Recently, much attention has been focused on the definition and structure of the hippocampus and its subfields, while the projections from the hippocampus have been relatively understudied. Here, we derive a reliable protocol for manual segmentation of hippocampal white matter regions (alveus, fimbria, and fornix) using high-resolution magnetic resonance images that are complementary to our previous definitions of the hippocampal subfields, both of which are freely available at https://github.com/cobralab/atlases. Our segmentation methods demonstrated high inter- and intra-rater reliability, were validated as inputs in automated segmentation, and were used to analyze the trajectory of these regions in both healthy aging (OASIS), and Alzheimer's disease (AD) and mild cognitive impairment (MCI; using ADNI). We observed significant bilateral decreases in the fornix in healthy aging while the alveus and cornu ammonis (CA) 1 were well preserved (all p's<0.006). MCI and AD demonstrated significant decreases in fimbriae and fornices. Many hippocampal subfields exhibited decreased volume in both MCI and AD, yet no significant differences were found between MCI and AD cohorts themselves. Our results suggest a neuroprotective or compensatory role for the alveus and CA1 in healthy aging and suggest that an improved understanding of the volumetric trajectories of these structures is required.

Linear registration to a standard space is one of the major steps in processing and analyzing magnetic resonance images (MRIs) of the brain. Here we present an overview of linear stereotaxic MRI registration and compare the performance of 5 publicly available and extensively used linear registration techniques in medical image analysis.

Neuroimaging studies typically adopt a common feature space for all data, which may obscure aspects of neuroanatomy only observable in subsets of a population, e.g. cortical folding patterns unique to individuals or shared by close relatives. Here, we propose to model individual variability using a distinctive keypoint signature: a set of unique, localized patterns, detected automatically in each image by a generic saliency operator. The similarity of an image pair is then quantified by the proportion of keypoints they share using a novel Jaccard-like measure of set overlap. Experiments demonstrate the keypoint method to be highly efficient and accurate, using a set of 7536 T1-weighted MRIs pooled from four public neuroimaging repositories, including twins, non-twin siblings, and 3334 unique subjects. All same-subject image pairs are identified by a similarity threshold despite confounds including aging and neurodegenerative disease progression. Outliers reveal previously unknown data labeling inconsistencies, demonstrating the usefulness of the keypoint signature as a computational tool for curating large neuroimage datasets.

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 in AD phenotypes. However, their utility in predicting the progression of pre-symptomatic AD in a population-based setting has not yet been investigated. In this study, we investigated if the disease timelines constructed in a case-controlled setting, with subjects stratified according to APOE status, are generalizable to a population-based cohort, and if progression along these disease timelines is predictive of AD. Seven volumetric biomarkers derived from structural MRI were considered. We estimated APOE-specific disease timelines of changes in these biomarkers using a recently proposed method called co-initialized discriminative event-based modeling (co-init DEBM). This method can also estimate a disease stage for new subjects by calculating their position along the disease 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 diagnostic and prognostic value of the disease stage in the two cohorts. Furthermore, we investigated if the rate of change of disease stage in RS participants with longitudinal MRI data was predictive of AD. In ADNI, the estimated disease timeslines 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 AUC of 0.83 in both APOEϵ4 non-carriers and carriers. In the RS cohort, we obtained an AUC of 0.83 and 0.85 in ϵ4 non-carriers and carriers, respectively. Progression along the disease timelines as estimated by the rate of change of disease stage showed a significant difference (p<0.005) for subjects with pre-symptomatic AD as compared to the 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 of change, could achieve. Our results suggest that co-init DEBM trained on case-controlled data is generalizable to a population-based cohort setting and that progression along the disease timelines is predictive of the development of AD in the general population. We expect that this approach can help to identify at-risk individuals from the general population for targeted clinical trials as well as to provide biomarker based objective assessment in such trials.

Alzheimer's disease (AD) is the most common form of dementia and is phenotypically heterogeneous. APOE is a triallelic gene which correlates with phenotypic heterogeneity in AD. In this work, we determined the effect of APOE alleles on the disease progression timeline of AD using a discriminative event-based model (DEBM). Since DEBM is a data-driven model, stratification into smaller disease subgroups would lead to more inaccurate models as compared to fitting the model on the entire dataset. Hence our secondary aim is to propose and evaluate novel approaches in which we split the different steps of DEBM into group-aspecific and group-specific parts, where the entire dataset is used to train the group-aspecific parts and only the data from a specific group is used to train the group-specific parts of the DEBM. We performed simulation experiments to benchmark the accuracy of the proposed approaches and to select the optimal approach. Subsequently, the chosen approach was applied to the baseline data of 417 cognitively normal, 235 mild cognitively impaired who convert to AD within 3 years, and 342 AD patients from the Alzheimers Disease Neuroimaging Initiative (ADNI) dataset to gain new insights into the effect of APOE carriership on the disease progression timeline of AD. In the ε4 carrier group, the model predicted with high confidence that CSF Amyloidβ42 and the cognitive score of Alzheimer's Disease Assessment Scale (ADAS) are early biomarkers. Hippocampus was the earliest volumetric biomarker to become abnormal, closely followed by the CSF Phosphorylated Tau181 (PTAU) biomarker. In the homozygous ε3 carrier group, the model predicted a similar ordering among CSF biomarkers. However, the volume of the fusiform gyrus was identified as one of the earliest volumetric biomarker. While the findings in the ε4 carrier and the homozygous ε3 carrier groups fit the current understanding of progression of AD, the finding in the ε2 carrier group did not. The model predicted, with relatively low confidence, CSF Neurogranin as one of the earliest biomarkers along with cognitive score of Mini-Mental State Examination (MMSE). Amyloid β42 was found to become abnormal after PTAU. The presented models could aid understanding of the disease, and in selecting homogeneous group of presymptomatic subjects at-risk of developing symptoms for clinical trials.

