction of the LC-TEC pathway in a cohort of 40 HCP subjects carefully selected based on rigorous quality control of the residual distortion artifacts in the brainstem. A probabilistic atlas of the LC-TEC pathway of both hemispheres is then developed in the MNI152 space and distributed publicly on the NITRC website. To apply our atlas on clinical imaging data, we develop an automated approach to calculate the medial core of the LC-TEC pathway for localized analysis of connectivity changes. In a cohort of 138 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI), we demonstrate the detection of the decreased fiber integrity in the LC-TEC pathways with increasing disease severity.

Automatic segmentation of brain anatomy has been a key processing step in quantitative neuroimaging analyses. An extensive body of literature has relied on Freesurfer segmentations. Yet, in recent years, the multi-atlas segmentation framework has consistently obtained results with superior accuracy in various evaluations. We compared brain anatomy segmentations from Freesurfer, which uses a single probabilistic atlas strategy, against segmentations from Multi-atlas region Segmentation utilizing Ensembles of registration algorithms and parameters and locally optimal atlas selection (MUSE), one of the leading ensemble-based methods that calculates a consensus segmentation through fusion of anatomical labels from multiple atlases and registrations. The focus of our evaluation was twofold. First, using manual ground-truth hippocampus segmentations, we found that Freesurfer segmentations showed a bias towards over-segmentation of larger hippocampi, and under-segmentation in older age. This bias was more pronounced in Freesurfer-v5.3, which has been used in multiple previous studies of aging, while the effect was mitigated in more recent Freesurfer-v6.0, albeit still present. Second, we evaluated inter-scanner segmentation stability using same day scan pairs from ADNI acquired on 1.5T and 3T scanners. We also found that MUSE obtains more consistent segmentations across scanners compared to Freesurfer, particularly in the deep structures.

Accurate and reliable whole-brain segmentation is critical to longitudinal neuroimaging studies. We undertake a comparative analysis of two subcortical segmentation methods, Automatic Segmentation (ASEG) and Sequence Adaptive Multimodal Segmentation (SAMSEG), recently provided in the open-source neuroimaging package FreeSurfer 7.1, with regard to reliability, bias, sensitivity to detect longitudinal change, and diagnostic sensitivity to Alzheimer’s disease. First, we assess intra- and inter-scanner reliability for eight bilateral subcortical structures: amygdala, caudate, hippocampus, lateral ventricles, nucleus accumbens, pallidum, putamen and thalamus. For intra-scanner analysis we use a large sample of participants (n = 1629) distributed across the lifespan (age range = 4–93 years) and acquired on a 1.5T Siemens Avanto (n = 774) and a 3T Siemens Skyra (n = 855) scanners. For inter-scanner analysis we use a sample of 24 participants scanned on the day with three models of Siemens scanners: 1.5T Avanto, 3T Skyra and 3T Prisma. Second, we test how each method detects volumetric age change using longitudinal follow up scans (n = 491 for Avanto and n = 245 for Skyra; interscan interval = 1–10 years). Finally, we test sensitivity to clinically relevant change. We compare annual rate of hippocampal atrophy in cognitively normal older adults (n = 20), patients with mild cognitive impairment (n = 20) and Alzheimer’s disease (n = 20). We find that both ASEG and SAMSEG are reliable and lead to the detection of within-person longitudinal change, although with notable differences between age-trajectories for most structures, including hippocampus and amygdala. In summary, SAMSEG yields significantly lower differences between repeated measures for intra- and inter-scanner analysis without compromising sensitivity to changes and demonstrating ability to detect clinically relevant longitudinal changes.

The volumetric and morphometric examination of hippocampus formation subfields in a longitudinal manner using in vivo MRI could lead to more sensitive biomarkers for neuropsychiatric disorders and diseases including Alzheimer’s disease, as the anatomical subregions are functionally specialised. Longitudinal processing allows for increased sensitivity due to reduced confounds of inter-subject variability and higher effect-sensitivity than cross-sectional designs. We examined the performance of a new longitudinal pipeline (Longitudinal Automatic Segmentation of Hippocampus Subfields [LASHiS]) against three freely available, published approaches. LASHiS automatically segments hippocampus formation subfields by propagating labels from cross-sectionally labelled time point scans using joint-label fusion to a non-linearly realigned ‘single subject template’, where image segmentation occurs free of bias to any individual time point. Our pipeline measures tissue characteristics available in in vivo high-resolution MRI scans, at both clinical (3 Tesla) and ultra-high field strength (7 Tesla) and differs from previous longitudinal segmentation pipelines in that it leverages multi-contrast information in the segmentation process. LASHiS produces robust and reliable automatic multi-contrast segmentations of hippocampus formation subfields, as measured by higher volume similarity coefficients and Dice coefficients for test-retest reliability and robust longitudinal Bayesian Linear Mixed Effects results at 7 T, while showing sound results at 3 T. All code for this project including the automatic pipeline is available at https://github.com/CAIsr/LASHiS

