Joint Prediction and Classification of Brain Image Evolution Trajectories from Baseline Brain Image with Application to Early Dementia

Abstract

Despite the large body of existing neuroimaging-based studies on brain dementia, in particular mild cognitive impairment (MCI), modeling and predicting the early dynamics of dementia onset and development in healthy brains is somewhat overlooked in the literature. The majority of computer-aided diagnosis tools developed for classifying healthy and demented brains mainly rely on either using single timepoint or longitudinal neuroimaging data. Longitudinal brain imaging data offer a larger time window to better capture subtle brain changes in early MCI development, and its utilization has been shown to improve classification and prediction results. However, typical longitudinal studies are challenged by a limited number of acquisition timepoints and the absence of inter-subject matching between timepoints. To address this limitation, we propose a novel framework that learns how to predict the developmental trajectory of a brain image from a single acquisition timepoint (i.e., baseline), while classifying the predicted trajectory as 'healthy' or 'demented'. To do so, we first rigidly align all training images, then extract 'landmark patches' from training images. Next, to predict the patch-wise trajectory evolution from baseline patch, we propose two novel strategies. The first strategy learns in a supervised manner to select a few training atlas patches that best boost the classification accuracy of the target testing patch. The second strategy learns in an unsupervised manner to select the set of most similar training atlas patches to the target testing patch using multi-kernel patch manifold learning. Finally, we train a linear classifier for each predicted patch trajectory. To identify the final label of the target subject, we use majority voting to aggregate the labels assigned by our model to all landmark patches' trajectories. Our image prediction model boosted the classification performance by 14% without further leveraging any enhancing methods such as feature selection.

Residual Embedding Similarity-Based Network Selection for Predicting Brain Network Evolution Trajectory from a Single Observation

“…For instance, [3] predicted the multishape trajectory of the baby brain using neonatal MRI data. Similarly, [4] and [5] used MR images to predict brain image evolution trajectories for early dementia detection. Although pioneering, such works mainly focused on Euclidean structured data (i.e, images), which is a flat representation of the brain and does not reflect the connectivity patterns existing among brain regions encoded in brain networks (i.e, connectomes).…”

Foreseeing Brain Graph Evolution Over Time Using Deep Adversarial Network Normalizer

“…Previous studies have developed shape-based and image-based prediction frameworks using morphological features derived from brain MRI scans to foresee the brain evolution trajectory [5,6]. For instance, [6] used a representative shape selection method to predict longitudinal development of cortical surfaces and white matter fibers assuming that similar shapes at baseline timepoint will have similar developmental trajectories.…”

“…For instance, [6] used a representative shape selection method to predict longitudinal development of cortical surfaces and white matter fibers assuming that similar shapes at baseline timepoint will have similar developmental trajectories. Such an assumption has been also adopted in a landmark study [5], demonstrating the reliability of exploring similarities between baseline training and testing samples for predicting the evolution of brain MR image trajectory in patients diagnosed with mild cognitive impairment. Although these works proposed successful predictive frameworks for image-based brain evolution trajectory prediction and classification, these were solely restricted to investigating the brain as a surface or a 3D image.…”

“…To address these limitations, we tap into the nascent field of geometric deep learning aiming to learn representations of non-Euclidean objects such as graphs while preserving their geometry. Drawing inspiration from previous brain evolution predictive frameworks [5,6,7], we also assume the preservation of local sample (i.e., brain graph) neighborhood across different timepoints. As such, by learning the similarities between samples at baseline timepoint, one can identify the most similar training brain graphs to a testing brain graph.…”

Deep EvoGraphNet Architecture For Time-Dependent Brain Graph Data Synthesis From a Single Timepoint

“…Thereby, it is vital to undertake an early diagnosis of brain diseases [2], especially for neurodegenerative diseases such as dementia which was found to be irreversible if discovered at a late stage [3]. In this context, recent landmark studies [4,5] have suggested using the robust predictive abilities of machine learning to predict the time-dependent (i.e., longitudinal) evolution of both the healthy and the disordered brain. However, such works only focus on predicting the brain when modeled as an image or a surface, thereby overlooking a wide spectrum of brain dysconnectivity disorders that can be pinned down by modeling the brain as a graph (also called connectome) [6], where the connectivity weight between pairs of anatomical regions of interest (ROIs) becomes the feature of interest.…”