We introduce a theoretical framework for estimating, comparing and interpreting mechanistic hypotheses on long term protein propagation across brain networks in neurodegenerative disorders (ND). The model is expressed within a Bayesian non-parametric regression setting, where mechanisms of protein dynamics are inferred by means of gradient matching on dynamical systems (DS). The Bayesian formalism, combined with stochastic variational inference, naturally allows for model comparison via assessment of model evidence, while providing uncertainty quantification of causal relationship underlying protein progressions. When applied to in-vivo AV45-PET brain imaging data measuring topographic amyloid deposition in Alzheimer's disease (AD), our model identified the mechanisms of accumulation, clearance and propagation as the best suited DS for bio-mechanical description of amyloid dynamics in AD, enabling realistic and accurate personalized simulation of amyloidosis.

Prior work on Alzheimer's Disease (AD) has demonstrated that convolutional neural networks (CNNs) can leverage the high-dimensional image information for diagnosing patients. Beside such data-driven approaches, many established biomarkers exist and are typically represented as tabular data, such as demographics, genetic alterations, or laboratory measurements from cerebrospinal fluid. However, little research has focused on the effective integration of tabular data into existing CNN architectures to improve patient diagnosis. We introduce the Dynamic Affine Feature Map Transform (DAFT), a general-purpose module for CNNs that incites or represses high-level concepts learned from a 3D image by conditioning feature maps of a convolutional layer on both a patient's image and tabular clinical information. This is achieved by using an auxiliary neural network that outputs a scaling factor and offset to dynamically apply an affine transformation to the feature maps of a convolutional layer. In our experiments on AD diagnosis and time-to-dementia prediction, we show that the DAFT is highly effective in combining 3D image and tabular information by achieving a mean balanced accuracy of 0.622 for diagnosis, and mean c-index of 0.748 for time-to-dementia prediction, thus outperforming all baseline methods. Finally, our extensive ablation study and empirical experiments reveal that the performance improvement due to the DAFT is robust with respect to many design choices.

Non-rigid cortical registration is an important and challenging task due to the geometric complexity of the human cortex and the high degree of inter-subject variability. A conventional solution is to use a spherical representation of surface properties and perform registration by aligning cortical folding patterns in that space. This strategy produces accurate spatial alignment, but often requires high computational cost. Recently, convolutional neural networks (CNNs) have demonstrated the potential to dramatically speed up volumetric registration. However, due to distortions introduced by projecting a sphere to a 2D plane, a direct application of recent learning-based methods to surfaces yields poor results. In this study, we present SphereMorph, a diffeomorphic registration framework for cortical surfaces using deep networks that addresses these issues. SphereMorph uses a UNet-style network associated with a spherical kernel to learn the displacement field and warps the sphere using a modified spatial transformer layer. We propose a resampling weight in computing the data fitting loss to account for distortions introduced by polar projection, and demonstrate the performance of our proposed method on two tasks, including cortical parcellation and group-wise functional area alignment. The experiments show that the proposed SphereMorph is capable of modeling the geometric registration problem in a CNN framework and demonstrate superior registration accuracy and computational efficiency. The source code of SphereMorph will be released to the public upon acceptance of this manuscript at https://github.com/voxelmorph/spheremorph.

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.

To perform a joint analysis of multivariate neuroimaging phenotypes and candidate genetic markers obtained from longitudinal studies, we develop a Bayesian longitudinal low-rank regression (L2R2) model. The L2R2 model integrates three key methodologies: a low-rank matrix for approximating the high-dimensional regression coefficient matrices corresponding to the genetic main effects and their interactions with time, penalized splines for characterizing the overall time effect, and a sparse factor analysis model coupled with random effects for capturing within-subject spatio-temporal correlations of longitudinal phenotypes. Posterior computation proceeds via an efficient Markov chain Monte Carlo algorithm. Simulations show that the L2R2 model outperforms several other competing methods. We apply the L2R2 model to investigate the effect of single nucleotide polymorphisms (SNPs) on the top 10 and top 40 previously reported Alzheimer disease-associated genes. We also identify associations between the interactions of these SNPs with patient age and the tissue volumes of 93 regions of interest from patients' brain images obtained from the Alzheimer's Disease Neuroimaging Initiative.

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.

While aggregation of neuroimaging datasets from multiple sites and scanners can yield increased statistical power, it also presents challenges due to systematic scanner effects. This unwanted technical variability can introduce noise and bias into estimation of biological variability of interest. We propose a method for harmonizing longitudinal multi-scanner imaging data based on ComBat, a method originally developed for genomics and later adapted to cross-sectional neuroimaging data. Using longitudinal cortical thickness measurements from 663 participants in the Alzheimer's Disease Neuroimaging Initiative (ADNI) study, we demonstrate the presence of additive and multiplicative scanner effects in various brain regions. We compare estimates of the association between diagnosis and change in cortical thickness over time using three versions of the ADNI data: unharmonized data, data harmonized using cross-sectional ComBat, and data harmonized using longitudinal ComBat. In simulation studies, we show that longitudinal ComBat is more powerful for detecting longitudinal change than cross-sectional ComBat and controls the type I error rate better than unharmonized data with scanner included as a covariate. The proposed method would be useful for other types of longitudinal data requiring harmonization, such as genomic data, or neuroimaging studies of neurodevelopment, psychiatric disorders, or other neurological diseases.