The association between white matter hyperintensities (WMH) and amyloid accumulation over time in cognitively normal, amyloid-negative elderly people remains largely unexplored. In order to study whether baseline WMH were associated with longitudinal subthreshold amyloid accumulation, 159 cognitively normal participants from the Alzheimer’s Disease Neuroimaging Initiative who were amyloid-negative at baseline were examined. All the participants underwent a T1 and a Fluid-Attenuated Inversion Recovery MRI scan at baseline. Amyloid PET imaging was performed at baseline and follow-up visits in 2-year intervals for up to 8 years. Partial volume correction was applied for quantifying cortical Standardised Uptake Value Ratios (SUVR). The associations between global and regional WMH burden and amyloid accumulation were assessed using linear mixed models adjusted by demographic characteristics and baseline SUVR. Partial volume correction increased the measured annual rate of change (+2.4%) compared to that obtained from non-corrected data (+0.5%). There were no significant correlations between baseline WMHs and baseline subthreshold cortical amyloid uptake. In a longitudinal analysis, increased baseline cortical SUVR and increased baseline burden of global (p ​= ​0.006), frontal (p ​= ​0.006), and parietal WMH (p ​= ​0.003) were associated with faster amyloid accumulation. WMH-related amyloid accumulation occurred in parietal, frontal, and, to a lesser extent, cingulate cortices. These results remained unchanged after a sensitivity analysis excluding participants with the highest cortical SUVRs. This is the first study to identify a specific spatial distribution of WMH which is associated with future amyloid accumulation in cognitively normal elderly subjects without PET-detectable amyloid pathology. These findings may have important implications in prevention trials for the early identification of amyloid accumulation.

Data-driven disease progression models have provided important insight into the timeline of brain changes inAD phenotypes. However, their utility in predicting the progression of pre-symptomatic AD in a populationbased setting has not yet been investigated. In this study, we investigated if the disease timelines constructedin a case-controlled setting, with subjects stratified according to APOE status, are generalizable to a populationbased cohort, and if progression along these disease timelines is predictive of AD. Seven volumetric biomarkersderived from structural MRI were considered. We estimated APOE-specific disease timelines of changes in thesebiomarkers using a recently proposed method called co-initialized discriminative event-based modeling (co-initDEBM). This method can also estimate a disease stage for new subjects by calculating their position along thedisease timelines. The model was trained and cross-validated on the Alzheimer’s Disease Neuroimaging Initiative(ADNI) dataset, and tested on the population-based Rotterdam Study (RS) cohort. We compared the diagnosticand prognostic value of the disease stage in the two cohorts. Furthermore, we investigated if the rate of change ofdisease stage in RS participants with longitudinal MRI data was predictive of AD. In ADNI, the estimated diseasetimeslines for 𝜖4 non-carriers and carriers were found to be significantly different from one another (𝑝 < 0.001).The estimate disease stage along the respective timelines distinguished AD subjects from controls with an AUCof 0.83 in both APOE 𝜖4 non-carriers and carriers. In the RS cohort, we obtained an AUC of 0.83 and 0.85 in 𝜖4non-carriers and carriers, respectively. Progression along the disease timelines as estimated by the rate of changeof disease stage showed a significant difference (𝑝 < 0.005) for subjects with pre-symptomatic AD as compared tothe general aging population in RS. It distinguished pre-symptomatic AD subjects with an AUC of 0.81 in APOE𝜖4 non-carriers and 0.88 in carriers, which was better than any individual volumetric biomarker, or its rate ofchange, could achieve. Our results suggest that co-init DEBM trained on case-controlled data is generalizable to apopulation-based cohort setting and that progression along the disease timelines is predictive of the developmentof AD in the general population. We expect that this approach can help to identify at-risk individuals from thegeneral population for targeted clinical trials as well as to provide biomarker based objective assessment in suchtrials.

