Thomas J. Grabowski

Sparse optimization in feature selection: application in neuroimaging

Feature selection plays an important role in the successful application of machine learning techniques to large real-world datasets. Avoiding model overfitting, especially when the number of features far exceeds the number of observations, requires selecting informative features and/or eliminating irrelevant ones. Searching for an optimal subset of features can be computationally expensive. Functional magnetic resonance imaging (fMRI) produces datasets with such characteristics creating challenges for applying machine learning techniques to classify cognitive states based on fMRI data. In this study, we present an embedded feature selection framework that integrates sparse optimization for regularization (or sparse regularization) and classification. This optimization approach attempts to maximize training accuracy while simultaneously enforcing sparsity by penalizing the objective function for the coefficients of the features. This process allows many coefficients to become zero, which effectively eliminates their corresponding features from the classification model. To demonstrate the utility of the approach, we apply our framework to three different real-world fMRI datasets. The results show that regularized classifiers yield better classification accuracy, especially when the number of initial features is large. The results further show that sparse regularization is key to achieving scientifically-relevant generalizability and functional localization of classifier features. The approach is thus highly suited for analysis of fMRI data.

Multimodal Characterization of the Late Effects of Traumatic Brain Injury: A Methodological Overview of the Late Effects of Traumatic Brain Injury Project

Epidemiological studies suggest that a single moderate-to-severe traumatic brain injury (TBI) is associated with an increased risk of neurodegenerative disease, including Alzheimers disease (AD) and Parkinsons disease (PD). Histopathological studies describe complex neurodegenerative pathologies in individuals exposed to single moderate-to-severe TBI or repetitive mild TBI, including chronic traumatic encephalopathy (CTE). However, the clinicopathological links between TBI and post-traumatic neurodegenerative diseases such as AD, PD, and CTE remain poorly understood. Here, we describe the methodology of the Late Effects of TBI (LETBI) study, whose goals are to characterize chronic post-traumatic neuropathology and to identify in vivo biomarkers of post-traumatic neurodegeneration. LETBI participants undergo extensive clinical evaluation using National Institutes of Health TBI Common Data Elements, proteomic and genomic analysis, structural and functional magnetic resonance imaging (MRI), and prospective consent for brain donation. Selected brain specimens undergo ultra-high resolution ex vivo MRI and histopathological evaluation including whole-mount analysis. Co-registration of ex vivo and in vivo MRI data enables identification of ex vivo lesions that were present during life. In vivo signatures of postmortem pathology are then correlated with cognitive and behavioral data to characterize the clinical phenotype(s) associated with pathological brain lesions. We illustrate the study methods and demonstrate proof of concept for this approach by reporting results from the first LETBI participant, who despite the presence of multiple in vivo and ex vivo pathoanatomic lesions had normal cognition and was functionally independent until her mid-80s. The LETBI project represents a multidisciplinary effort to characterize post-traumatic neuropathology and identify in vivo signatures of postmortem pathology in a prospective study.

Findings of Vascular Brain Injury and Structural Loss from Cranial Magnetic Resonance Imaging in Elderly American Indians: The Strong Heart Study.

Background: The Cerebrovascular Disease and its Consequences in American Indians study conducted cranial MRI examination of surviving participants of the Strong Heart Study, a longitudinal cohort of elderly American Indians. Methods: Of the 1,033 recruited participants, some were unable to complete the MRI (n = 22), some scans were unusable due to participant motion or technical errors (n = 13), and one community withdrew consent after data collection (n = 209), leaving 789 interpretable MRI scan images. Six image sequences were obtained in contiguous slices on 1.5T scanners. Neuroradiologists graded white matter hyperintensities (WMH), sulci, and ventricles on a 0- to 9-point scale, and recorded the presence of infarcts and hemorrhages. Intracranial, brain, hippocampal, and WMH volumes were estimated by automated image processing. Results: The median scores for graded measures were 2 (WMH) and 3 (sulci, ventricles). About one-third of participants had lacunar (20%) or other infarcts (13%); few had hemorrhages (5.7%). Findings of cortical atrophy were also prevalent. Statistical analyses indicated significant associations between older age and findings of vascular injury and atrophy; male gender was associated with findings of cortical atrophy. Conclusions: Vascular brain injury is the likely explanation in this elderly American Indian population for brain infarcts, hemorrhages, WMH grade, and WMH volume. Although vascular brain injury may play a role in other findings, independent degenerative other disease processes may underlie abnormal sulcal widening, ventricular enlargement, hippocampal volume, and total brain volume. Further examination of risk factors and outcomes with these findings may expand the understanding of neurological conditions in this understudied population.

Quantitative cerebrovascular pathology in a community-based cohort of older adults

Cerebrovascular disease, especially small vessel pathology, is the leading comorbidity in degenerative disorders. We applied arterial spin labeling and cerebrovascular reserve (CVR) imaging to quantify small vessel disease and study its effect on cognitive symptoms in nondemented older adults from a community-based cohort. We evaluated baseline cerebral blood flow (CBF) using arterial spin labeling and percent signal change as a marker of CVR using blood-oxygen level-dependent imaging following a breath-hold stimulus. Measurements were performed in and near white matter hyperintensities, which are currently the standard to assess severity of vascular pathology. We show that similar to other studies (1) CBF and CVR are markedly reduced in the hyperintensities as well as in the tissue surrounding them, indicating susceptibility to infarction; (2) low CBF and CVR are significantly correlated with poor cognitive performance; and (3) in addition, compared to a 58.4% reduction in CBF, larger exhaustion (79.3%) of CVR was observed in the hyperintensities with a faster, nonlinear rate of decline. We conclude that CVR may be a more sensitive biomarker of small vessel disease than CBF.

