Statistical Methods for High-dimensional Data with Complex Correlation Structure Applied to the Brain Dynamic Functional Connectivity Study PDF Download

Author: Maria Aleksandra Kudela
Publisher:
ISBN:
Category :
Languages : en
Pages : 238

View

Book Description
A popular non-invasive brain activity measurement method is based on the functional magnetic resonance imaging (fMRI). Such data are frequently used to study functional connectivity (FC) defined as statistical association among two or more anatomically distinct fMRI signals (Friston, 1994). FC has emerged in recent years as a valuable tool for providing a deeper understanding of neurodegenerative diseases and neuropsychiatric disorders, such as Alzheimer's disease and autism. Information about complex association structure in high-dimensional fMRI data is often discarded by a calculating an average across complex spatiotemporal processes without providing an uncertainty measure around it. First, we propose a non-parametric approach to estimate the uncertainty of dynamic FC (dFC) estimates. Our method is based on three components: an extension of a boot strapping method for multivariate time series, recently introduced by Jentsch and Politis (2015); sliding window correlation estimation; and kernel smoothing. Second, we propose a two-step approach to analyze and summarize dFC estimates from a task-based fMRI study of social-to-heavy alcohol drinkers during stimulation with avors. In the first step, we apply our method from the first paper to estimate dFC for each region subject combination. In the second step, we use semiparametric additive mixed models to account for complex correlation structure and model dFC on a population level following the study's experimental design. Third, we propose to utilize the estimated dFC to study the system's modularity defined as the mutually exclusive division of brain regions into blocks with intra-connectivity greater than the one obtained by chance. As a result, we obtain brain partition suggesting the existence of common functionally-based brain organization. The main contribution of our work stems from the combination of the methods from the fields of statistics, machine learning and network theory to provide statistical tools for studying brain connectivity from a holistic, multi-disciplinary perspective.

Author: Antonio Canale
Publisher: Springer
ISBN: 3030000397
Category : Mathematics
Languages : en
Pages : 156

View

Book Description
This volume presents a collection of peer-reviewed contributions arising from StartUp Research: a stimulating research experience in which twenty-eight early-career researchers collaborated with seven senior international professors in order to develop novel statistical methods for complex brain imaging data. During this meeting, which was held on June 25–27, 2017 in Siena (Italy), the research groups focused on recent multimodality imaging datasets measuring brain function and structure, and proposed a wide variety of methods for network analysis, spatial inference, graphical modeling, multiple testing, dynamic inference, data fusion, tensor factorization, object-oriented analysis and others. The results of their studies are gathered here, along with a final contribution by Michele Guindani and Marina Vannucci that opens new research directions in this field. The book offers a valuable resource for all researchers in Data Science and Neuroscience who are interested in the promising intersections of these two fundamental disciplines.

Author: Nicole Lazar
Publisher: Springer Science & Business Media
ISBN: 0387781919
Category : Medical
Languages : en
Pages : 302

View

Book Description
The study of brain function is one of the most fascinating pursuits of m- ern science. Functional neuroimaging is an important component of much of the current research in cognitive, clinical, and social psychology. The exci- ment of studying the brain is recognized in both the popular press and the scienti?c community. In the pages of mainstream publications, including The New York Times and Wired, readers can learn about cutting-edge research into topics such as understanding how customers react to products and - vertisements (“If your brain has a ‘buy button,’ what pushes it?”, The New York Times,October19,2004),howviewersrespondtocampaignads(“Using M. R. I. ’s to see politics on the brain,” The New York Times, April 20, 2004; “This is your brain on Hillary: Political neuroscience hits new low,” Wired, November 12,2007),howmen and womenreactto sexualstimulation (“Brain scans arouse researchers,”Wired, April 19, 2004), distinguishing lies from the truth (“Duped,” The New Yorker, July 2, 2007; “Woman convicted of child abuse hopes fMRI can prove her innocence,” Wired, November 5, 2007), and even what separates “cool” people from “nerds” (“If you secretly like Michael Bolton, we’ll know,” Wired, October 2004). Reports on pathologies such as autism, in which neuroimaging plays a large role, are also common (for - stance, a Time magazine cover story from May 6, 2002, entitled “Inside the world of autism”).