Most existing algorithms for automatic 3D morphometry of human brain MRI scans are designed for data with near-isotropic voxels at approximately 1 mm resolution, and frequently have contrast constraints as well-typically requiring T1-weighted images (e.g., MP-RAGE scans). This limitation prevents the analysis of millions of MRI scans acquired with large inter-slice spacing in clinical settings every year. In turn, the inability to quantitatively analyze these scans hinders the adoption of quantitative neuro imaging in healthcare, and also precludes research studies that could attain huge sample sizes and hence greatly improve our understanding of the human brain. Recent advances in convolutional neural networks (CNNs) are producing outstanding results in super-resolution and contrast synthesis of MRI. However, these approaches are very sensitive to the specific combination of contrast, resolution and orientation of the input images, and thus do not generalize to diverse clinical acquisition protocols – even within sites. In this article, we present SynthSR, a method to train a CNN that receives one or more scans with spaced slices, acquired with different contrast, resolution and orientation, and produces an isotropic scan of canonical contrast (typically a 1 mm MP-RAGE). The presented method does not require any preprocessing, beyond rigid coregistration of the input scans. Crucially, SynthSR trains on synthetic input images generated from 3D segmentations, and can thus be used to train CNNs for any combination of contrasts, resolutions and orientations without high-resolution real images of the input contrasts. We test the images generated with SynthSR in an array of common downstream analyses, and show that they can be reliably used for subcortical segmentation and volumetry, image registration (e.g., for tensor-based morphometry), and, if some image quality requirements are met, even cortical thickness morphometry. The source code is publicly available at https://github.com/BBillot/SynthSR .

Alzheimer's disease (AD) is known as one of the major causes of dementia and is characterized by slow progression over several years, with no treatments or available medicines. In this regard, there have been efforts to identify the risk of developing AD in its earliest time. While many of the previous works considered cross-sectional analysis, more recent studies have focused on the diagnosis and prognosis of AD with longitudinal or time series data in a way of disease progression modeling. Under the same problem settings, in this work, we propose a novel computational framework that can predict the phenotypic measurements of MRI biomarkers and trajectories of clinical status along with cognitive scores at multiple future time points. However, in handling time series data, it generally faces many unexpected missing observations. In regard to such an unfavorable situation, we define a secondary problem of estimating those missing values and tackle it in a systematic way by taking account of temporal and multivariate relations inherent in time series data. Concretely, we propose a deep recurrent network that jointly tackles the four problems of (i) missing value imputation, (ii) phenotypic measurements forecasting, (iii) trajectory estimation of a cognitive score, and (iv) clinical status prediction of a subject based on his/her longitudinal imaging biomarkers. Notably, the learnable parameters of all the modules in our predictive models are trained in an end-to-end manner by taking the morphological features and cognitive scores as input, with our circumspectly defined loss function. In our experiments over The Alzheimers Disease Prediction Of Longitudinal Evolution (TADPOLE) challenge cohort, we measured performance for various metrics and compared our method to competing methods in the literature. Exhaustive analyses and ablation studies were also conducted to better confirm the effectiveness of our method.

The hippocampus has been extensively studied in various neuropsychiatric disorders throughout the lifespan. However, inconsistent results have been reported with respect to which subfield volumes are most related to age. Here, we investigate whether these discrepancies may be explained by experimental design differences that exist between studies. Multiple datasets were used to collect 1690 magnetic resonance scans from healthy individuals aged 18-95 years old. Standard T1-weighted (T1w; MPRAGE sequence, 1 mm3 voxels), high-resolution T2-weighted (T2w; SPACE sequence, 0.64 mm3 voxels) and slab T2-weighted (Slab; 2D turbo spin echo, 0.4 x 0.4 x 2 mm3 voxels) images were acquired. The MAGeT Brain algorithm was used for segmentation of the hippocampal grey matter (GM) subfields and peri-hippocampal white matter (WM) subregions. Linear mixed-effect models and Akaike information criterion were used to examine linear, second or third order natural splines relationship between hippocampal volumes and age. We demonstrated that stratum radiatum/lacunosum/moleculare and fornix subregions expressed the highest relative volumetric decrease, while the cornus ammonis 1 presented a relative volumetric preservation of its volume with age. We also found that volumes extracted from slab images were often underestimated and demonstrated different age-related relationships compared to volumes extracted from T1w and T2w images. The current work suggests that although T1w, T2w and slab derived subfield volumetric outputs are largely homologous, modality choice plays a meaningful role in the volumetric estimation of the hippocampal subfields.