Automated connectivity-based cortical mapping using registration-constrained classification

An important goal in neuroscience has been to map the surface of the human brain, and many researchers have developed sophisticated methods to parcellate the cortex. However, many of these methods stop short of developing a framework to apply existing cortical maps to new subjects in a consistent fashion. The computationally complex step is often the initial mapping of a large set of brains, and it is inefficient to repeat these processes for every new data sample. In this analysis, we propose the use of a library of training brains to build a statistical model of the parcellated cortical surface and to act as templates for mapping new data. We train classifiers on training data sampled from local neighborhoods on the cortical surface, using features derived from training brain connectivity information, and apply these classifiers to map the surfaces of previously unseen brains. We demonstrate the performance of 3 different classifiers, each trained on 3 different types of training features, to accurately predict the map of new brain surfaces.

Leptomeninges-Derived Induced Pluripotent Stem Cells and Directly Converted Neurons From Autopsy Cases With Varying Neuropathologic Backgrounds

Abstract Patient-specific stem cell technology from skin and other biopsy sources has transformed in vitro models of neurodegenerative disease, permitting interrogation of the effects of complex human genetics on neurotoxicity. However, the neuropathologic changes that underlie cognitive and behavioral phenotypes can only be determined at autopsy. To better correlate the biology of derived neurons with age-related and neurodegenerative changes, we generated leptomeningeal cell lines from well-characterized research subjects that have undergone comprehensive postmortem neuropathologic examinations. In a series of proof of principle experiments, we reprogrammed autopsy leptomeningeal cell lines to human-induced pluripotent stem cells (hiPSCs) and differentiated these into neurons. We show that leptomeningeal-derived hiPSC lines can be generated from fresh and frozen leptomeninges, are pluripotent, and retain the karyotype of the starting cell population. Additionally, neurons differentiated from these hiPSCs are functional and produce measurable Alzheimer disease-relevant analytes (Aβ and Tau). Finally, we used direct conversion protocols to transdifferentiate leptomeningeal cells to neurons. These resources allow the generation of in vitro models to test mechanistic hypotheses as well as diagnostic and therapeutic strategies in association with neuropathology, clinical and cognitive data, and biomarker studies, aiding in the study of late-onset Alzheimer disease and other age-related neurodegenerative diseases.

Celebrating the 125th anniversary of the American Psychological Association: A quarter century of neuropsychology.

OBJECTIVEnThe American Psychological Association (APA) celebrated its 125th anniversary in 2017. As part of this celebration, the APA journal Neuropsychology has published in its November 2017 issue 11 papers describing some of the advances in the field of neuropsychology over the past 25 years.nnnMETHODnThe papers address three broad topics: assessment and intervention, brain imaging, and theory and methods.nnnRESULTSnThe papers describe the rise of new assessment and intervention technologies, the impact of evidence for neuroplasticity on neurorehabilitation. Examples of the use of mathematical models of cognition to investigate latent neurobehavioral processes, the development of the field of neuropsychology in select international countries, the increasing sophistication of brain imaging methods, the recent evidence for localizationist and connectionist accounts of neurobehavioral functioning, the advances in neurobehavioral genomics, and descriptions of newly developed statistical models of longitudinal change.nnnCONCLUSIONnTogether the papers convey evidence of the vibrant growth in the field of neuropsychology over the quarter century since APAs 100th anniversary in 1992. (PsycINFO Database Record

Electrocorticography and the early maturation of high-frequency suppression within the default mode network

OBJECTIVE The acquisition and refinement of cognitive and behavioral skills during development is associated with the maturation of various brain oscillatory activities. Most developmental investigations have identified distinct patterns of low-frequency electrophysiological activity that are characteristic of various behavioral milestones. In this investigation, the authors focused on the cross-sectional developmental properties of high-frequency spectral power from the brains default mode network (DMN) during goal-directed behavior. METHODS The authors contrasted regionally specific, time-evolving high gamma power (HGP) in the lateral DMN cortex between 3 young children (age range 3-6 years) and 3 adults by use of electrocorticography (ECoG) recordings over the left perisylvian cortex during a picture-naming task. RESULTS Across all participants, a nearly identical and consistent response suppression of HGP, which is a functional signature of the DMN, was observed during task performance recordings acquired from ECoG electrodes placed over the lateral DMN cortex. This finding provides evidence of relatively early maturation of the DMN. Furthermore, only HGP relative to evoked alpha and beta band power showed this level of consistency across all participants. CONCLUSIONS Regionally specific, task-evoked suppression of the high-frequency components of the cortical power spectrum is established early in brain development, and this response may reflect the early maturation of specific cognitive and/or computational mechanisms.

Network Optimization of Functional Connectivity Within Default Mode Network Regions to Detect Cognitive Decline

The rapid aging of the world’s population is causing an increase in the prevalence of cognitive decline and degenerative brain disease in the elderly. Current diagnoses of amnestic and nonamnestic mild cognitive impairment, which may represent early stage Alzheimer’s disease or related degenerative conditions, are based on clinical grounds. The recent emergence of advanced network analyses of functional magnetic resonance imaging (fMRI) data taken at cognitive rest has provided insight that declining functional connectivity of the default mode network (DMN) may be correlated with neurological disorders, and particularly prodromal Alzheimer’s disease. The goal of this paper is to develop a network analysis technique using fMRI data to characterize transition stages from healthy brain aging to cognitive decline. Previous studies primarily focused on inter-nodal connectivity of the DMN and often assume functional homogeneity within each DMN region. In this paper, we develop a technique that focuses on identifying critical intra-nodal DMN connectivity by incorporating sparsity into connectivity modeling of the

Running Neuroimaging Applications on Amazon Web Services: How, When, and at What Cost?

k