Author: Anita Meghan Sinha
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description
Magnetic resonance imaging (MRI) plays an integral role in the study, diagnosis and treatment of neurological diseases. Neuroimaging analyses involve high-dimensional, large-scale data that contain rich spatial and temporal information about the dynamic and integrated systems in the brain. Therefore, it has become imperative to develop and optimize analytical approaches drawn from engineering and mathematics to more precisely model these complex patterns and interactions, which will advance our understanding of functional brain organization in health and disease. Chapter 1 provides an overview and background of MRI, with a particular focus on the use of resting-state functional magnetic resonance imaging (rs-fMRI) to capture and characterize brain connectivity. Previous work of statistical methods developed for fMRI analysis are reviewed. Chapter 2 presents an analysis of changes in functional connectivity and behavioral outcomes in patients of stroke who undergo brain-computer interface (BCI) interventional therapy. This work employs a widely used network-based inference method for fMRI analysis that serves as motivation for subsequent work to overcome statistical challenges associated with its use to more effectively model and characterize brain network dynamics and organization in a robust manner. Chapter 3 presents a novel application of differential covariance trajectory analysis as promising framework for brain network modeling using rs-fMRI data. The proposed algorithm models functional connectivity as trajectories on the manifold and employs a localization procedure to search over and identify subsets of first- and second-order differences in brain connectivity features between patients with Temporal Lobe Epilepsy (TLE) and healthy control subjects. Chapter 4 extends the work presented in the previous chapter to apply the combined differential covariance trajectory and scan statistics framework to characterize the Alzheimer's Disease connectome. We demonstrate the utility and robustness of this method to study altered brain network organization in large-scale functional networks in a different and older clinical population, which is notably of smaller sample size, where the statistical signal may be weak. Chapter 5 discusses conclusions and key takeaways of the work, along with potential future avenues of research.

Author: Matthias Bethge
Publisher: Frontiers E-books
ISBN: 2889190129
Category :
Languages : en
Pages : 209

View

Book Description
Modern recording techniques such as multi-electrode arrays and 2-photon imaging are capable of simultaneously monitoring the activity of large neuronal ensembles at single cell resolution. This makes it possible to study the dynamics of neural populations of considerable size, and to gain insights into their computations and functional organization. The key challenge with multi-electrode recordings is their high-dimensional nature. Understanding this kind of data requires powerful statistical techniques for capturing the structure of the neural population responses and their relation with external stimuli or behavioral observations. Contributions to this Research Topic should advance statistical modeling of neural populations. Questions of particular interest include: 1. What classes of statistical methods are most useful for modeling population activity? 2. What are the main limitations of current approaches, and what can be done to overcome them? 3. How can statistical methods be used to empirically test existing models of (probabilistic) population coding? 4. What role can statistical methods play in formulating novel hypotheses about the principles of information processing in neural populations? This Research Topic is connected to a one day workshop at the Computational Neuroscience Meeting 2009 in Berlin (http://www.cnsorg.org/2009/workshops.shtml and http://www.kyb.tuebingen.mpg.de/bethge/workshops/cns2009/)

Author: Petter Holme
Publisher: Springer
ISBN: 3642364616
Category : Science
Languages : en
Pages : 356

View

Book Description
The concept of temporal networks is an extension of complex networks as a modeling framework to include information on when interactions between nodes happen. Many studies of the last decade examine how the static network structure affect dynamic systems on the network. In this traditional approach the temporal aspects are pre-encoded in the dynamic system model. Temporal-network methods, on the other hand, lift the temporal information from the level of system dynamics to the mathematical representation of the contact network itself. This framework becomes particularly useful for cases where there is a lot of structure and heterogeneity both in the timings of interaction events and the network topology. The advantage compared to common static network approaches is the ability to design more accurate models in order to explain and predict large-scale dynamic phenomena (such as, e.g., epidemic outbreaks and other spreading phenomena). On the other hand, temporal network methods are mathematically and conceptually more challenging. This book is intended as a first introduction and state-of-the art overview of this rapidly emerging field.

Author: Vito Paolo Pastore
Publisher: Springer Nature
ISBN: 3030590429
Category : Technology & Engineering
Languages : en
Pages : 87

View

Book Description
This book describes a set of novel statistical algorithms designed to infer functional connectivity of large-scale neural assemblies. The algorithms are developed with the aim of maximizing computational accuracy and efficiency, while faithfully reconstructing both the inhibitory and excitatory functional links. The book reports on statistical methods to compute the most significant functional connectivity graph, and shows how to use graph theory to extract the topological features of the computed network. A particular feature is that the methods used and extended at the purpose of this work are reported in a fairly completed, yet concise manner, together with the necessary mathematical fundamentals and explanations to understand their application. Furthermore, all these methods have been embedded in the user-friendly open source software named SpiCoDyn, which is also introduced here. All in all, this book provides researchers and graduate students in bioengineering, neurophysiology and computer science, with a set of simplified and reduced models for studying functional connectivity in in silico biological neuronal networks, thus overcoming the complexity of brain circuits.

Author: Ke Zeng
Publisher:
ISBN:
Category :
Languages : en
Pages : 256

View

Book Description
Building quantitative models to summarize the structural variability of the human brain is an essential task in brain image analysis. Such quantitative models can be used to measure the normative variation of healthy brains, to capture their change over time, and to find imaging patterns of a diseased group. These model can be further applied to individual brain scans for tissue segmentation, lesion delineation, abnormality detection and image registration. A common approach to derive a representation of a population is through the use of atlases (i.e., characteristic brains) that are either manually determined or automatically inferred. However, atlases are first-order statistical measures that do not convey information about the amount and direction of variability within a population and are therefore inadequate for many applications. Most previous works on statistical modeling of imaging data have resorted to voxel-based constructions in which image values at different voxels are assumed to be statistically independent. Although voxel-based methods can identify structural variations that are well localized, they are myopic to correlations between different regions and cannot capture any global patterns of the underlying data. Contrarily, classical multivariate statistical methods can be useful for finding the most dominant trends of variability. However, they are incapable of providing a statistically consistent estimate of the full covariance structure or the joint probabilistic density function of high-dimensional image data with a limited amount of samples. In this thesis, we introduce a multivariate framework for learning probability distributions over high-dimensional image data to capture the inter-subject structural variability of the brain. Specifically, we adopt the divide and conquer strategy by breaking the challenging task of learning high-dimensional image data into a collection of smaller, more tractable problems. In Chapter 2, we present a generative model built upon the aforementioned strategy to capture normative variations of image appearance. The model is incorporated within a novel framework for locating imaging abnormalities. In particular, a 3-Dimensional image volume is modeled as an ensemble of overlapping local regions. A sparse probabilistic model is used to approximate the marginal distribution of local intensity patterns, while pairwise potentials are incorporated to account for correlations across local regions. To tackle the difficulties associated with registering an image of a healthy brain to a scan of a diseased brain, we develop an iterative procedure that interleaves abnormality detection with registration. The method was evaluated using simulated data and tested using images with real lesions. Experimental results demonstrate that the framework can achieve accurate registration and abnormality detection simultaneously.In Chapter 3, we introduce a generative probabilistic model of high-dimensional spatial transformations. To make use of linear statistical methods while preserving diffeomorphisms, we adopt the Log-Euclidean framework and parametrize diffeomorphisms as exponentials of stationary velocity fields. Following the divide and conquer principle, we treat a velocity field as a collection of local velocities that reside in much lower-dimensional sub-spaces. Differing from the model for image appearances, principal component analysis is used to estimate the covariance structure for each local velocity and canonical correlation analysis is used to learn the dependencies between pairs of local velocities. The learned model is used as the foundation of a statistically constrained diffeomorphic registration algorithm. The method was tested using both simulated and real data. The results indicate that the proposed model is able to capture the normative variations of deformations with sub-millimeter accuracy and that the learned statistical constraints lead to substantially more robust registration results in the presence of abnormalities. Lastly, in Chapter 4, we shift our attention to the segmentation of specific pathological structures in a supervised setting. In particular, we demonstrate how a generative model similar to the one described in Chapter 2 can be combined with discriminative learning techniques to form a hybrid segmentation framework. The hybrid method was validated using 132 scans of patients with high-grade gliomas. Quantitative evaluation of the segmentation shows that the hybrid approach outperforms both the baseline generative method and the baseline discriminative model.

Author: F. Gregory Ashby
Publisher: MIT Press
ISBN: 0262042681
Category : Medical
Languages : en
Pages : 569

View

Book Description
A guide to all aspects of experimental design and data analysis for fMRI experiments, completely revised and updated for the second edition. Functional magnetic resonance imaging (fMRI), which allows researchers to observe neural activity in the human brain noninvasively, has revolutionized the scientific study of the mind. An fMRI experiment produces massive amounts of highly complex data for researchers to analyze. This book describes all aspects of experimental design and data analysis for fMRI experiments, covering every step—from preprocessing to advanced methods for assessing functional connectivity—as well as the most popular multivariate approaches. The goal is not to describe which buttons to push in the popular software packages but to help researchers understand the basic underlying logic, the assumptions, the strengths and weaknesses, and the appropriateness of each method. The field of fMRI research has advanced dramatically in recent years, in both methodology and technology, and this second edition has been completely revised and updated. Six new chapters cover experimental design, functional connectivity analysis through the methods of psychophysiological interactions and beta-series regression, decoding using multi-voxel pattern analysis, dynamic causal modeling, and representational similarity analysis. Other chapters offer new material on recently discovered problems related to head movements, the multivariate GLM, meta-analysis, and other topics. All complex derivations now appear at the end of the relevant chapter to improve readability. A new appendix describes how to build a design matrix with effect coding for group analysis. As in the first edition, MATLAB code is provided with which readers can implement many of the methods described.

Author: Chang Yan
Publisher: Frontiers Media SA
ISBN: 2832540082
Category : Science
Languages : en
Pages : 127

View

Book Description
Wearable health devices have been an emerging technology that enables an ambulatory acquisition of physiological signals to monitor health status over a long time (hours/days/weeks/years) inside and outside clinical environments. Big data and deep learning, in particular, are receiving a lot of attention in this rapidly growing digital health community. A key benefit of deep learning is to analyze and learn massive amounts of data, which makes it especially valuable in healthcare since raw data is largely gathered from personalized wearable health devices. A wide range of users may benefit from unobstructed and even remote monitoring of pertinent or vital signs, which makes it easier to detect life-threatening diseases early, track the progression of pathologies and stress levels, evaluate the efficacy of therapies, provide low-cost and reliable diagnoses, etc. Today’s personal health devices have provided an amazing insight into people’s health and wellness, which allow clinicians to use these smart wearables to collect and analyze measuring data like electroencephalogram (EEG), electrocardiogram (ECG or EKG), respiration, heart rate, temperature level, blood oxygen, and blood pressure for health monitoring or clinical trials. This Research Topic mainly focuses on the technical revolution in wearable health systems, which aims to design more smart and useful wearables, contributing to a substantial change in the methodologies, applications, and algorithms of machine learning for wearable health devices. With the help of deep learning and sensor fusion capabilities from wearable health platforms, this data will be used more effectively, which can help to construct smart, novel, specific solutions to improve the quality of healthcare and capabilities of utilizing new deep learning technologies.